* [PATCH v2 01/62] x86/svm: Offer CPUID Faulting to AMD HVM guests as well
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 02/62] xen/x86: report domain id on cpuid Wei Liu
` (60 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
CPUID Faulting can be virtulised for HVM guests without hardware support,
meaning it can be offered to SVM guests.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
xen/arch/x86/hvm/svm/svm.c | 6 ++++++
xen/arch/x86/msr.c | 3 ++-
2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/xen/arch/x86/hvm/svm/svm.c b/xen/arch/x86/hvm/svm/svm.c
index c48fdfaa5d..9d9ad777d1 100644
--- a/xen/arch/x86/hvm/svm/svm.c
+++ b/xen/arch/x86/hvm/svm/svm.c
@@ -1786,6 +1786,12 @@ static void svm_vmexit_do_cpuid(struct cpu_user_regs *regs)
if ( (inst_len = __get_instruction_length(curr, INSTR_CPUID)) == 0 )
return;
+ if ( hvm_check_cpuid_faulting(curr) )
+ {
+ hvm_inject_hw_exception(TRAP_gp_fault, 0);
+ return;
+ }
+
guest_cpuid(curr, regs->eax, regs->ecx, &res);
HVMTRACE_5D(CPUID, regs->eax, res.a, res.b, res.c, res.d);
diff --git a/xen/arch/x86/msr.c b/xen/arch/x86/msr.c
index 31983edc54..187f8623a5 100644
--- a/xen/arch/x86/msr.c
+++ b/xen/arch/x86/msr.c
@@ -39,7 +39,8 @@ static void __init calculate_hvm_max_policy(void)
return;
/* 0x000000ce MSR_INTEL_PLATFORM_INFO */
- if ( boot_cpu_data.x86_vendor == X86_VENDOR_INTEL )
+ if ( boot_cpu_data.x86_vendor == X86_VENDOR_INTEL ||
+ boot_cpu_data.x86_vendor == X86_VENDOR_AMD )
{
dp->plaform_info.available = true;
dp->plaform_info.cpuid_faulting = true;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 02/62] xen/x86: report domain id on cpuid
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
2018-01-12 11:28 ` [PATCH v2 01/62] x86/svm: Offer CPUID Faulting to AMD HVM guests as well Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 03/62] tools/libxc: remove extraneous newline in xc_dom_load_acpi Wei Liu
` (59 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Use the ECX register of the hypervisor leaf 5. The EAX register on
this leaf is a flags field that can be used to notice the presence of
the domain id in ECX. Note that this is only available to HVM guests.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Use leaf 5 instead.
---
xen/arch/x86/traps.c | 5 +++++
xen/include/public/arch-x86/cpuid.h | 2 ++
2 files changed, 7 insertions(+)
diff --git a/xen/arch/x86/traps.c b/xen/arch/x86/traps.c
index db16a44417..a896053b2f 100644
--- a/xen/arch/x86/traps.c
+++ b/xen/arch/x86/traps.c
@@ -924,6 +924,11 @@ void cpuid_hypervisor_leaves(const struct vcpu *v, uint32_t leaf,
/* Indicate presence of vcpu id and set it in ebx */
res->a |= XEN_HVM_CPUID_VCPU_ID_PRESENT;
res->b = v->vcpu_id;
+
+ /* Indicate presence of domain id and set it in ecx */
+ res->a |= XEN_HVM_CPUID_DOMID_PRESENT;
+ res->c = d->domain_id;
+
break;
case 5: /* PV-specific parameters */
diff --git a/xen/include/public/arch-x86/cpuid.h b/xen/include/public/arch-x86/cpuid.h
index eb76875d0e..665c4b644d 100644
--- a/xen/include/public/arch-x86/cpuid.h
+++ b/xen/include/public/arch-x86/cpuid.h
@@ -94,12 +94,14 @@
* HVM-specific features
* Sub-leaf 0: EAX: Features
* Sub-leaf 0: EBX: vcpu id (iff EAX has XEN_HVM_CPUID_VCPU_ID_PRESENT flag)
+ * Sub-leaf 0: ECX: domain id (iff EAX has XEN_HVM_CPUID_DOMID_PRESENT flag)
*/
#define XEN_HVM_CPUID_APIC_ACCESS_VIRT (1u << 0) /* Virtualized APIC registers */
#define XEN_HVM_CPUID_X2APIC_VIRT (1u << 1) /* Virtualized x2APIC accesses */
/* Memory mapped from other domains has valid IOMMU entries */
#define XEN_HVM_CPUID_IOMMU_MAPPINGS (1u << 2)
#define XEN_HVM_CPUID_VCPU_ID_PRESENT (1u << 3) /* vcpu id is present in EBX */
+#define XEN_HVM_CPUID_DOMID_PRESENT (1u << 4) /* domid is present in ECX */
/*
* Leaf 6 (0x40000x05)
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 03/62] tools/libxc: remove extraneous newline in xc_dom_load_acpi
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
2018-01-12 11:28 ` [PATCH v2 01/62] x86/svm: Offer CPUID Faulting to AMD HVM guests as well Wei Liu
2018-01-12 11:28 ` [PATCH v2 02/62] xen/x86: report domain id on cpuid Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 04/62] tools/libelf: fix elf notes check for PVH guest Wei Liu
` (58 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
tools/libxc/xc_dom_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/libxc/xc_dom_core.c b/tools/libxc/xc_dom_core.c
index b5f316a1dc..303cb971e8 100644
--- a/tools/libxc/xc_dom_core.c
+++ b/tools/libxc/xc_dom_core.c
@@ -1078,7 +1078,7 @@ static int xc_dom_load_acpi(struct xc_dom_image *dom)
while ( (i < MAX_ACPI_MODULES) && dom->acpi_modules[i].length )
{
- DOMPRINTF("%s: %d bytes at address %" PRIx64 "\n", __FUNCTION__,
+ DOMPRINTF("%s: %d bytes at address %" PRIx64, __FUNCTION__,
dom->acpi_modules[i].length,
dom->acpi_modules[i].guest_addr_out);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 04/62] tools/libelf: fix elf notes check for PVH guest
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (2 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 03/62] tools/libxc: remove extraneous newline in xc_dom_load_acpi Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 05/62] tools/libxc: Multi modules support Wei Liu
` (57 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
PVH only requires PHYS32_ENTRY to be set. Return immediately if that's
the case.
Also remove the printk in pvh_load_kernel.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
xen/arch/x86/hvm/dom0_build.c | 4 ----
xen/common/libelf/libelf-dominfo.c | 9 ++++++++-
2 files changed, 8 insertions(+), 5 deletions(-)
diff --git a/xen/arch/x86/hvm/dom0_build.c b/xen/arch/x86/hvm/dom0_build.c
index d696d4cd97..7d177a1d7d 100644
--- a/xen/arch/x86/hvm/dom0_build.c
+++ b/xen/arch/x86/hvm/dom0_build.c
@@ -484,10 +484,6 @@ static int __init pvh_load_kernel(struct domain *d, const module_t *image,
return -EINVAL;
}
- printk("OS: %s version: %s loader: %s bitness: %s\n", parms.guest_os,
- parms.guest_ver, parms.loader,
- elf_64bit(&elf) ? "64-bit" : "32-bit");
-
/* Copy the OS image and free temporary buffer. */
elf.dest_base = (void *)(parms.virt_kstart - parms.virt_base);
elf.dest_size = parms.virt_kend - parms.virt_kstart;
diff --git a/xen/common/libelf/libelf-dominfo.c b/xen/common/libelf/libelf-dominfo.c
index 829d5176a9..508f08db42 100644
--- a/xen/common/libelf/libelf-dominfo.c
+++ b/xen/common/libelf/libelf-dominfo.c
@@ -381,6 +381,13 @@ static elf_errorstatus elf_xen_note_check(struct elf_binary *elf,
return 0;
}
+ /* PVH only requires one ELF note to be set */
+ if ( parms->phys_entry != UNSET_ADDR32 )
+ {
+ elf_msg(elf, "ELF: Found PVH image\n");
+ return 0;
+ }
+
/* Check the contents of the Xen notes or guest string. */
if ( ((strlen(parms->loader) == 0) ||
strncmp(parms->loader, "generic", 7)) &&
@@ -389,7 +396,7 @@ static elf_errorstatus elf_xen_note_check(struct elf_binary *elf,
{
elf_err(elf,
"ERROR: Will only load images built for the generic loader or Linux images"
- " (Not '%.*s' and '%.*s')\n",
+ " (Not '%.*s' and '%.*s') or with PHYS32_ENTRY set\n",
(int)sizeof(parms->loader), parms->loader,
(int)sizeof(parms->guest_os), parms->guest_os);
return -1;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 05/62] tools/libxc: Multi modules support
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (3 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 04/62] tools/libelf: fix elf notes check for PVH guest Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 06/62] xen/common: Widen the guest logging buffer slightly Wei Liu
` (56 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Jonathan Ludlam <jonathan.ludlam@citrix.com>
Signed-off-by: Jonathan Ludlam <jonathan.ludlam@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
stubdom/grub/kexec.c | 7 +-
tools/helpers/init-xenstore-domain.c | 4 +-
tools/libxc/include/xc_dom.h | 48 ++++++-----
tools/libxc/xc_dom_compat_linux.c | 2 +-
tools/libxc/xc_dom_core.c | 152 +++++++++++++++++++++++------------
tools/libxc/xc_dom_x86.c | 65 ++++++++-------
tools/libxl/libxl_dom.c | 10 +--
7 files changed, 175 insertions(+), 113 deletions(-)
diff --git a/stubdom/grub/kexec.c b/stubdom/grub/kexec.c
index 437a0a96e9..61ca082d42 100644
--- a/stubdom/grub/kexec.c
+++ b/stubdom/grub/kexec.c
@@ -202,7 +202,7 @@ static void tpm_hash2pcr(struct xc_dom_image *dom, char *cmdline)
ASSERT(rv == 0 && resp->status == 0);
cmd.pcr = bswap_32(5); // PCR #5 for initrd
- sha1(dom->ramdisk_blob, dom->ramdisk_size, cmd.hash);
+ sha1(dom->modules[0].blob, dom->modules[0].size, cmd.hash);
rv = tpmfront_cmd(tpm, (void*)&cmd, sizeof(cmd), (void*)&resp, &resplen);
ASSERT(rv == 0 && resp->status == 0);
@@ -231,13 +231,12 @@ void kexec(void *kernel, long kernel_size, void *module, long module_size, char
/* We are using guest owned memory, therefore no limits. */
xc_dom_kernel_max_size(dom, 0);
- xc_dom_ramdisk_max_size(dom, 0);
+ xc_dom_module_max_size(dom, 0);
dom->kernel_blob = kernel;
dom->kernel_size = kernel_size;
- dom->ramdisk_blob = module;
- dom->ramdisk_size = module_size;
+ xc_dom_module_mem(dom, module, module_size, NULL);
dom->flags = flags;
dom->console_evtchn = start_info.console.domU.evtchn;
diff --git a/tools/helpers/init-xenstore-domain.c b/tools/helpers/init-xenstore-domain.c
index 047ad0cb1d..8453be283b 100644
--- a/tools/helpers/init-xenstore-domain.c
+++ b/tools/helpers/init-xenstore-domain.c
@@ -145,10 +145,10 @@ static int build(xc_interface *xch)
if ( ramdisk )
{
- rv = xc_dom_ramdisk_file(dom, ramdisk);
+ rv = xc_dom_module_file(dom, ramdisk, NULL);
if ( rv )
{
- fprintf(stderr, "xc_dom_ramdisk_file failed\n");
+ fprintf(stderr, "xc_dom_module_file failed\n");
goto err;
}
}
diff --git a/tools/libxc/include/xc_dom.h b/tools/libxc/include/xc_dom.h
index cdcdd07d2b..08be8a8f3f 100644
--- a/tools/libxc/include/xc_dom.h
+++ b/tools/libxc/include/xc_dom.h
@@ -22,6 +22,7 @@
#define INVALID_PFN ((xen_pfn_t)-1)
#define X86_HVM_NR_SPECIAL_PAGES 8
#define X86_HVM_END_SPECIAL_REGION 0xff000u
+#define XG_MAX_MODULES 2
/* --- typedefs and structs ---------------------------------------- */
@@ -56,17 +57,32 @@ struct xc_dom_phys {
xen_pfn_t count;
};
+struct xc_dom_module {
+ void *blob;
+ size_t size;
+ void *cmdline;
+ /* If seg.vstart is non zero then the module will be loaded at that
+ * address, otherwise it will automatically placed.
+ *
+ * If automatic placement is used and the module is gzip
+ * compressed then it will be decompressed as it is loaded. If the
+ * module has been explicitly placed then it is loaded as is
+ * otherwise decompressing risks undoing the manual placement.
+ */
+ struct xc_dom_seg seg;
+};
+
struct xc_dom_image {
/* files */
void *kernel_blob;
size_t kernel_size;
- void *ramdisk_blob;
- size_t ramdisk_size;
+ unsigned int num_modules;
+ struct xc_dom_module modules[XG_MAX_MODULES];
void *devicetree_blob;
size_t devicetree_size;
size_t max_kernel_size;
- size_t max_ramdisk_size;
+ size_t max_module_size;
size_t max_devicetree_size;
/* arguments and parameters */
@@ -80,15 +96,6 @@ struct xc_dom_image {
/* memory layout */
struct xc_dom_seg kernel_seg;
- /* If ramdisk_seg.vstart is non zero then the ramdisk will be
- * loaded at that address, otherwise it will automatically placed.
- *
- * If automatic placement is used and the ramdisk is gzip
- * compressed then it will be decompressed as it is loaded. If the
- * ramdisk has been explicitly placed then it is loaded as is
- * otherwise decompressing risks undoing the manual placement.
- */
- struct xc_dom_seg ramdisk_seg;
struct xc_dom_seg p2m_seg;
struct xc_dom_seg pgtables_seg;
struct xc_dom_seg devicetree_seg;
@@ -277,12 +284,12 @@ void xc_dom_release(struct xc_dom_image *dom);
int xc_dom_rambase_init(struct xc_dom_image *dom, uint64_t rambase);
int xc_dom_mem_init(struct xc_dom_image *dom, unsigned int mem_mb);
-/* Set this larger if you have enormous ramdisks/kernels. Note that
+/* Set this larger if you have enormous modules/kernels. Note that
* you should trust all kernels not to be maliciously large (e.g. to
* exhaust all dom0 memory) if you do this (see CVE-2012-4544 /
* XSA-25). You can also set the default independently for
- * ramdisks/kernels in xc_dom_allocate() or call
- * xc_dom_{kernel,ramdisk}_max_size.
+ * modules/kernels in xc_dom_allocate() or call
+ * xc_dom_{kernel,module}_max_size.
*/
#ifndef XC_DOM_DECOMPRESS_MAX
#define XC_DOM_DECOMPRESS_MAX (1024*1024*1024) /* 1GB */
@@ -291,8 +298,8 @@ int xc_dom_mem_init(struct xc_dom_image *dom, unsigned int mem_mb);
int xc_dom_kernel_check_size(struct xc_dom_image *dom, size_t sz);
int xc_dom_kernel_max_size(struct xc_dom_image *dom, size_t sz);
-int xc_dom_ramdisk_check_size(struct xc_dom_image *dom, size_t sz);
-int xc_dom_ramdisk_max_size(struct xc_dom_image *dom, size_t sz);
+int xc_dom_module_check_size(struct xc_dom_image *dom, size_t sz);
+int xc_dom_module_max_size(struct xc_dom_image *dom, size_t sz);
int xc_dom_devicetree_max_size(struct xc_dom_image *dom, size_t sz);
@@ -303,11 +310,12 @@ int xc_dom_do_gunzip(xc_interface *xch,
int xc_dom_try_gunzip(struct xc_dom_image *dom, void **blob, size_t * size);
int xc_dom_kernel_file(struct xc_dom_image *dom, const char *filename);
-int xc_dom_ramdisk_file(struct xc_dom_image *dom, const char *filename);
+int xc_dom_module_file(struct xc_dom_image *dom, const char *filename,
+ const char *cmdline);
int xc_dom_kernel_mem(struct xc_dom_image *dom, const void *mem,
size_t memsize);
-int xc_dom_ramdisk_mem(struct xc_dom_image *dom, const void *mem,
- size_t memsize);
+int xc_dom_module_mem(struct xc_dom_image *dom, const void *mem,
+ size_t memsize, const char *cmdline);
int xc_dom_devicetree_file(struct xc_dom_image *dom, const char *filename);
int xc_dom_devicetree_mem(struct xc_dom_image *dom, const void *mem,
size_t memsize);
diff --git a/tools/libxc/xc_dom_compat_linux.c b/tools/libxc/xc_dom_compat_linux.c
index c922c61e90..b3d43feed9 100644
--- a/tools/libxc/xc_dom_compat_linux.c
+++ b/tools/libxc/xc_dom_compat_linux.c
@@ -56,7 +56,7 @@ int xc_linux_build(xc_interface *xch, uint32_t domid,
if ( (rc = xc_dom_kernel_file(dom, image_name)) != 0 )
goto out;
if ( initrd_name && strlen(initrd_name) &&
- ((rc = xc_dom_ramdisk_file(dom, initrd_name)) != 0) )
+ ((rc = xc_dom_module_file(dom, initrd_name, NULL)) != 0) )
goto out;
dom->flags |= flags;
diff --git a/tools/libxc/xc_dom_core.c b/tools/libxc/xc_dom_core.c
index 303cb971e8..3e65aff22b 100644
--- a/tools/libxc/xc_dom_core.c
+++ b/tools/libxc/xc_dom_core.c
@@ -314,16 +314,16 @@ int xc_dom_kernel_check_size(struct xc_dom_image *dom, size_t sz)
return 0;
}
-int xc_dom_ramdisk_check_size(struct xc_dom_image *dom, size_t sz)
+int xc_dom_module_check_size(struct xc_dom_image *dom, size_t sz)
{
/* No limit */
- if ( !dom->max_ramdisk_size )
+ if ( !dom->max_module_size )
return 0;
- if ( sz > dom->max_ramdisk_size )
+ if ( sz > dom->max_module_size )
{
xc_dom_panic(dom->xch, XC_INVALID_KERNEL,
- "ramdisk image too large");
+ "module image too large");
return 1;
}
@@ -764,7 +764,7 @@ struct xc_dom_image *xc_dom_allocate(xc_interface *xch,
dom->xch = xch;
dom->max_kernel_size = XC_DOM_DECOMPRESS_MAX;
- dom->max_ramdisk_size = XC_DOM_DECOMPRESS_MAX;
+ dom->max_module_size = XC_DOM_DECOMPRESS_MAX;
dom->max_devicetree_size = XC_DOM_DECOMPRESS_MAX;
if ( cmdline )
@@ -797,10 +797,10 @@ int xc_dom_kernel_max_size(struct xc_dom_image *dom, size_t sz)
return 0;
}
-int xc_dom_ramdisk_max_size(struct xc_dom_image *dom, size_t sz)
+int xc_dom_module_max_size(struct xc_dom_image *dom, size_t sz)
{
- DOMPRINTF("%s: ramdisk_max_size=%zx", __FUNCTION__, sz);
- dom->max_ramdisk_size = sz;
+ DOMPRINTF("%s: module_max_size=%zx", __FUNCTION__, sz);
+ dom->max_module_size = sz;
return 0;
}
@@ -821,16 +821,30 @@ int xc_dom_kernel_file(struct xc_dom_image *dom, const char *filename)
return xc_dom_try_gunzip(dom, &dom->kernel_blob, &dom->kernel_size);
}
-int xc_dom_ramdisk_file(struct xc_dom_image *dom, const char *filename)
+int xc_dom_module_file(struct xc_dom_image *dom, const char *filename, const char *cmdline)
{
+ unsigned int mod = dom->num_modules++;
+
DOMPRINTF("%s: filename=\"%s\"", __FUNCTION__, filename);
- dom->ramdisk_blob =
- xc_dom_malloc_filemap(dom, filename, &dom->ramdisk_size,
- dom->max_ramdisk_size);
+ dom->modules[mod].blob =
+ xc_dom_malloc_filemap(dom, filename, &dom->modules[mod].size,
+ dom->max_module_size);
- if ( dom->ramdisk_blob == NULL )
+ if ( dom->modules[mod].blob == NULL )
return -1;
-// return xc_dom_try_gunzip(dom, &dom->ramdisk_blob, &dom->ramdisk_size);
+
+ if ( cmdline )
+ {
+ dom->modules[mod].cmdline = xc_dom_strdup(dom, cmdline);
+
+ if ( dom->modules[mod].cmdline == NULL )
+ return -1;
+ }
+ else
+ {
+ dom->modules[mod].cmdline = NULL;
+ }
+
return 0;
}
@@ -859,13 +873,28 @@ int xc_dom_kernel_mem(struct xc_dom_image *dom, const void *mem, size_t memsize)
return xc_dom_try_gunzip(dom, &dom->kernel_blob, &dom->kernel_size);
}
-int xc_dom_ramdisk_mem(struct xc_dom_image *dom, const void *mem,
- size_t memsize)
+int xc_dom_module_mem(struct xc_dom_image *dom, const void *mem,
+ size_t memsize, const char *cmdline)
{
+ unsigned int mod = dom->num_modules++;
+
DOMPRINTF_CALLED(dom->xch);
- dom->ramdisk_blob = (void *)mem;
- dom->ramdisk_size = memsize;
-// return xc_dom_try_gunzip(dom, &dom->ramdisk_blob, &dom->ramdisk_size);
+
+ dom->modules[mod].blob = (void *)mem;
+ dom->modules[mod].size = memsize;
+
+ if ( cmdline )
+ {
+ dom->modules[mod].cmdline = xc_dom_strdup(dom, cmdline);
+
+ if ( dom->modules[mod].cmdline == NULL )
+ return -1;
+ }
+ else
+ {
+ dom->modules[mod].cmdline = NULL;
+ }
+
return 0;
}
@@ -990,41 +1019,42 @@ int xc_dom_update_guest_p2m(struct xc_dom_image *dom)
return 0;
}
-static int xc_dom_build_ramdisk(struct xc_dom_image *dom)
+static int xc_dom_build_module(struct xc_dom_image *dom, unsigned int mod)
{
- size_t unziplen, ramdisklen;
- void *ramdiskmap;
+ size_t unziplen, modulelen;
+ void *modulemap;
+ char name[10];
- if ( !dom->ramdisk_seg.vstart )
+ if ( !dom->modules[mod].seg.vstart )
{
unziplen = xc_dom_check_gzip(dom->xch,
- dom->ramdisk_blob, dom->ramdisk_size);
- if ( xc_dom_ramdisk_check_size(dom, unziplen) != 0 )
+ dom->modules[mod].blob, dom->modules[mod].size);
+ if ( xc_dom_module_check_size(dom, unziplen) != 0 )
unziplen = 0;
}
else
unziplen = 0;
- ramdisklen = unziplen ? unziplen : dom->ramdisk_size;
-
- if ( xc_dom_alloc_segment(dom, &dom->ramdisk_seg, "ramdisk",
- dom->ramdisk_seg.vstart, ramdisklen) != 0 )
+ modulelen = unziplen ? unziplen : dom->modules[mod].size;
+ snprintf(name, sizeof(name), "module%u", mod);
+ if ( xc_dom_alloc_segment(dom, &dom->modules[mod].seg, name,
+ dom->modules[mod].seg.vstart, modulelen) != 0 )
goto err;
- ramdiskmap = xc_dom_seg_to_ptr(dom, &dom->ramdisk_seg);
- if ( ramdiskmap == NULL )
+ modulemap = xc_dom_seg_to_ptr(dom, &dom->modules[mod].seg);
+ if ( modulemap == NULL )
{
- DOMPRINTF("%s: xc_dom_seg_to_ptr(dom, &dom->ramdisk_seg) => NULL",
- __FUNCTION__);
+ DOMPRINTF("%s: xc_dom_seg_to_ptr(dom, &dom->modules[%u].seg) => NULL",
+ __FUNCTION__, mod);
goto err;
}
if ( unziplen )
{
- if ( xc_dom_do_gunzip(dom->xch, dom->ramdisk_blob, dom->ramdisk_size,
- ramdiskmap, ramdisklen) == -1 )
+ if ( xc_dom_do_gunzip(dom->xch, dom->modules[mod].blob, dom->modules[mod].size,
+ modulemap, modulelen) == -1 )
goto err;
}
else
- memcpy(ramdiskmap, dom->ramdisk_blob, dom->ramdisk_size);
+ memcpy(modulemap, dom->modules[mod].blob, dom->modules[mod].size);
return 0;
@@ -1131,6 +1161,7 @@ int xc_dom_build_image(struct xc_dom_image *dom)
{
unsigned int page_size;
bool unmapped_initrd;
+ unsigned int mod;
DOMPRINTF_CALLED(dom->xch);
@@ -1154,15 +1185,24 @@ int xc_dom_build_image(struct xc_dom_image *dom)
if ( dom->kernel_loader->loader(dom) != 0 )
goto err;
- /* Don't load ramdisk now if no initial mapping required. */
- unmapped_initrd = dom->parms.unmapped_initrd && !dom->ramdisk_seg.vstart;
-
- if ( dom->ramdisk_blob && !unmapped_initrd )
+ /* Don't load ramdisk / other modules now if no initial mapping required. */
+ for ( mod = 0; mod < dom->num_modules; mod++ )
{
- if ( xc_dom_build_ramdisk(dom) != 0 )
- goto err;
- dom->initrd_start = dom->ramdisk_seg.vstart;
- dom->initrd_len = dom->ramdisk_seg.vend - dom->ramdisk_seg.vstart;
+ unmapped_initrd = (dom->parms.unmapped_initrd &&
+ !dom->modules[mod].seg.vstart);
+
+ if ( dom->modules[mod].blob && !unmapped_initrd )
+ {
+ if ( xc_dom_build_module(dom, mod) != 0 )
+ goto err;
+
+ if ( mod == 0 )
+ {
+ dom->initrd_start = dom->modules[mod].seg.vstart;
+ dom->initrd_len =
+ dom->modules[mod].seg.vend - dom->modules[mod].seg.vstart;
+ }
+ }
}
/* load devicetree */
@@ -1216,14 +1256,24 @@ int xc_dom_build_image(struct xc_dom_image *dom)
if ( dom->virt_pgtab_end && xc_dom_alloc_pad(dom, dom->virt_pgtab_end) )
return -1;
- /* Load ramdisk if no initial mapping required. */
- if ( dom->ramdisk_blob && unmapped_initrd )
+ for ( mod = 0; mod < dom->num_modules; mod++ )
{
- if ( xc_dom_build_ramdisk(dom) != 0 )
- goto err;
- dom->flags |= SIF_MOD_START_PFN;
- dom->initrd_start = dom->ramdisk_seg.pfn;
- dom->initrd_len = page_size * dom->ramdisk_seg.pages;
+ unmapped_initrd = (dom->parms.unmapped_initrd &&
+ !dom->modules[mod].seg.vstart);
+
+ /* Load ramdisk / other modules if no initial mapping required. */
+ if ( dom->modules[mod].blob && unmapped_initrd )
+ {
+ if ( xc_dom_build_module(dom, mod) != 0 )
+ goto err;
+
+ if ( mod == 0 )
+ {
+ dom->flags |= SIF_MOD_START_PFN;
+ dom->initrd_start = dom->modules[mod].seg.pfn;
+ dom->initrd_len = page_size * dom->modules[mod].seg.pages;
+ }
+ }
}
/* Allocate p2m list if outside of initial kernel mapping. */
diff --git a/tools/libxc/xc_dom_x86.c b/tools/libxc/xc_dom_x86.c
index bff68a011f..0b65dab4bc 100644
--- a/tools/libxc/xc_dom_x86.c
+++ b/tools/libxc/xc_dom_x86.c
@@ -70,8 +70,8 @@
#define round_up(addr, mask) ((addr) | (mask))
#define round_pg_up(addr) (((addr) + PAGE_SIZE_X86 - 1) & ~(PAGE_SIZE_X86 - 1))
-#define HVMLOADER_MODULE_MAX_COUNT 1
-#define HVMLOADER_MODULE_NAME_SIZE 10
+#define HVMLOADER_MODULE_MAX_COUNT 2
+#define HVMLOADER_MODULE_CMDLINE_SIZE MAX_GUEST_CMDLINE
struct xc_dom_params {
unsigned levels;
@@ -627,6 +627,12 @@ static int alloc_magic_pages_hvm(struct xc_dom_image *dom)
xc_hvm_param_set(xch, domid, HVM_PARAM_SHARING_RING_PFN,
special_pfn(SPECIALPAGE_SHARING));
+ start_info_size +=
+ sizeof(struct hvm_modlist_entry) * HVMLOADER_MODULE_MAX_COUNT;
+
+ start_info_size +=
+ HVMLOADER_MODULE_CMDLINE_SIZE * HVMLOADER_MODULE_MAX_COUNT;
+
if ( !dom->device_model )
{
if ( dom->cmdline )
@@ -634,22 +640,9 @@ static int alloc_magic_pages_hvm(struct xc_dom_image *dom)
dom->cmdline_size = ROUNDUP(strlen(dom->cmdline) + 1, 8);
start_info_size += dom->cmdline_size;
}
-
- /* Limited to one module. */
- if ( dom->ramdisk_blob )
- start_info_size += sizeof(struct hvm_modlist_entry);
}
else
{
- start_info_size +=
- sizeof(struct hvm_modlist_entry) * HVMLOADER_MODULE_MAX_COUNT;
- /*
- * Add extra space to write modules name.
- * The HVMLOADER_MODULE_NAME_SIZE accounts for NUL byte.
- */
- start_info_size +=
- HVMLOADER_MODULE_NAME_SIZE * HVMLOADER_MODULE_MAX_COUNT;
-
/*
* Allocate and clear additional ioreq server pages. The default
* server will use the IOREQ and BUFIOREQ special pages above.
@@ -749,7 +742,7 @@ static int start_info_x86_32(struct xc_dom_image *dom)
start_info->console.domU.mfn = xc_dom_p2m(dom, dom->console_pfn);
start_info->console.domU.evtchn = dom->console_evtchn;
- if ( dom->ramdisk_blob )
+ if ( dom->modules[0].blob )
{
start_info->mod_start = dom->initrd_start;
start_info->mod_len = dom->initrd_len;
@@ -800,7 +793,7 @@ static int start_info_x86_64(struct xc_dom_image *dom)
start_info->console.domU.mfn = xc_dom_p2m(dom, dom->console_pfn);
start_info->console.domU.evtchn = dom->console_evtchn;
- if ( dom->ramdisk_blob )
+ if ( dom->modules[0].blob )
{
start_info->mod_start = dom->initrd_start;
start_info->mod_len = dom->initrd_len;
@@ -1237,7 +1230,7 @@ static int meminit_hvm(struct xc_dom_image *dom)
unsigned long target_pages = dom->target_pages;
unsigned long cur_pages, cur_pfn;
int rc;
- unsigned long stat_normal_pages = 0, stat_2mb_pages = 0,
+ unsigned long stat_normal_pages = 0, stat_2mb_pages = 0,
stat_1gb_pages = 0;
unsigned int memflags = 0;
int claim_enabled = dom->claim_enabled;
@@ -1303,6 +1296,8 @@ static int meminit_hvm(struct xc_dom_image *dom)
p2m_size = 0;
for ( i = 0; i < nr_vmemranges; i++ )
{
+ DOMPRINTF("range: start=0x%"PRIx64" end=0x%"PRIx64, vmemranges[i].start, vmemranges[i].end);
+
total_pages += ((vmemranges[i].end - vmemranges[i].start)
>> PAGE_SHIFT);
p2m_size = p2m_size > (vmemranges[i].end >> PAGE_SHIFT) ?
@@ -1633,7 +1628,7 @@ static int alloc_pgtables_hvm(struct xc_dom_image *dom)
*/
static void add_module_to_list(struct xc_dom_image *dom,
struct xc_hvm_firmware_module *module,
- const char *name,
+ const char *cmdline,
struct hvm_modlist_entry *modlist,
struct hvm_start_info *start_info)
{
@@ -1648,16 +1643,20 @@ static void add_module_to_list(struct xc_dom_image *dom,
return;
assert(start_info->nr_modules < HVMLOADER_MODULE_MAX_COUNT);
- assert(strnlen(name, HVMLOADER_MODULE_NAME_SIZE)
- < HVMLOADER_MODULE_NAME_SIZE);
modlist[index].paddr = module->guest_addr_out;
modlist[index].size = module->length;
- strncpy(modules_cmdline_start + HVMLOADER_MODULE_NAME_SIZE * index,
- name, HVMLOADER_MODULE_NAME_SIZE);
+ if ( cmdline )
+ {
+ assert(strnlen(cmdline, HVMLOADER_MODULE_CMDLINE_SIZE)
+ < HVMLOADER_MODULE_CMDLINE_SIZE);
+ strncpy(modules_cmdline_start + HVMLOADER_MODULE_CMDLINE_SIZE * index,
+ cmdline, HVMLOADER_MODULE_CMDLINE_SIZE);
+ }
+
modlist[index].cmdline_paddr =
- modules_cmdline_paddr + HVMLOADER_MODULE_NAME_SIZE * index;
+ modules_cmdline_paddr + HVMLOADER_MODULE_CMDLINE_SIZE * index;
start_info->nr_modules++;
}
@@ -1669,10 +1668,10 @@ static int bootlate_hvm(struct xc_dom_image *dom)
struct hvm_start_info *start_info;
size_t start_info_size;
struct hvm_modlist_entry *modlist;
+ unsigned int i;
start_info_size = sizeof(*start_info) + dom->cmdline_size;
- if ( dom->ramdisk_blob )
- start_info_size += sizeof(struct hvm_modlist_entry);
+ start_info_size += sizeof(struct hvm_modlist_entry) * dom->num_modules;
if ( start_info_size >
dom->start_info_seg.pages << XC_DOM_PAGE_SHIFT(dom) )
@@ -1703,12 +1702,18 @@ static int bootlate_hvm(struct xc_dom_image *dom)
((uintptr_t)cmdline - (uintptr_t)start_info);
}
- if ( dom->ramdisk_blob )
+ for ( i = 0; i < dom->num_modules; i++ )
{
+ struct xc_hvm_firmware_module mod;
+
+ DOMPRINTF("Adding module %u", i);
+ mod.guest_addr_out =
+ dom->modules[i].seg.vstart - dom->parms.virt_base;
+ mod.length =
+ dom->modules[i].seg.vend - dom->modules[i].seg.vstart;
- modlist[0].paddr = dom->ramdisk_seg.vstart - dom->parms.virt_base;
- modlist[0].size = dom->ramdisk_seg.vend - dom->ramdisk_seg.vstart;
- start_info->nr_modules = 1;
+ add_module_to_list(dom, &mod, dom->modules[i].cmdline,
+ modlist, start_info);
}
/* ACPI module 0 is the RSDP */
diff --git a/tools/libxl/libxl_dom.c b/tools/libxl/libxl_dom.c
index ef834e652d..fbbdb9ec2f 100644
--- a/tools/libxl/libxl_dom.c
+++ b/tools/libxl/libxl_dom.c
@@ -796,12 +796,12 @@ int libxl__build_pv(libxl__gc *gc, uint32_t domid,
if ( state->pv_ramdisk.path && strlen(state->pv_ramdisk.path) ) {
if (state->pv_ramdisk.mapped) {
- if ( (ret = xc_dom_ramdisk_mem(dom, state->pv_ramdisk.data, state->pv_ramdisk.size)) != 0 ) {
+ if ( (ret = xc_dom_module_mem(dom, state->pv_ramdisk.data, state->pv_ramdisk.size, NULL)) != 0 ) {
LOGE(ERROR, "xc_dom_ramdisk_mem failed");
goto out;
}
} else {
- if ( (ret = xc_dom_ramdisk_file(dom, state->pv_ramdisk.path)) != 0 ) {
+ if ( (ret = xc_dom_module_file(dom, state->pv_ramdisk.path, NULL)) != 0 ) {
LOGE(ERROR, "xc_dom_ramdisk_file failed");
goto out;
}
@@ -1043,14 +1043,14 @@ static int libxl__domain_firmware(libxl__gc *gc,
if (state->pv_ramdisk.path && strlen(state->pv_ramdisk.path)) {
if (state->pv_ramdisk.mapped) {
- rc = xc_dom_ramdisk_mem(dom, state->pv_ramdisk.data,
- state->pv_ramdisk.size);
+ rc = xc_dom_module_mem(dom, state->pv_ramdisk.data,
+ state->pv_ramdisk.size, NULL);
if (rc) {
LOGE(ERROR, "xc_dom_ramdisk_mem failed");
goto out;
}
} else {
- rc = xc_dom_ramdisk_file(dom, state->pv_ramdisk.path);
+ rc = xc_dom_module_file(dom, state->pv_ramdisk.path, NULL);
if (rc) {
LOGE(ERROR, "xc_dom_ramdisk_file failed");
goto out;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 06/62] xen/common: Widen the guest logging buffer slightly
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (4 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 05/62] tools/libxc: Multi modules support Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 07/62] x86/time: Print a more helpful error when a platform timer can't be found Wei Liu
` (55 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
This reduces the amount of line wrapping from guests; Xen in particular likes
to print lines longer than 80 characters.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Wei Liu <wei.liu2@citrix.com>
---
xen/include/xen/sched.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
index 002ba29d6d..64abc1df6c 100644
--- a/xen/include/xen/sched.h
+++ b/xen/include/xen/sched.h
@@ -427,7 +427,7 @@ struct domain
xen_domain_handle_t handle;
/* hvm_print_line() and guest_console_write() logging. */
-#define DOMAIN_PBUF_SIZE 80
+#define DOMAIN_PBUF_SIZE 200
char *pbuf;
unsigned pbuf_idx;
spinlock_t pbuf_lock;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 07/62] x86/time: Print a more helpful error when a platform timer can't be found
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (5 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 06/62] xen/common: Widen the guest logging buffer slightly Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 08/62] x86/link: Introduce and use SECTION_ALIGN Wei Liu
` (54 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Wei Liu <wei.liu2@citrix.com>
---
xen/arch/x86/time.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
index 2a879502a2..3b654d7b7d 100644
--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -708,7 +708,8 @@ static u64 __init init_platform_timer(void)
}
}
- BUG_ON(rc <= 0);
+ if ( rc <= 0 )
+ panic("Unable to find usable platform timer");
printk("Platform timer is %s %s\n",
freq_string(pts->frequency), pts->name);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 08/62] x86/link: Introduce and use SECTION_ALIGN
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (6 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 07/62] x86/time: Print a more helpful error when a platform timer can't be found Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 09/62] ACPICA: Make ACPI Power Management Timer (PM Timer) optional Wei Liu
` (53 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
... to reduce the quantity of #ifdef EFI.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Wei Liu <wei.liu2@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
---
xen/arch/x86/xen.lds.S | 50 +++++++++++++-------------------------------------
1 file changed, 13 insertions(+), 37 deletions(-)
diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S
index d5e8821d41..6164ad094f 100644
--- a/xen/arch/x86/xen.lds.S
+++ b/xen/arch/x86/xen.lds.S
@@ -12,12 +12,14 @@
#define FORMAT "pei-x86-64"
#undef __XEN_VIRT_START
#define __XEN_VIRT_START __image_base__
+#define SECTION_ALIGN MB(2)
ENTRY(efi_start)
#else /* !EFI */
#define FORMAT "elf64-x86-64"
+#define SECTION_ALIGN PAGE_SIZE
ENTRY(start)
@@ -67,11 +69,7 @@ SECTIONS
_etext = .; /* End of text section */
} :text = 0x9090
-#ifdef EFI
- . = ALIGN(MB(2));
-#else
- . = ALIGN(PAGE_SIZE);
-#endif
+ . = ALIGN(SECTION_ALIGN);
__2M_text_end = .;
__2M_rodata_start = .; /* Start of 2M superpages, mapped RO. */
@@ -149,11 +147,7 @@ SECTIONS
#endif
_erodata = .;
-#ifdef EFI
- . = ALIGN(MB(2));
-#else
- . = ALIGN(PAGE_SIZE);
-#endif
+ . = ALIGN(SECTION_ALIGN);
__2M_rodata_end = .;
__2M_init_start = .; /* Start of 2M superpages, mapped RWX (boot only). */
@@ -215,11 +209,7 @@ SECTIONS
__ctors_end = .;
} :text
-#ifdef EFI
- . = ALIGN(MB(2));
-#else
- . = ALIGN(PAGE_SIZE);
-#endif
+ . = ALIGN(SECTION_ALIGN);
__init_end = .;
__2M_init_end = .;
@@ -257,11 +247,7 @@ SECTIONS
} :text
_end = . ;
-#ifdef EFI
- . = ALIGN(MB(2));
-#else
- . = ALIGN(PAGE_SIZE);
-#endif
+ . = ALIGN(SECTION_ALIGN);
__2M_rwdata_end = .;
#ifdef EFI
@@ -310,23 +296,13 @@ ASSERT(__image_base__ > XEN_VIRT_START ||
ASSERT(kexec_reloc_size - kexec_reloc <= PAGE_SIZE, "kexec_reloc is too large")
#endif
-#ifdef EFI
-ASSERT(IS_ALIGNED(__2M_text_end, MB(2)), "__2M_text_end misaligned")
-ASSERT(IS_ALIGNED(__2M_rodata_start, MB(2)), "__2M_rodata_start misaligned")
-ASSERT(IS_ALIGNED(__2M_rodata_end, MB(2)), "__2M_rodata_end misaligned")
-ASSERT(IS_ALIGNED(__2M_init_start, MB(2)), "__2M_init_start misaligned")
-ASSERT(IS_ALIGNED(__2M_init_end, MB(2)), "__2M_init_end misaligned")
-ASSERT(IS_ALIGNED(__2M_rwdata_start, MB(2)), "__2M_rwdata_start misaligned")
-ASSERT(IS_ALIGNED(__2M_rwdata_end, MB(2)), "__2M_rwdata_end misaligned")
-#else
-ASSERT(IS_ALIGNED(__2M_text_end, PAGE_SIZE), "__2M_text_end misaligned")
-ASSERT(IS_ALIGNED(__2M_rodata_start, PAGE_SIZE), "__2M_rodata_start misaligned")
-ASSERT(IS_ALIGNED(__2M_rodata_end, PAGE_SIZE), "__2M_rodata_end misaligned")
-ASSERT(IS_ALIGNED(__2M_init_start, PAGE_SIZE), "__2M_init_start misaligned")
-ASSERT(IS_ALIGNED(__2M_init_end, PAGE_SIZE), "__2M_init_end misaligned")
-ASSERT(IS_ALIGNED(__2M_rwdata_start, PAGE_SIZE), "__2M_rwdata_start misaligned")
-ASSERT(IS_ALIGNED(__2M_rwdata_end, PAGE_SIZE), "__2M_rwdata_end misaligned")
-#endif
+ASSERT(IS_ALIGNED(__2M_text_end, SECTION_ALIGN), "__2M_text_end misaligned")
+ASSERT(IS_ALIGNED(__2M_rodata_start, SECTION_ALIGN), "__2M_rodata_start misaligned")
+ASSERT(IS_ALIGNED(__2M_rodata_end, SECTION_ALIGN), "__2M_rodata_end misaligned")
+ASSERT(IS_ALIGNED(__2M_init_start, SECTION_ALIGN), "__2M_init_start misaligned")
+ASSERT(IS_ALIGNED(__2M_init_end, SECTION_ALIGN), "__2M_init_end misaligned")
+ASSERT(IS_ALIGNED(__2M_rwdata_start, SECTION_ALIGN), "__2M_rwdata_start misaligned")
+ASSERT(IS_ALIGNED(__2M_rwdata_end, SECTION_ALIGN), "__2M_rwdata_end misaligned")
ASSERT(IS_ALIGNED(cpu0_stack, STACK_SIZE), "cpu0_stack misaligned")
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 09/62] ACPICA: Make ACPI Power Management Timer (PM Timer) optional.
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (7 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 08/62] x86/link: Introduce and use SECTION_ALIGN Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 10/62] xen/domctl: Return arch_config via getdomaininfo Wei Liu
` (52 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Bob Moore <robert.moore@intel.com>
PM Timer is now optional.
This support is already in Windows8 and "SHOULD" come out in ACPI 5.0A
(if all goes well).
The change doesn't affect Xen directly, because it does not rely
on the presence of the PM timer.
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
[ported to Xen]
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
xen/drivers/acpi/tables/tbfadt.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/xen/drivers/acpi/tables/tbfadt.c b/xen/drivers/acpi/tables/tbfadt.c
index d62d8d5cb9..f11fd5a900 100644
--- a/xen/drivers/acpi/tables/tbfadt.c
+++ b/xen/drivers/acpi/tables/tbfadt.c
@@ -95,7 +95,8 @@ static struct acpi_fadt_info __initdata fadt_info_table[] = {
{"PmTimerBlock", ACPI_FADT_OFFSET(xpm_timer_block),
ACPI_FADT_OFFSET(pm_timer_block),
- ACPI_FADT_OFFSET(pm_timer_length), ACPI_FADT_REQUIRED},
+ ACPI_FADT_OFFSET(pm_timer_length),
+ ACPI_FADT_SEPARATE_LENGTH}, /* ACPI 5.0A: Timer is optional */
{"Gpe0Block", ACPI_FADT_OFFSET(xgpe0_block),
ACPI_FADT_OFFSET(gpe0_block),
@@ -437,7 +438,7 @@ static void __init acpi_tb_validate_fadt(void)
if (fadt_info_table[i].type & ACPI_FADT_REQUIRED) {
/*
- * Field is required (Pm1a_event, Pm1a_control, pm_timer).
+ * Field is required (Pm1a_event, Pm1a_control).
* Both the address and length must be non-zero.
*/
if (!address64->address || !length) {
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 10/62] xen/domctl: Return arch_config via getdomaininfo
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (8 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 09/62] ACPICA: Make ACPI Power Management Timer (PM Timer) optional Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 11/62] tools/ocaml: Expose arch_config in domaininfo Wei Liu
` (51 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
This allows toolstack software to distinguish HVM from PVH guests.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
v2: bump domctl version number
---
tools/libxc/include/xenctrl.h | 1 +
tools/libxc/xc_domain.c | 1 +
xen/arch/x86/domctl.c | 2 ++
xen/include/public/domctl.h | 3 ++-
4 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/tools/libxc/include/xenctrl.h b/tools/libxc/include/xenctrl.h
index 09e1363be2..7d7a7c5999 100644
--- a/tools/libxc/include/xenctrl.h
+++ b/tools/libxc/include/xenctrl.h
@@ -456,6 +456,7 @@ typedef struct xc_dominfo {
unsigned int max_vcpu_id;
xen_domain_handle_t handle;
unsigned int cpupool;
+ struct xen_arch_domainconfig arch_config;
} xc_dominfo_t;
typedef xen_domctl_getdomaininfo_t xc_domaininfo_t;
diff --git a/tools/libxc/xc_domain.c b/tools/libxc/xc_domain.c
index 3ccd27f101..8169284dc1 100644
--- a/tools/libxc/xc_domain.c
+++ b/tools/libxc/xc_domain.c
@@ -421,6 +421,7 @@ int xc_domain_getinfo(xc_interface *xch,
info->nr_online_vcpus = domctl.u.getdomaininfo.nr_online_vcpus;
info->max_vcpu_id = domctl.u.getdomaininfo.max_vcpu_id;
info->cpupool = domctl.u.getdomaininfo.cpupool;
+ info->arch_config = domctl.u.getdomaininfo.arch_config;
memcpy(info->handle, domctl.u.getdomaininfo.handle,
sizeof(xen_domain_handle_t));
diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c
index 5973d9fdaa..2585d4e14b 100644
--- a/xen/arch/x86/domctl.c
+++ b/xen/arch/x86/domctl.c
@@ -348,6 +348,8 @@ void arch_get_domain_info(const struct domain *d,
{
if ( paging_mode_hap(d) )
info->flags |= XEN_DOMINF_hap;
+
+ info->arch_config.emulation_flags = d->arch.emulation_flags;
}
#define MAX_IOPORTS 0x10000
diff --git a/xen/include/public/domctl.h b/xen/include/public/domctl.h
index 9ae7295919..a8921dd7ed 100644
--- a/xen/include/public/domctl.h
+++ b/xen/include/public/domctl.h
@@ -38,7 +38,7 @@
#include "hvm/save.h"
#include "memory.h"
-#define XEN_DOMCTL_INTERFACE_VERSION 0x0000000e
+#define XEN_DOMCTL_INTERFACE_VERSION 0x0000000f
/*
* NB. xen_domctl.domain is an IN/OUT parameter for this operation.
@@ -116,6 +116,7 @@ struct xen_domctl_getdomaininfo {
uint32_t ssidref;
xen_domain_handle_t handle;
uint32_t cpupool;
+ struct xen_arch_domainconfig arch_config;
};
typedef struct xen_domctl_getdomaininfo xen_domctl_getdomaininfo_t;
DEFINE_XEN_GUEST_HANDLE(xen_domctl_getdomaininfo_t);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 11/62] tools/ocaml: Expose arch_config in domaininfo
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (9 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 10/62] xen/domctl: Return arch_config via getdomaininfo Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 12/62] tools/ocaml: Extend domain_create() to take arch_domainconfig Wei Liu
` (50 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
tools/ocaml/libs/xc/xenctrl.ml | 29 +++++++++++++++++++++++++++++
tools/ocaml/libs/xc/xenctrl.mli | 28 ++++++++++++++++++++++++++++
tools/ocaml/libs/xc/xenctrl_stubs.c | 26 ++++++++++++++++++++++++--
3 files changed, 81 insertions(+), 2 deletions(-)
diff --git a/tools/ocaml/libs/xc/xenctrl.ml b/tools/ocaml/libs/xc/xenctrl.ml
index 70a325b0e9..d549068d60 100644
--- a/tools/ocaml/libs/xc/xenctrl.ml
+++ b/tools/ocaml/libs/xc/xenctrl.ml
@@ -28,6 +28,34 @@ type vcpuinfo =
cpumap: int32;
}
+type xen_arm_arch_domainconfig =
+{
+ gic_version: int;
+ nr_spis: int;
+ clock_frequency: int32;
+}
+
+type x86_arch_emulation_flags =
+ | X86_EMU_LAPIC
+ | X86_EMU_HPET
+ | X86_EMU_PM
+ | X86_EMU_RTC
+ | X86_EMU_IOAPIC
+ | X86_EMU_PIC
+ | X86_EMU_VGA
+ | X86_EMU_IOMMU
+ | X86_EMU_PIT
+ | X86_EMU_USE_PIRQ
+
+type xen_x86_arch_domainconfig =
+{
+ emulation_flags: x86_arch_emulation_flags list;
+}
+
+type arch_domainconfig =
+ | ARM of xen_arm_arch_domainconfig
+ | X86 of xen_x86_arch_domainconfig
+
type domaininfo =
{
domid : domid;
@@ -46,6 +74,7 @@ type domaininfo =
max_vcpu_id : int;
ssidref : int32;
handle : int array;
+ arch_config : arch_domainconfig;
}
type sched_control =
diff --git a/tools/ocaml/libs/xc/xenctrl.mli b/tools/ocaml/libs/xc/xenctrl.mli
index 702d8a7ab8..08f1fd26ae 100644
--- a/tools/ocaml/libs/xc/xenctrl.mli
+++ b/tools/ocaml/libs/xc/xenctrl.mli
@@ -22,6 +22,33 @@ type vcpuinfo = {
cputime : int64;
cpumap : int32;
}
+
+type xen_arm_arch_domainconfig = {
+ gic_version: int;
+ nr_spis: int;
+ clock_frequency: int32;
+}
+
+type x86_arch_emulation_flags =
+ | X86_EMU_LAPIC
+ | X86_EMU_HPET
+ | X86_EMU_PM
+ | X86_EMU_RTC
+ | X86_EMU_IOAPIC
+ | X86_EMU_PIC
+ | X86_EMU_VGA
+ | X86_EMU_IOMMU
+ | X86_EMU_PIT
+ | X86_EMU_USE_PIRQ
+
+type xen_x86_arch_domainconfig = {
+ emulation_flags: x86_arch_emulation_flags list;
+}
+
+type arch_domainconfig =
+ | ARM of xen_arm_arch_domainconfig
+ | X86 of xen_x86_arch_domainconfig
+
type domaininfo = {
domid : domid;
dying : bool;
@@ -39,6 +66,7 @@ type domaininfo = {
max_vcpu_id : int;
ssidref : int32;
handle : int array;
+ arch_config : arch_domainconfig;
}
type sched_control = { weight : int; cap : int; }
type physinfo_cap_flag = CAP_HVM | CAP_DirectIO
diff --git a/tools/ocaml/libs/xc/xenctrl_stubs.c b/tools/ocaml/libs/xc/xenctrl_stubs.c
index c66732f67c..124aa34fe8 100644
--- a/tools/ocaml/libs/xc/xenctrl_stubs.c
+++ b/tools/ocaml/libs/xc/xenctrl_stubs.c
@@ -273,10 +273,10 @@ CAMLprim value stub_xc_domain_shutdown(value xch, value domid, value reason)
static value alloc_domaininfo(xc_domaininfo_t * info)
{
CAMLparam0();
- CAMLlocal2(result, tmp);
+ CAMLlocal5(result, tmp, arch_config, x86_arch_config, emul_list);
int i;
- result = caml_alloc_tuple(16);
+ result = caml_alloc_tuple(17);
Store_field(result, 0, Val_int(info->domain));
Store_field(result, 1, Val_bool(info->flags & XEN_DOMINF_dying));
@@ -302,6 +302,28 @@ static value alloc_domaininfo(xc_domaininfo_t * info)
Store_field(result, 15, tmp);
+ /* emulation_flags: x86_arch_emulation_flags list; */
+ tmp = emul_list = Val_emptylist;
+ for (i = 0; i < 10; i++) {
+ if ((info->arch_config.emulation_flags >> i) & 1) {
+ tmp = caml_alloc_small(2, Tag_cons);
+ Field(tmp, 0) = Val_int(i);
+ Field(tmp, 1) = emul_list;
+ emul_list = tmp;
+ }
+ }
+
+ /* xen_x86_arch_domainconfig */
+ x86_arch_config = caml_alloc_tuple(1);
+ Store_field(x86_arch_config, 0, emul_list);
+
+ /* arch_config: arch_domainconfig */
+ arch_config = caml_alloc_small(1, 1);
+
+ Store_field(arch_config, 0, x86_arch_config);
+
+ Store_field(result, 16, arch_config);
+
CAMLreturn(result);
}
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 12/62] tools/ocaml: Extend domain_create() to take arch_domainconfig
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (10 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 11/62] tools/ocaml: Expose arch_config in domaininfo Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 13/62] x86/fixmap: Modify fix_to_virt() to return a void pointer Wei Liu
` (49 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Jon Ludlam <jonathan.ludlam@citrix.com>
No longer passing NULL into xc_domain_create() allows for the creation
of PVH guests.
Signed-off-by: Jon Ludlam <jonathan.ludlam@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
tools/ocaml/libs/xc/xenctrl.ml | 2 +-
tools/ocaml/libs/xc/xenctrl.mli | 2 +-
tools/ocaml/libs/xc/xenctrl_stubs.c | 22 ++++++++++++++++++++--
3 files changed, 22 insertions(+), 4 deletions(-)
diff --git a/tools/ocaml/libs/xc/xenctrl.ml b/tools/ocaml/libs/xc/xenctrl.ml
index d549068d60..9116aa222c 100644
--- a/tools/ocaml/libs/xc/xenctrl.ml
+++ b/tools/ocaml/libs/xc/xenctrl.ml
@@ -143,7 +143,7 @@ let with_intf f =
interface_close xc;
r
-external _domain_create: handle -> int32 -> domain_create_flag list -> int array -> domid
+external _domain_create: handle -> int32 -> domain_create_flag list -> int array -> arch_domainconfig -> domid
= "stub_xc_domain_create"
let int_array_of_uuid_string s =
diff --git a/tools/ocaml/libs/xc/xenctrl.mli b/tools/ocaml/libs/xc/xenctrl.mli
index 08f1fd26ae..54c099c88f 100644
--- a/tools/ocaml/libs/xc/xenctrl.mli
+++ b/tools/ocaml/libs/xc/xenctrl.mli
@@ -102,7 +102,7 @@ external sizeof_xen_pfn : unit -> int = "stub_sizeof_xen_pfn"
external interface_open : unit -> handle = "stub_xc_interface_open"
external interface_close : handle -> unit = "stub_xc_interface_close"
val with_intf : (handle -> 'a) -> 'a
-val domain_create : handle -> int32 -> domain_create_flag list -> string -> domid
+val domain_create : handle -> int32 -> domain_create_flag list -> string -> arch_domainconfig -> domid
val domain_sethandle : handle -> domid -> string -> unit
external domain_max_vcpus : handle -> domid -> int -> unit
= "stub_xc_domain_max_vcpus"
diff --git a/tools/ocaml/libs/xc/xenctrl_stubs.c b/tools/ocaml/libs/xc/xenctrl_stubs.c
index 124aa34fe8..0b5a2361c0 100644
--- a/tools/ocaml/libs/xc/xenctrl_stubs.c
+++ b/tools/ocaml/libs/xc/xenctrl_stubs.c
@@ -144,7 +144,8 @@ static int domain_create_flag_table[] = {
};
CAMLprim value stub_xc_domain_create(value xch, value ssidref,
- value flags, value handle)
+ value flags, value handle,
+ value domconfig)
{
CAMLparam4(xch, ssidref, flags, handle);
@@ -155,6 +156,7 @@ CAMLprim value stub_xc_domain_create(value xch, value ssidref,
uint32_t c_ssidref = Int32_val(ssidref);
unsigned int c_flags = 0;
value l;
+ xc_domain_configuration_t config = {};
if (Wosize_val(handle) != 16)
caml_invalid_argument("Handle not a 16-integer array");
@@ -168,8 +170,24 @@ CAMLprim value stub_xc_domain_create(value xch, value ssidref,
c_flags |= domain_create_flag_table[v];
}
+ switch(Tag_val(domconfig)) {
+ case 0: /* ARM - nothing to do */
+ caml_failwith("Unhandled: ARM");
+ break;
+
+ case 1: /* X86 - emulation flags in the block */
+ for (l = Field(Field(domconfig, 0), 0);
+ l != Val_none;
+ l = Field(l, 1))
+ config.emulation_flags |= 1u << Int_val(Field(l, 0));
+ break;
+
+ default:
+ caml_failwith("Unhandled domconfig type");
+ }
+
caml_enter_blocking_section();
- result = xc_domain_create(_H(xch), c_ssidref, h, c_flags, &domid, NULL);
+ result = xc_domain_create(_H(xch), c_ssidref, h, c_flags, &domid, &config);
caml_leave_blocking_section();
if (result < 0)
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 13/62] x86/fixmap: Modify fix_to_virt() to return a void pointer
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (11 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 12/62] tools/ocaml: Extend domain_create() to take arch_domainconfig Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 14/62] x86: Common cpuid faulting support Wei Liu
` (48 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Almost all users of fix_to_virt() actually want a pointer. Include the cast
within the definition, so the callers don't need to.
Two users which need the integer value are switched to using __fix_to_virt()
directly. A few users stay fully unchanged, due to GCC's void pointer
arithmetic extension causing the same behaviour. Most users however have
their explicit casting dropped.
Since __iomem is not used consistently in Xen, we drop it too.
No functional change.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Reviewed-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2: update commit message and remove unnecessary parentheses.
---
xen/arch/x86/acpi/lib.c | 2 +-
xen/arch/x86/mm.c | 4 ++--
xen/arch/x86/mpparse.c | 2 +-
xen/arch/x86/msi.c | 3 +--
xen/arch/x86/tboot.c | 4 ++--
xen/drivers/acpi/apei/apei-io.c | 2 +-
xen/drivers/char/ehci-dbgp.c | 2 +-
xen/drivers/char/ns16550.c | 2 +-
xen/include/asm-x86/apicdef.h | 2 +-
xen/include/asm-x86/fixmap.h | 2 +-
10 files changed, 12 insertions(+), 13 deletions(-)
diff --git a/xen/arch/x86/acpi/lib.c b/xen/arch/x86/acpi/lib.c
index 7d7c71848b..265b9ad819 100644
--- a/xen/arch/x86/acpi/lib.c
+++ b/xen/arch/x86/acpi/lib.c
@@ -49,7 +49,7 @@ char *__acpi_map_table(paddr_t phys, unsigned long size)
offset = phys & (PAGE_SIZE - 1);
mapped_size = PAGE_SIZE - offset;
set_fixmap(FIX_ACPI_END, phys);
- base = fix_to_virt(FIX_ACPI_END);
+ base = __fix_to_virt(FIX_ACPI_END);
/*
* Most cases can be covered by the below.
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index a56f875d45..f73fee225e 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -5227,12 +5227,12 @@ void __set_fixmap(
enum fixed_addresses idx, unsigned long mfn, unsigned long flags)
{
BUG_ON(idx >= __end_of_fixed_addresses);
- map_pages_to_xen(fix_to_virt(idx), mfn, 1, flags);
+ map_pages_to_xen(__fix_to_virt(idx), mfn, 1, flags);
}
void *__init arch_vmap_virt_end(void)
{
- return (void *)fix_to_virt(__end_of_fixed_addresses);
+ return fix_to_virt(__end_of_fixed_addresses);
}
void __iomem *ioremap(paddr_t pa, size_t len)
diff --git a/xen/arch/x86/mpparse.c b/xen/arch/x86/mpparse.c
index a1a0738a19..49140e46f0 100644
--- a/xen/arch/x86/mpparse.c
+++ b/xen/arch/x86/mpparse.c
@@ -703,7 +703,7 @@ static void __init efi_check_config(void)
return;
__set_fixmap(FIX_EFI_MPF, PFN_DOWN(efi.mps), __PAGE_HYPERVISOR);
- mpf = (void *)fix_to_virt(FIX_EFI_MPF) + ((long)efi.mps & (PAGE_SIZE-1));
+ mpf = fix_to_virt(FIX_EFI_MPF) + ((long)efi.mps & (PAGE_SIZE-1));
if (memcmp(mpf->mpf_signature, "_MP_", 4) == 0 &&
mpf->mpf_length == 1 &&
diff --git a/xen/arch/x86/msi.c b/xen/arch/x86/msi.c
index 095bd3cae7..8c89f072a8 100644
--- a/xen/arch/x86/msi.c
+++ b/xen/arch/x86/msi.c
@@ -960,8 +960,7 @@ static int msix_capability_init(struct pci_dev *dev,
xfree(entry);
return idx;
}
- base = (void *)(fix_to_virt(idx) +
- ((unsigned long)entry_paddr & (PAGE_SIZE - 1)));
+ base = fix_to_virt(idx) + (entry_paddr & (PAGE_SIZE - 1));
/* Mask interrupt here */
writel(1, base + PCI_MSIX_ENTRY_VECTOR_CTRL_OFFSET);
diff --git a/xen/arch/x86/tboot.c b/xen/arch/x86/tboot.c
index 59d7c477f4..d36bf33407 100644
--- a/xen/arch/x86/tboot.c
+++ b/xen/arch/x86/tboot.c
@@ -82,7 +82,7 @@ static void __init tboot_copy_memory(unsigned char *va, uint32_t size,
{
map_base = PFN_DOWN(pa + i);
set_fixmap(FIX_TBOOT_MAP_ADDRESS, map_base << PAGE_SHIFT);
- map_addr = (unsigned char *)fix_to_virt(FIX_TBOOT_MAP_ADDRESS);
+ map_addr = fix_to_virt(FIX_TBOOT_MAP_ADDRESS);
}
va[i] = map_addr[pa + i - (map_base << PAGE_SHIFT)];
}
@@ -98,7 +98,7 @@ void __init tboot_probe(void)
/* Map and check for tboot UUID. */
set_fixmap(FIX_TBOOT_SHARED_BASE, opt_tboot_pa);
- tboot_shared = (tboot_shared_t *)fix_to_virt(FIX_TBOOT_SHARED_BASE);
+ tboot_shared = fix_to_virt(FIX_TBOOT_SHARED_BASE);
if ( tboot_shared == NULL )
return;
if ( memcmp(&tboot_shared_uuid, (uuid_t *)tboot_shared, sizeof(uuid_t)) )
diff --git a/xen/drivers/acpi/apei/apei-io.c b/xen/drivers/acpi/apei/apei-io.c
index 8955de935e..89b70f45ef 100644
--- a/xen/drivers/acpi/apei/apei-io.c
+++ b/xen/drivers/acpi/apei/apei-io.c
@@ -92,7 +92,7 @@ static void __iomem *__init apei_range_map(paddr_t paddr, unsigned long size)
apei_range_nr++;
}
- return (void __iomem *)fix_to_virt(FIX_APEI_RANGE_BASE + start_nr);
+ return fix_to_virt(FIX_APEI_RANGE_BASE + start_nr);
}
/*
diff --git a/xen/drivers/char/ehci-dbgp.c b/xen/drivers/char/ehci-dbgp.c
index d48e777c34..d0071d3114 100644
--- a/xen/drivers/char/ehci-dbgp.c
+++ b/xen/drivers/char/ehci-dbgp.c
@@ -1327,7 +1327,7 @@ static void __init ehci_dbgp_init_preirq(struct serial_port *port)
* than enough. 1k is the biggest that was seen.
*/
set_fixmap_nocache(FIX_EHCI_DBGP, dbgp->bar_val);
- ehci_bar = (void __iomem *)fix_to_virt(FIX_EHCI_DBGP);
+ ehci_bar = fix_to_virt(FIX_EHCI_DBGP);
ehci_bar += dbgp->bar_val & ~PAGE_MASK;
dbgp_printk("ehci_bar: %p\n", ehci_bar);
diff --git a/xen/drivers/char/ns16550.c b/xen/drivers/char/ns16550.c
index e0f8199f98..f32dbd3247 100644
--- a/xen/drivers/char/ns16550.c
+++ b/xen/drivers/char/ns16550.c
@@ -697,7 +697,7 @@ static void __init ns16550_init_preirq(struct serial_port *port)
enum fixed_addresses idx = FIX_COM_BEGIN + (uart - ns16550_com);
set_fixmap_nocache(idx, uart->io_base);
- uart->remapped_io_base = (void __iomem *)fix_to_virt(idx);
+ uart->remapped_io_base = fix_to_virt(idx);
uart->remapped_io_base += uart->io_base & ~PAGE_MASK;
#else
uart->remapped_io_base = (char *)ioremap(uart->io_base, uart->io_size);
diff --git a/xen/include/asm-x86/apicdef.h b/xen/include/asm-x86/apicdef.h
index eed504a31a..2fa0b77a8a 100644
--- a/xen/include/asm-x86/apicdef.h
+++ b/xen/include/asm-x86/apicdef.h
@@ -119,7 +119,7 @@
/* Only available in x2APIC mode */
#define APIC_SELF_IPI 0x3F0
-#define APIC_BASE (fix_to_virt(FIX_APIC_BASE))
+#define APIC_BASE __fix_to_virt(FIX_APIC_BASE)
/* It's only used in x2APIC mode of an x2APIC unit. */
#define APIC_MSR_BASE 0x800
diff --git a/xen/include/asm-x86/fixmap.h b/xen/include/asm-x86/fixmap.h
index 89bf6cb611..51b0e7e945 100644
--- a/xen/include/asm-x86/fixmap.h
+++ b/xen/include/asm-x86/fixmap.h
@@ -79,7 +79,7 @@ extern void __set_fixmap(
#define __fix_to_virt(x) (FIXADDR_TOP - ((x) << PAGE_SHIFT))
#define __virt_to_fix(x) ((FIXADDR_TOP - ((x)&PAGE_MASK)) >> PAGE_SHIFT)
-#define fix_to_virt(x) (__fix_to_virt(x))
+#define fix_to_virt(x) ((void *)__fix_to_virt(x))
static inline unsigned long virt_to_fix(const unsigned long vaddr)
{
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 14/62] x86: Common cpuid faulting support
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (12 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 13/62] x86/fixmap: Modify fix_to_virt() to return a void pointer Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 15/62] x86/Kconfig: Options for Xen and PVH support Wei Liu
` (47 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
With CPUID Faulting offered to SVM guests, move Xen's faulting code to being
common rather than Intel specific.
This is necessary for nested Xen (inc. pv-shim mode) to prevent PV guests from
finding the outer HVM Xen leaves via native cpuid.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
xen/arch/x86/cpu/amd.c | 16 +++++---
xen/arch/x86/cpu/common.c | 76 ++++++++++++++++++++++++++++++++++++--
xen/arch/x86/cpu/intel.c | 81 +++++++----------------------------------
xen/include/asm-x86/cpuid.h | 3 --
xen/include/asm-x86/processor.h | 4 +-
5 files changed, 98 insertions(+), 82 deletions(-)
diff --git a/xen/arch/x86/cpu/amd.c b/xen/arch/x86/cpu/amd.c
index 5f36ac75a7..2bff3ee377 100644
--- a/xen/arch/x86/cpu/amd.c
+++ b/xen/arch/x86/cpu/amd.c
@@ -198,11 +198,12 @@ static void __init noinline probe_masking_msrs(void)
}
/*
- * Context switch levelling state to the next domain. A parameter of NULL is
- * used to context switch to the default host state (by the cpu bringup-code,
- * crash path, etc).
+ * Context switch CPUID masking state to the next domain. Only called if
+ * CPUID Faulting isn't available, but masking MSRs have been detected. A
+ * parameter of NULL is used to context switch to the default host state (by
+ * the cpu bringup-code, crash path, etc).
*/
-static void amd_ctxt_switch_levelling(const struct vcpu *next)
+static void amd_ctxt_switch_masking(const struct vcpu *next)
{
struct cpuidmasks *these_masks = &this_cpu(cpuidmasks);
const struct domain *nextd = next ? next->domain : NULL;
@@ -263,6 +264,9 @@ static void __init noinline amd_init_levelling(void)
{
const struct cpuidmask *m = NULL;
+ if (probe_cpuid_faulting())
+ return;
+
probe_masking_msrs();
if (*opt_famrev != '\0') {
@@ -352,7 +356,7 @@ static void __init noinline amd_init_levelling(void)
}
if (levelling_caps)
- ctxt_switch_levelling = amd_ctxt_switch_levelling;
+ ctxt_switch_masking = amd_ctxt_switch_masking;
}
/*
@@ -518,7 +522,7 @@ static void early_init_amd(struct cpuinfo_x86 *c)
if (c == &boot_cpu_data)
amd_init_levelling();
- amd_ctxt_switch_levelling(NULL);
+ ctxt_switch_levelling(NULL);
}
static void init_amd(struct cpuinfo_x86 *c)
diff --git a/xen/arch/x86/cpu/common.c b/xen/arch/x86/cpu/common.c
index 06e0eab132..1ff121887c 100644
--- a/xen/arch/x86/cpu/common.c
+++ b/xen/arch/x86/cpu/common.c
@@ -113,12 +113,80 @@ static const struct cpu_dev default_cpu = {
};
static const struct cpu_dev *this_cpu = &default_cpu;
-static void default_ctxt_switch_levelling(const struct vcpu *next)
+static DEFINE_PER_CPU(uint64_t, msr_misc_features);
+void (* __read_mostly ctxt_switch_masking)(const struct vcpu *next);
+
+bool __init probe_cpuid_faulting(void)
+{
+ uint64_t val;
+
+ if (rdmsr_safe(MSR_INTEL_PLATFORM_INFO, val) ||
+ !(val & MSR_PLATFORM_INFO_CPUID_FAULTING) ||
+ rdmsr_safe(MSR_INTEL_MISC_FEATURES_ENABLES,
+ this_cpu(msr_misc_features)))
+ {
+ setup_clear_cpu_cap(X86_FEATURE_CPUID_FAULTING);
+ return false;
+ }
+
+ expected_levelling_cap |= LCAP_faulting;
+ levelling_caps |= LCAP_faulting;
+ setup_force_cpu_cap(X86_FEATURE_CPUID_FAULTING);
+
+ return true;
+}
+
+static void set_cpuid_faulting(bool enable)
+{
+ uint64_t *this_misc_features = &this_cpu(msr_misc_features);
+ uint64_t val = *this_misc_features;
+
+ if (!!(val & MSR_MISC_FEATURES_CPUID_FAULTING) == enable)
+ return;
+
+ val ^= MSR_MISC_FEATURES_CPUID_FAULTING;
+
+ wrmsrl(MSR_INTEL_MISC_FEATURES_ENABLES, val);
+ *this_misc_features = val;
+}
+
+void ctxt_switch_levelling(const struct vcpu *next)
{
- /* Nop */
+ const struct domain *nextd = next ? next->domain : NULL;
+
+ if (cpu_has_cpuid_faulting) {
+ /*
+ * No need to alter the faulting setting if we are switching
+ * to idle; it won't affect any code running in idle context.
+ */
+ if (nextd && is_idle_domain(nextd))
+ return;
+ /*
+ * We *should* be enabling faulting for the control domain.
+ *
+ * Unfortunately, the domain builder (having only ever been a
+ * PV guest) expects to be able to see host cpuid state in a
+ * native CPUID instruction, to correctly build a CPUID policy
+ * for HVM guests (notably the xstate leaves).
+ *
+ * This logic is fundimentally broken for HVM toolstack
+ * domains, and faulting causes PV guests to behave like HVM
+ * guests from their point of view.
+ *
+ * Future development plans will move responsibility for
+ * generating the maximum full cpuid policy into Xen, at which
+ * this problem will disappear.
+ */
+ set_cpuid_faulting(nextd && !is_control_domain(nextd) &&
+ (is_pv_domain(nextd) ||
+ next->arch.msr->
+ misc_features_enables.cpuid_faulting));
+ return;
+ }
+
+ if (ctxt_switch_masking)
+ ctxt_switch_masking(next);
}
-void (* __read_mostly ctxt_switch_levelling)(const struct vcpu *next) =
- default_ctxt_switch_levelling;
bool_t opt_cpu_info;
boolean_param("cpuinfo", opt_cpu_info);
diff --git a/xen/arch/x86/cpu/intel.c b/xen/arch/x86/cpu/intel.c
index 8311952f1f..0888f76161 100644
--- a/xen/arch/x86/cpu/intel.c
+++ b/xen/arch/x86/cpu/intel.c
@@ -15,40 +15,6 @@
#include "cpu.h"
-static bool __init probe_intel_cpuid_faulting(void)
-{
- uint64_t x;
-
- if (rdmsr_safe(MSR_INTEL_PLATFORM_INFO, x) ||
- !(x & MSR_PLATFORM_INFO_CPUID_FAULTING))
- return 0;
-
- expected_levelling_cap |= LCAP_faulting;
- levelling_caps |= LCAP_faulting;
- setup_force_cpu_cap(X86_FEATURE_CPUID_FAULTING);
- return 1;
-}
-
-DEFINE_PER_CPU(bool, cpuid_faulting_enabled);
-
-static void set_cpuid_faulting(bool enable)
-{
- bool *this_enabled = &this_cpu(cpuid_faulting_enabled);
- uint32_t hi, lo;
-
- ASSERT(cpu_has_cpuid_faulting);
-
- if (*this_enabled == enable)
- return;
-
- rdmsr(MSR_INTEL_MISC_FEATURES_ENABLES, lo, hi);
- lo &= ~MSR_MISC_FEATURES_CPUID_FAULTING;
- if (enable)
- lo |= MSR_MISC_FEATURES_CPUID_FAULTING;
- wrmsr(MSR_INTEL_MISC_FEATURES_ENABLES, lo, hi);
-
- *this_enabled = enable;
-}
/*
* Set caps in expected_levelling_cap, probe a specific masking MSR, and set
@@ -145,40 +111,17 @@ static void __init probe_masking_msrs(void)
}
/*
- * Context switch levelling state to the next domain. A parameter of NULL is
- * used to context switch to the default host state (by the cpu bringup-code,
- * crash path, etc).
+ * Context switch CPUID masking state to the next domain. Only called if
+ * CPUID Faulting isn't available, but masking MSRs have been detected. A
+ * parameter of NULL is used to context switch to the default host state (by
+ * the cpu bringup-code, crash path, etc).
*/
-static void intel_ctxt_switch_levelling(const struct vcpu *next)
+static void intel_ctxt_switch_masking(const struct vcpu *next)
{
struct cpuidmasks *these_masks = &this_cpu(cpuidmasks);
const struct domain *nextd = next ? next->domain : NULL;
- const struct cpuidmasks *masks;
-
- if (cpu_has_cpuid_faulting) {
- /*
- * We *should* be enabling faulting for the control domain.
- *
- * Unfortunately, the domain builder (having only ever been a
- * PV guest) expects to be able to see host cpuid state in a
- * native CPUID instruction, to correctly build a CPUID policy
- * for HVM guests (notably the xstate leaves).
- *
- * This logic is fundimentally broken for HVM toolstack
- * domains, and faulting causes PV guests to behave like HVM
- * guests from their point of view.
- *
- * Future development plans will move responsibility for
- * generating the maximum full cpuid policy into Xen, at which
- * this problem will disappear.
- */
- set_cpuid_faulting(nextd && !is_control_domain(nextd) &&
- (is_pv_domain(nextd) ||
- next->arch.msr->misc_features_enables.cpuid_faulting));
- return;
- }
-
- masks = (nextd && is_pv_domain(nextd) && nextd->arch.pv_domain.cpuidmasks)
+ const struct cpuidmasks *masks =
+ (nextd && is_pv_domain(nextd) && nextd->arch.pv_domain.cpuidmasks)
? nextd->arch.pv_domain.cpuidmasks : &cpuidmask_defaults;
if (msr_basic) {
@@ -223,8 +166,10 @@ static void intel_ctxt_switch_levelling(const struct vcpu *next)
*/
static void __init noinline intel_init_levelling(void)
{
- if (!probe_intel_cpuid_faulting())
- probe_masking_msrs();
+ if (probe_cpuid_faulting())
+ return;
+
+ probe_masking_msrs();
if (msr_basic) {
uint32_t ecx, edx, tmp;
@@ -278,7 +223,7 @@ static void __init noinline intel_init_levelling(void)
}
if (levelling_caps)
- ctxt_switch_levelling = intel_ctxt_switch_levelling;
+ ctxt_switch_masking = intel_ctxt_switch_masking;
}
static void early_init_intel(struct cpuinfo_x86 *c)
@@ -316,7 +261,7 @@ static void early_init_intel(struct cpuinfo_x86 *c)
if (c == &boot_cpu_data)
intel_init_levelling();
- intel_ctxt_switch_levelling(NULL);
+ ctxt_switch_levelling(NULL);
}
/*
diff --git a/xen/include/asm-x86/cpuid.h b/xen/include/asm-x86/cpuid.h
index d2dd841e15..74d6f123e5 100644
--- a/xen/include/asm-x86/cpuid.h
+++ b/xen/include/asm-x86/cpuid.h
@@ -58,9 +58,6 @@ DECLARE_PER_CPU(struct cpuidmasks, cpuidmasks);
/* Default masking MSR values, calculated at boot. */
extern struct cpuidmasks cpuidmask_defaults;
-/* Whether or not cpuid faulting is available for the current domain. */
-DECLARE_PER_CPU(bool, cpuid_faulting_enabled);
-
#define CPUID_GUEST_NR_BASIC (0xdu + 1)
#define CPUID_GUEST_NR_FEAT (0u + 1)
#define CPUID_GUEST_NR_CACHE (5u + 1)
diff --git a/xen/include/asm-x86/processor.h b/xen/include/asm-x86/processor.h
index 41a8d8c32f..c9601b2fb2 100644
--- a/xen/include/asm-x86/processor.h
+++ b/xen/include/asm-x86/processor.h
@@ -151,7 +151,9 @@ extern struct cpuinfo_x86 boot_cpu_data;
extern struct cpuinfo_x86 cpu_data[];
#define current_cpu_data cpu_data[smp_processor_id()]
-extern void (*ctxt_switch_levelling)(const struct vcpu *next);
+extern bool probe_cpuid_faulting(void);
+extern void ctxt_switch_levelling(const struct vcpu *next);
+extern void (*ctxt_switch_masking)(const struct vcpu *next);
extern u64 host_pat;
extern bool_t opt_cpu_info;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 15/62] x86/Kconfig: Options for Xen and PVH support
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (13 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 14/62] x86: Common cpuid faulting support Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 16/62] x86/link: Relocate program headers Wei Liu
` (46 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Introduce two options. One to detect whether the binary is running on
Xen, the other enables PVH ABI support.
The former will be useful to PV in HVM approach. Both will be used by
PV in PVH approach.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2:
Write commit message. Didn't change the config option value as it
requires a lot of changes in later patches.
---
xen/arch/x86/Kconfig | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/xen/arch/x86/Kconfig b/xen/arch/x86/Kconfig
index 7c4582922f..c0b0bcdcb3 100644
--- a/xen/arch/x86/Kconfig
+++ b/xen/arch/x86/Kconfig
@@ -117,6 +117,23 @@ config TBOOT
Technology (TXT)
If unsure, say Y.
+
+config XEN_GUEST
+ def_bool n
+ prompt "Xen Guest"
+ ---help---
+ Support for Xen detecting when it is running under Xen.
+
+ If unsure, say N.
+
+config PVH_GUEST
+ def_bool n
+ prompt "PVH Guest"
+ depends on XEN_GUEST
+ ---help---
+ Support booting using the PVH ABI.
+
+ If unsure, say N.
endmenu
source "common/Kconfig"
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 16/62] x86/link: Relocate program headers
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (14 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 15/62] x86/Kconfig: Options for Xen and PVH support Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 17/62] x86: introduce ELFNOTE macro Wei Liu
` (45 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
When the xen binary is loaded by libelf (in the future) we rely on the
elf loader to load the binary accordingly. Specify the load address so
that the resulting binary can make p_vaddr and p_paddr have different
values.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2:
Clarify commit message. Haven't tested grub1 boot.
---
xen/arch/x86/xen.lds.S | 22 +++++++++++++---------
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S
index 6164ad094f..400d8a56c4 100644
--- a/xen/arch/x86/xen.lds.S
+++ b/xen/arch/x86/xen.lds.S
@@ -13,6 +13,7 @@
#undef __XEN_VIRT_START
#define __XEN_VIRT_START __image_base__
#define SECTION_ALIGN MB(2)
+#define DECL_SECTION(x) x :
ENTRY(efi_start)
@@ -20,8 +21,9 @@ ENTRY(efi_start)
#define FORMAT "elf64-x86-64"
#define SECTION_ALIGN PAGE_SIZE
+#define DECL_SECTION(x) x : AT(ADDR(x) - __XEN_VIRT_START)
-ENTRY(start)
+ENTRY(start_pa)
#endif /* EFI */
@@ -56,9 +58,11 @@ SECTIONS
__2M_text_start = .; /* Start of 2M superpages, mapped RX. */
#endif
+ start_pa = ABSOLUTE(start - __XEN_VIRT_START);
+
. = __XEN_VIRT_START + XEN_IMG_OFFSET;
_start = .;
- .text : {
+ DECL_SECTION(.text) {
_stext = .; /* Text and read-only data */
*(.text)
*(.text.cold)
@@ -73,7 +77,7 @@ SECTIONS
__2M_text_end = .;
__2M_rodata_start = .; /* Start of 2M superpages, mapped RO. */
- .rodata : {
+ DECL_SECTION(.rodata) {
_srodata = .;
/* Bug frames table */
__start_bug_frames = .;
@@ -132,13 +136,13 @@ SECTIONS
* compiler may want to inject other things in the .note which we don't care
* about - hence this unique name.
*/
- .note.gnu.build-id : {
+ DECL_SECTION(.note.gnu.build-id) {
__note_gnu_build_id_start = .;
*(.note.gnu.build-id)
__note_gnu_build_id_end = .;
} :note :text
#elif defined(BUILD_ID_EFI)
- .buildid : {
+ DECL_SECTION(.buildid) {
__note_gnu_build_id_start = .;
*(.buildid)
__note_gnu_build_id_end = .;
@@ -153,7 +157,7 @@ SECTIONS
__2M_init_start = .; /* Start of 2M superpages, mapped RWX (boot only). */
. = ALIGN(PAGE_SIZE); /* Init code and data */
__init_begin = .;
- .init : {
+ DECL_SECTION(.init) {
_sinittext = .;
*(.init.text)
/*
@@ -215,7 +219,7 @@ SECTIONS
__2M_rwdata_start = .; /* Start of 2M superpages, mapped RW. */
. = ALIGN(SMP_CACHE_BYTES);
- .data.read_mostly : {
+ DECL_SECTION(.data.read_mostly) {
*(.data.read_mostly)
. = ALIGN(8);
__start_schedulers_array = .;
@@ -223,7 +227,7 @@ SECTIONS
__end_schedulers_array = .;
} :text
- .data : { /* Data */
+ DECL_SECTION(.data) {
*(.data.page_aligned)
*(.data)
*(.data.rel)
@@ -231,7 +235,7 @@ SECTIONS
CONSTRUCTORS
} :text
- .bss : { /* BSS */
+ DECL_SECTION(.bss) {
__bss_start = .;
*(.bss.stack_aligned)
*(.bss.page_aligned*)
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 17/62] x86: introduce ELFNOTE macro
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (15 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 16/62] x86/link: Relocate program headers Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 18/62] x86: produce a binary that can be booted as PVH Wei Liu
` (44 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
It is needed later for introducing PVH entry point.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2:
1. Specify section attribute and type.
2. Use p2align.
3. Align instructions.
4. Haven't used .L or turned it into assembly macro.
---
xen/include/asm-x86/asm_defns.h | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/xen/include/asm-x86/asm_defns.h b/xen/include/asm-x86/asm_defns.h
index 7e8838e49b..eb25097de5 100644
--- a/xen/include/asm-x86/asm_defns.h
+++ b/xen/include/asm-x86/asm_defns.h
@@ -336,4 +336,16 @@ static always_inline void stac(void)
#define REX64_PREFIX "rex64/"
#endif
+#define ELFNOTE(name, type, desc) \
+ .pushsection .note.name, "a", @note ; \
+ .p2align 2 ; \
+ .long 2f - 1f /* namesz */ ; \
+ .long 4f - 3f /* descsz */ ; \
+ .long type /* type */ ; \
+1: .asciz #name /* name */ ; \
+2: .p2align 2 ; \
+3: desc /* desc */ ; \
+4: .p2align 2 ; \
+ .popsection
+
#endif /* __X86_ASM_DEFNS_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 18/62] x86: produce a binary that can be booted as PVH
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (16 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 17/62] x86: introduce ELFNOTE macro Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 19/62] x86/entry: Early PVH boot code Wei Liu
` (43 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Produce a binary that can be booted as PVH. It doesn't do much yet.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
v2:
1. Remove shim-y dependency.
2. Remove extraneous blank line.
3. Fix bugs in xen.lds.S.
4. Haven't split code into pvh.S because that will break later
patches.
---
.gitignore | 1 +
xen/arch/x86/Makefile | 8 ++++++++
xen/arch/x86/boot/head.S | 9 +++++++++
xen/arch/x86/xen.lds.S | 9 ++++++++-
4 files changed, 26 insertions(+), 1 deletion(-)
diff --git a/.gitignore b/.gitignore
index d64b03d06c..8da67daf31 100644
--- a/.gitignore
+++ b/.gitignore
@@ -323,6 +323,7 @@ xen/xsm/flask/xenpolicy-*
tools/flask/policy/policy.conf
tools/flask/policy/xenpolicy-*
xen/xen
+xen/xen-shim
xen/xen-syms
xen/xen-syms.map
xen/xen.*
diff --git a/xen/arch/x86/Makefile b/xen/arch/x86/Makefile
index d5d58a205e..01d1178530 100644
--- a/xen/arch/x86/Makefile
+++ b/xen/arch/x86/Makefile
@@ -75,6 +75,8 @@ efi-y := $(shell if [ ! -r $(BASEDIR)/include/xen/compile.h -o \
-O $(BASEDIR)/include/xen/compile.h ]; then \
echo '$(TARGET).efi'; fi)
+shim-$(CONFIG_PVH_GUEST) := $(TARGET)-shim
+
ifneq ($(build_id_linker),)
notes_phdrs = --notes
else
@@ -144,6 +146,11 @@ $(TARGET)-syms: prelink.o xen.lds $(BASEDIR)/common/symbols-dummy.o
>$(@D)/$(@F).map
rm -f $(@D)/.$(@F).[0-9]*
+# Use elf32-x86-64 if toolchain support exists, elf32-i386 otherwise.
+$(TARGET)-shim: FORMAT = $(firstword $(filter elf32-x86-64,$(shell $(OBJCOPY) --help)) elf32-i386)
+$(TARGET)-shim: $(TARGET)-syms
+ $(OBJCOPY) -O $(FORMAT) $< $@
+
note.o: $(TARGET)-syms
$(OBJCOPY) -O binary --only-section=.note.gnu.build-id $(BASEDIR)/xen-syms $@.bin
$(OBJCOPY) -I binary -O elf64-x86-64 -B i386:x86-64 \
@@ -224,5 +231,6 @@ clean::
rm -f asm-offsets.s *.lds boot/*.o boot/*~ boot/core boot/mkelf32
rm -f $(BASEDIR)/.xen-syms.[0-9]* boot/.*.d
rm -f $(BASEDIR)/.xen.efi.[0-9]* efi/*.efi efi/disabled efi/mkreloc
+ rm -f $(BASEDIR)/xen-shim
rm -f boot/cmdline.S boot/reloc.S boot/*.lnk boot/*.bin
rm -f note.o
diff --git a/xen/arch/x86/boot/head.S b/xen/arch/x86/boot/head.S
index 475c678f2c..472255a4b7 100644
--- a/xen/arch/x86/boot/head.S
+++ b/xen/arch/x86/boot/head.S
@@ -7,6 +7,7 @@
#include <asm/page.h>
#include <asm/msr.h>
#include <asm/cpufeature.h>
+#include <public/elfnote.h>
.text
.code32
@@ -374,6 +375,14 @@ cs32_switch:
/* Jump to earlier loaded address. */
jmp *%edi
+#ifdef CONFIG_PVH_GUEST
+ELFNOTE(Xen, XEN_ELFNOTE_PHYS32_ENTRY, .long sym_offs(__pvh_start))
+
+__pvh_start:
+ ud2a
+
+#endif /* CONFIG_PVH_GUEST */
+
__start:
cld
cli
diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S
index 400d8a56c4..2023f971e4 100644
--- a/xen/arch/x86/xen.lds.S
+++ b/xen/arch/x86/xen.lds.S
@@ -34,7 +34,7 @@ OUTPUT_ARCH(i386:x86-64)
PHDRS
{
text PT_LOAD ;
-#if defined(BUILD_ID) && !defined(EFI)
+#if (defined(BUILD_ID) || defined (CONFIG_PVH_GUEST)) && !defined(EFI)
note PT_NOTE ;
#endif
}
@@ -128,6 +128,12 @@ SECTIONS
__param_end = .;
} :text
+#if defined(CONFIG_PVH_GUEST) && !defined(EFI)
+ DECL_SECTION(.note.Xen) {
+ *(.note.Xen)
+ } :note :text
+#endif
+
#if defined(BUILD_ID)
#if !defined(EFI)
/*
@@ -279,6 +285,7 @@ SECTIONS
#ifdef EFI
*(.comment)
*(.comment.*)
+ *(.note.Xen)
#endif
}
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 19/62] x86/entry: Early PVH boot code
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (17 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 18/62] x86: produce a binary that can be booted as PVH Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 20/62] x86/boot: Map more than the first 16MB Wei Liu
` (42 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
v2:
1. Fix comment.
2. Use cmpb $0.
3. Address comments on pvh-boot.c.
4. Haven't changed the pritnk modifiers to accommodate future changes.
5. Missing a prerequisite patch to relocate pvh_info to make __va work reliably.
[BLOCKER].
---
xen/arch/x86/Makefile | 1 +
xen/arch/x86/boot/head.S | 40 +++++++++++-
xen/arch/x86/boot/x86_64.S | 2 +-
xen/arch/x86/guest/Makefile | 1 +
xen/arch/x86/guest/pvh-boot.c | 119 +++++++++++++++++++++++++++++++++++
xen/arch/x86/setup.c | 18 +++++-
xen/include/asm-x86/guest.h | 34 ++++++++++
xen/include/asm-x86/guest/pvh-boot.h | 57 +++++++++++++++++
8 files changed, 268 insertions(+), 4 deletions(-)
create mode 100644 xen/arch/x86/guest/Makefile
create mode 100644 xen/arch/x86/guest/pvh-boot.c
create mode 100644 xen/include/asm-x86/guest.h
create mode 100644 xen/include/asm-x86/guest/pvh-boot.h
diff --git a/xen/arch/x86/Makefile b/xen/arch/x86/Makefile
index 01d1178530..ac91e13606 100644
--- a/xen/arch/x86/Makefile
+++ b/xen/arch/x86/Makefile
@@ -1,6 +1,7 @@
subdir-y += acpi
subdir-y += cpu
subdir-y += genapic
+subdir-$(CONFIG_XEN_GUEST) += guest
subdir-$(CONFIG_HVM) += hvm
subdir-y += mm
subdir-$(CONFIG_XENOPROF) += oprofile
diff --git a/xen/arch/x86/boot/head.S b/xen/arch/x86/boot/head.S
index 472255a4b7..4fe5a776b1 100644
--- a/xen/arch/x86/boot/head.S
+++ b/xen/arch/x86/boot/head.S
@@ -379,7 +379,39 @@ cs32_switch:
ELFNOTE(Xen, XEN_ELFNOTE_PHYS32_ENTRY, .long sym_offs(__pvh_start))
__pvh_start:
- ud2a
+ cld
+ cli
+
+ /*
+ * We need one push/pop to determine load address. Use the same
+ * absolute stack address as the native path, for lack of a better
+ * alternative.
+ */
+ mov $0x1000, %esp
+
+ /* Calculate the load base address. */
+ call 1f
+1: pop %esi
+ sub $sym_offs(1b), %esi
+
+ /* Set up stack. */
+ lea STACK_SIZE + sym_esi(cpu0_stack), %esp
+
+ mov %ebx, sym_esi(pvh_start_info_pa)
+
+ /* Prepare gdt and segments */
+ add %esi, sym_esi(gdt_boot_base)
+ lgdt sym_esi(gdt_boot_descr)
+
+ mov $BOOT_DS, %ecx
+ mov %ecx, %ds
+ mov %ecx, %es
+ mov %ecx, %ss
+
+ /* Skip bootloader setup and bios setup, go straight to trampoline */
+ movb $1, sym_esi(pvh_boot)
+ movb $1, sym_esi(skip_realmode)
+ jmp trampoline_setup
#endif /* CONFIG_PVH_GUEST */
@@ -543,12 +575,18 @@ trampoline_setup:
/* Get bottom-most low-memory stack address. */
add $TRAMPOLINE_SPACE,%ecx
+#ifdef CONFIG_PVH_GUEST
+ cmpb $0, sym_fs(pvh_boot)
+ jne 1f
+#endif
+
/* Save the Multiboot info struct (after relocation) for later use. */
push %ecx /* Bottom-most low-memory stack address. */
push %ebx /* Multiboot information address. */
push %eax /* Multiboot magic. */
call reloc
mov %eax,sym_fs(multiboot_ptr)
+1:
/*
* Now trampoline_phys points to the following structure (lowest address
diff --git a/xen/arch/x86/boot/x86_64.S b/xen/arch/x86/boot/x86_64.S
index 925fd4bb0a..cf47e019f5 100644
--- a/xen/arch/x86/boot/x86_64.S
+++ b/xen/arch/x86/boot/x86_64.S
@@ -31,7 +31,7 @@ ENTRY(__high_start)
test %ebx,%ebx
jnz start_secondary
- /* Pass off the Multiboot info structure to C land. */
+ /* Pass off the Multiboot info structure to C land (if applicable). */
mov multiboot_ptr(%rip),%edi
call __start_xen
BUG /* __start_xen() shouldn't return. */
diff --git a/xen/arch/x86/guest/Makefile b/xen/arch/x86/guest/Makefile
new file mode 100644
index 0000000000..a5f1625ab1
--- /dev/null
+++ b/xen/arch/x86/guest/Makefile
@@ -0,0 +1 @@
+obj-bin-$(CONFIG_PVH_GUEST) += pvh-boot.init.o
diff --git a/xen/arch/x86/guest/pvh-boot.c b/xen/arch/x86/guest/pvh-boot.c
new file mode 100644
index 0000000000..186e332657
--- /dev/null
+++ b/xen/arch/x86/guest/pvh-boot.c
@@ -0,0 +1,119 @@
+/******************************************************************************
+ * arch/x86/guest/pvh-boot.c
+ *
+ * PVH boot time support
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+#include <xen/init.h>
+#include <xen/lib.h>
+#include <xen/mm.h>
+
+#include <asm/guest.h>
+
+#include <public/arch-x86/hvm/start_info.h>
+
+/* Initialised in head.S, before .bss is zeroed. */
+bool __initdata pvh_boot;
+uint32_t __initdata pvh_start_info_pa;
+
+static multiboot_info_t __initdata pvh_mbi;
+static module_t __initdata pvh_mbi_mods[8];
+static const char *__initdata pvh_loader = "PVH Directboot";
+
+static void __init convert_pvh_info(void)
+{
+ const struct hvm_start_info *pvh_info = __va(pvh_start_info_pa);
+ const struct hvm_modlist_entry *entry;
+ module_t *mod;
+ unsigned int i;
+
+ ASSERT(pvh_info->magic == XEN_HVM_START_MAGIC_VALUE);
+
+ /*
+ * Turn hvm_start_info into mbi. Luckily all modules are placed under 4GB
+ * boundary on x86.
+ */
+ pvh_mbi.flags = MBI_CMDLINE | MBI_MODULES | MBI_LOADERNAME;
+
+ BUG_ON(pvh_info->cmdline_paddr >> 32);
+ pvh_mbi.cmdline = pvh_info->cmdline_paddr;
+ pvh_mbi.boot_loader_name = __pa(pvh_loader);
+
+ BUG_ON(pvh_info->nr_modules >= ARRAY_SIZE(pvh_mbi_mods));
+ pvh_mbi.mods_count = pvh_info->nr_modules;
+ pvh_mbi.mods_addr = __pa(pvh_mbi_mods);
+
+ mod = pvh_mbi_mods;
+ entry = __va(pvh_info->modlist_paddr);
+ for ( i = 0; i < pvh_info->nr_modules; i++ )
+ {
+ BUG_ON(entry[i].paddr >> 32);
+ BUG_ON(entry[i].cmdline_paddr >> 32);
+
+ mod[i].mod_start = entry[i].paddr;
+ mod[i].mod_end = entry[i].paddr + entry[i].size;
+ mod[i].string = entry[i].cmdline_paddr;
+ }
+}
+
+multiboot_info_t *__init pvh_init(void)
+{
+ convert_pvh_info();
+
+ return &pvh_mbi;
+}
+
+void __init pvh_print_info(void)
+{
+ const struct hvm_start_info *pvh_info = __va(pvh_start_info_pa);
+ const struct hvm_modlist_entry *entry;
+ unsigned int i;
+
+ ASSERT(pvh_info->magic == XEN_HVM_START_MAGIC_VALUE);
+
+ printk("PVH start info: (pa %08x)\n", pvh_start_info_pa);
+ printk(" version: %u\n", pvh_info->version);
+ printk(" flags: %#"PRIx32"\n", pvh_info->flags);
+ printk(" nr_modules: %u\n", pvh_info->nr_modules);
+ printk(" modlist_pa: %016"PRIx64"\n", pvh_info->modlist_paddr);
+ printk(" cmdline_pa: %016"PRIx64"\n", pvh_info->cmdline_paddr);
+ if ( pvh_info->cmdline_paddr )
+ printk(" cmdline: '%s'\n", (char *)__va(pvh_info->cmdline_paddr));
+ printk(" rsdp_pa: %016"PRIx64"\n", pvh_info->rsdp_paddr);
+
+ entry = __va(pvh_info->modlist_paddr);
+ for ( i = 0; i < pvh_info->nr_modules; i++ )
+ {
+ printk(" mod[%u].pa: %016"PRIx64"\n", i, entry[i].paddr);
+ printk(" mod[%u].size: %016"PRIu64"\n", i, entry[i].size);
+ printk(" mod[%u].cmdline_pa: %016"PRIx64"\n",
+ i, entry[i].cmdline_paddr);
+ if ( entry[i].cmdline_paddr )
+ printk(" mod[%1u].cmdline: '%s'\n", i,
+ (char *)__va(entry[i].cmdline_paddr));
+ }
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 65170fe54c..f507d22a2d 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -51,6 +51,7 @@
#include <asm/alternative.h>
#include <asm/mc146818rtc.h>
#include <asm/cpuid.h>
+#include <asm/guest.h>
/* opt_nosmp: If true, secondary processors are ignored. */
static bool __initdata opt_nosmp;
@@ -649,8 +650,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
char *memmap_type = NULL;
char *cmdline, *kextra, *loader;
unsigned int initrdidx, domcr_flags = DOMCRF_s3_integrity;
- multiboot_info_t *mbi = __va(mbi_p);
- module_t *mod = (module_t *)__va(mbi->mods_addr);
+ multiboot_info_t *mbi;
+ module_t *mod;
unsigned long nr_pages, raw_max_page, modules_headroom, *module_map;
int i, j, e820_warn = 0, bytes = 0;
bool acpi_boot_table_init_done = false, relocated = false;
@@ -680,6 +681,16 @@ void __init noreturn __start_xen(unsigned long mbi_p)
/* Full exception support from here on in. */
+ if ( pvh_boot )
+ {
+ ASSERT(mbi_p == 0);
+ mbi = pvh_init();
+ }
+ else
+ mbi = __va(mbi_p);
+
+ mod = __va(mbi->mods_addr);
+
loader = (mbi->flags & MBI_LOADERNAME)
? (char *)__va(mbi->boot_loader_name) : "unknown";
@@ -719,6 +730,9 @@ void __init noreturn __start_xen(unsigned long mbi_p)
ehci_dbgp_init();
console_init_preirq();
+ if ( pvh_boot )
+ pvh_print_info();
+
printk("Bootloader: %s\n", loader);
printk("Command line: %s\n", cmdline);
diff --git a/xen/include/asm-x86/guest.h b/xen/include/asm-x86/guest.h
new file mode 100644
index 0000000000..630c092c25
--- /dev/null
+++ b/xen/include/asm-x86/guest.h
@@ -0,0 +1,34 @@
+/******************************************************************************
+ * asm-x86/guest.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#ifndef __X86_GUEST_H__
+#define __X86_GUEST_H__
+
+#include <asm/guest/pvh-boot.h>
+
+#endif /* __X86_GUEST_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/asm-x86/guest/pvh-boot.h b/xen/include/asm-x86/guest/pvh-boot.h
new file mode 100644
index 0000000000..1b429f9401
--- /dev/null
+++ b/xen/include/asm-x86/guest/pvh-boot.h
@@ -0,0 +1,57 @@
+/******************************************************************************
+ * asm-x86/guest/pvh-boot.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#ifndef __X86_PVH_BOOT_H__
+#define __X86_PVH_BOOT_H__
+
+#include <xen/multiboot.h>
+
+#ifdef CONFIG_PVH_GUEST
+
+extern bool pvh_boot;
+
+multiboot_info_t *pvh_init(void);
+void pvh_print_info(void);
+
+#else
+
+#define pvh_boot 0
+
+static inline multiboot_info_t *pvh_init(void)
+{
+ ASSERT_UNREACHABLE();
+ return NULL;
+}
+
+static inline void pvh_print_info(void)
+{
+ ASSERT_UNREACHABLE();
+}
+
+#endif /* CONFIG_PVH_GUEST */
+#endif /* __X86_PVH_BOOT_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 20/62] x86/boot: Map more than the first 16MB
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (18 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 19/62] x86/entry: Early PVH boot code Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 21/62] x86/entry: Probe for Xen early during boot Wei Liu
` (41 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
TODO: Replace somehow (bootstrap_map() ?)
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
xen/arch/x86/boot/x86_64.S | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/xen/arch/x86/boot/x86_64.S b/xen/arch/x86/boot/x86_64.S
index cf47e019f5..42636cf334 100644
--- a/xen/arch/x86/boot/x86_64.S
+++ b/xen/arch/x86/boot/x86_64.S
@@ -114,11 +114,10 @@ GLOBAL(__page_tables_start)
GLOBAL(l2_identmap)
.quad sym_offs(l1_identmap) + __PAGE_HYPERVISOR
idx = 1
- .rept 7
+ .rept 4 * L2_PAGETABLE_ENTRIES - 1
.quad (idx << L2_PAGETABLE_SHIFT) | PAGE_HYPERVISOR | _PAGE_PSE
idx = idx + 1
.endr
- .fill 4 * L2_PAGETABLE_ENTRIES - 8, 8, 0
.size l2_identmap, . - l2_identmap
/*
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 21/62] x86/entry: Probe for Xen early during boot
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (19 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 20/62] x86/boot: Map more than the first 16MB Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 22/62] x86/guest: Hypercall support Wei Liu
` (40 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
v2: Add __read_mostly.
---
xen/arch/x86/guest/Makefile | 2 ++
xen/arch/x86/guest/xen.c | 75 +++++++++++++++++++++++++++++++++++++++++
xen/arch/x86/setup.c | 2 ++
xen/include/asm-x86/guest.h | 1 +
xen/include/asm-x86/guest/xen.h | 47 ++++++++++++++++++++++++++
5 files changed, 127 insertions(+)
create mode 100644 xen/arch/x86/guest/xen.c
create mode 100644 xen/include/asm-x86/guest/xen.h
diff --git a/xen/arch/x86/guest/Makefile b/xen/arch/x86/guest/Makefile
index a5f1625ab1..1345a60c81 100644
--- a/xen/arch/x86/guest/Makefile
+++ b/xen/arch/x86/guest/Makefile
@@ -1 +1,3 @@
+obj-y += xen.o
+
obj-bin-$(CONFIG_PVH_GUEST) += pvh-boot.init.o
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
new file mode 100644
index 0000000000..8507757841
--- /dev/null
+++ b/xen/arch/x86/guest/xen.c
@@ -0,0 +1,75 @@
+/******************************************************************************
+ * arch/x86/guest/xen.c
+ *
+ * Support for detecting and running under Xen.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+#include <xen/init.h>
+#include <xen/types.h>
+
+#include <asm/guest.h>
+#include <asm/processor.h>
+
+#include <public/arch-x86/cpuid.h>
+
+bool __read_mostly xen_guest;
+
+static __read_mostly uint32_t xen_cpuid_base;
+
+static void __init find_xen_leaves(void)
+{
+ uint32_t eax, ebx, ecx, edx, base;
+
+ for ( base = XEN_CPUID_FIRST_LEAF;
+ base < XEN_CPUID_FIRST_LEAF + 0x10000; base += 0x100 )
+ {
+ cpuid(base, &eax, &ebx, &ecx, &edx);
+
+ if ( (ebx == XEN_CPUID_SIGNATURE_EBX) &&
+ (ecx == XEN_CPUID_SIGNATURE_ECX) &&
+ (edx == XEN_CPUID_SIGNATURE_EDX) &&
+ ((eax - base) >= 2) )
+ {
+ xen_cpuid_base = base;
+ break;
+ }
+ }
+}
+
+void __init probe_hypervisor(void)
+{
+ /* Too early to use cpu_has_hypervisor */
+ if ( !(cpuid_ecx(1) & cpufeat_mask(X86_FEATURE_HYPERVISOR)) )
+ return;
+
+ find_xen_leaves();
+
+ if ( !xen_cpuid_base )
+ return;
+
+ xen_guest = true;
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index f507d22a2d..53ffccb9c5 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -715,6 +715,8 @@ void __init noreturn __start_xen(unsigned long mbi_p)
* allocing any xenheap structures wanted in lower memory. */
kexec_early_calculations();
+ probe_hypervisor();
+
parse_video_info();
rdmsrl(MSR_EFER, this_cpu(efer));
diff --git a/xen/include/asm-x86/guest.h b/xen/include/asm-x86/guest.h
index 630c092c25..8d91f81451 100644
--- a/xen/include/asm-x86/guest.h
+++ b/xen/include/asm-x86/guest.h
@@ -20,6 +20,7 @@
#define __X86_GUEST_H__
#include <asm/guest/pvh-boot.h>
+#include <asm/guest/xen.h>
#endif /* __X86_GUEST_H__ */
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
new file mode 100644
index 0000000000..97a7c8d531
--- /dev/null
+++ b/xen/include/asm-x86/guest/xen.h
@@ -0,0 +1,47 @@
+/******************************************************************************
+ * asm-x86/guest/xen.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#ifndef __X86_GUEST_XEN_H__
+#define __X86_GUEST_XEN_H__
+
+#include <xen/types.h>
+
+#ifdef CONFIG_XEN_GUEST
+
+extern bool xen_guest;
+
+void probe_hypervisor(void);
+
+#else
+
+#define xen_guest 0
+
+static inline void probe_hypervisor(void) {};
+
+#endif /* CONFIG_XEN_GUEST */
+#endif /* __X86_GUEST_XEN_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 22/62] x86/guest: Hypercall support
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (20 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 21/62] x86/entry: Probe for Xen early during boot Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 23/62] x86/shutdown: Support for using SCHEDOP_{shutdown, reboot} Wei Liu
` (39 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2: append underscores to tmp.
---
xen/arch/x86/guest/Makefile | 1 +
xen/arch/x86/guest/hypercall_page.S | 79 ++++++++++++++++++++++++++++++
xen/arch/x86/guest/xen.c | 5 ++
xen/arch/x86/xen.lds.S | 1 +
xen/include/asm-x86/guest.h | 1 +
xen/include/asm-x86/guest/hypercall.h | 92 +++++++++++++++++++++++++++++++++++
6 files changed, 179 insertions(+)
create mode 100644 xen/arch/x86/guest/hypercall_page.S
create mode 100644 xen/include/asm-x86/guest/hypercall.h
diff --git a/xen/arch/x86/guest/Makefile b/xen/arch/x86/guest/Makefile
index 1345a60c81..26fb4b1007 100644
--- a/xen/arch/x86/guest/Makefile
+++ b/xen/arch/x86/guest/Makefile
@@ -1,3 +1,4 @@
+obj-y += hypercall_page.o
obj-y += xen.o
obj-bin-$(CONFIG_PVH_GUEST) += pvh-boot.init.o
diff --git a/xen/arch/x86/guest/hypercall_page.S b/xen/arch/x86/guest/hypercall_page.S
new file mode 100644
index 0000000000..fdd2e72272
--- /dev/null
+++ b/xen/arch/x86/guest/hypercall_page.S
@@ -0,0 +1,79 @@
+#include <asm/page.h>
+#include <asm/asm_defns.h>
+#include <public/xen.h>
+
+ .section ".text.page_aligned", "ax", @progbits
+ .p2align PAGE_SHIFT
+
+GLOBAL(hypercall_page)
+ /* Poisoned with `ret` for safety before hypercalls are set up. */
+ .fill PAGE_SIZE, 1, 0xc3
+ .type hypercall_page, STT_OBJECT
+ .size hypercall_page, PAGE_SIZE
+
+/*
+ * Identify a specific hypercall in the hypercall page
+ * @param name Hypercall name.
+ */
+#define DECLARE_HYPERCALL(name) \
+ .globl HYPERCALL_ ## name; \
+ .set HYPERCALL_ ## name, hypercall_page + __HYPERVISOR_ ## name * 32; \
+ .type HYPERCALL_ ## name, STT_FUNC; \
+ .size HYPERCALL_ ## name, 32
+
+DECLARE_HYPERCALL(set_trap_table)
+DECLARE_HYPERCALL(mmu_update)
+DECLARE_HYPERCALL(set_gdt)
+DECLARE_HYPERCALL(stack_switch)
+DECLARE_HYPERCALL(set_callbacks)
+DECLARE_HYPERCALL(fpu_taskswitch)
+DECLARE_HYPERCALL(sched_op_compat)
+DECLARE_HYPERCALL(platform_op)
+DECLARE_HYPERCALL(set_debugreg)
+DECLARE_HYPERCALL(get_debugreg)
+DECLARE_HYPERCALL(update_descriptor)
+DECLARE_HYPERCALL(memory_op)
+DECLARE_HYPERCALL(multicall)
+DECLARE_HYPERCALL(update_va_mapping)
+DECLARE_HYPERCALL(set_timer_op)
+DECLARE_HYPERCALL(event_channel_op_compat)
+DECLARE_HYPERCALL(xen_version)
+DECLARE_HYPERCALL(console_io)
+DECLARE_HYPERCALL(physdev_op_compat)
+DECLARE_HYPERCALL(grant_table_op)
+DECLARE_HYPERCALL(vm_assist)
+DECLARE_HYPERCALL(update_va_mapping_otherdomain)
+DECLARE_HYPERCALL(iret)
+DECLARE_HYPERCALL(vcpu_op)
+DECLARE_HYPERCALL(set_segment_base)
+DECLARE_HYPERCALL(mmuext_op)
+DECLARE_HYPERCALL(xsm_op)
+DECLARE_HYPERCALL(nmi_op)
+DECLARE_HYPERCALL(sched_op)
+DECLARE_HYPERCALL(callback_op)
+DECLARE_HYPERCALL(xenoprof_op)
+DECLARE_HYPERCALL(event_channel_op)
+DECLARE_HYPERCALL(physdev_op)
+DECLARE_HYPERCALL(hvm_op)
+DECLARE_HYPERCALL(sysctl)
+DECLARE_HYPERCALL(domctl)
+DECLARE_HYPERCALL(kexec_op)
+DECLARE_HYPERCALL(tmem_op)
+DECLARE_HYPERCALL(xc_reserved_op)
+DECLARE_HYPERCALL(xenpmu_op)
+
+DECLARE_HYPERCALL(arch_0)
+DECLARE_HYPERCALL(arch_1)
+DECLARE_HYPERCALL(arch_2)
+DECLARE_HYPERCALL(arch_3)
+DECLARE_HYPERCALL(arch_4)
+DECLARE_HYPERCALL(arch_5)
+DECLARE_HYPERCALL(arch_6)
+DECLARE_HYPERCALL(arch_7)
+
+/*
+ * Local variables:
+ * tab-width: 8
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 8507757841..10b90d0f61 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -22,6 +22,7 @@
#include <xen/types.h>
#include <asm/guest.h>
+#include <asm/msr.h>
#include <asm/processor.h>
#include <public/arch-x86/cpuid.h>
@@ -29,6 +30,7 @@
bool __read_mostly xen_guest;
static __read_mostly uint32_t xen_cpuid_base;
+extern char hypercall_page[];
static void __init find_xen_leaves(void)
{
@@ -61,6 +63,9 @@ void __init probe_hypervisor(void)
if ( !xen_cpuid_base )
return;
+ /* Fill the hypercall page. */
+ wrmsrl(cpuid_ebx(xen_cpuid_base + 2), __pa(hypercall_page));
+
xen_guest = true;
}
diff --git a/xen/arch/x86/xen.lds.S b/xen/arch/x86/xen.lds.S
index 2023f971e4..509f176913 100644
--- a/xen/arch/x86/xen.lds.S
+++ b/xen/arch/x86/xen.lds.S
@@ -65,6 +65,7 @@ SECTIONS
DECL_SECTION(.text) {
_stext = .; /* Text and read-only data */
*(.text)
+ *(.text.page_aligned)
*(.text.cold)
*(.text.unlikely)
*(.fixup)
diff --git a/xen/include/asm-x86/guest.h b/xen/include/asm-x86/guest.h
index 8d91f81451..5abdb8c433 100644
--- a/xen/include/asm-x86/guest.h
+++ b/xen/include/asm-x86/guest.h
@@ -19,6 +19,7 @@
#ifndef __X86_GUEST_H__
#define __X86_GUEST_H__
+#include <asm/guest/hypercall.h>
#include <asm/guest/pvh-boot.h>
#include <asm/guest/xen.h>
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
new file mode 100644
index 0000000000..d959c3dd8a
--- /dev/null
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -0,0 +1,92 @@
+/******************************************************************************
+ * asm-x86/guest/hypercall.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#ifndef __X86_XEN_HYPERCALL_H__
+#define __X86_XEN_HYPERCALL_H__
+
+#ifdef CONFIG_XEN_GUEST
+
+/*
+ * Hypercall primatives for 64bit
+ *
+ * Inputs: %rdi, %rsi, %rdx, %r10, %r8, %r9 (arguments 1-6)
+ */
+
+#define _hypercall64_1(type, hcall, a1) \
+ ({ \
+ long res, tmp__; \
+ asm volatile ( \
+ "call hypercall_page + %c[offset]" \
+ : "=a" (res), "=D" (tmp__) \
+ : [offset] "i" (hcall * 32), \
+ "1" ((long)(a1)) \
+ : "memory" ); \
+ (type)res; \
+ })
+
+#define _hypercall64_2(type, hcall, a1, a2) \
+ ({ \
+ long res, tmp__; \
+ asm volatile ( \
+ "call hypercall_page + %c[offset]" \
+ : "=a" (res), "=D" (tmp__), "=S" (tmp__) \
+ : [offset] "i" (hcall * 32), \
+ "1" ((long)(a1)), "2" ((long)(a2)) \
+ : "memory" ); \
+ (type)res; \
+ })
+
+#define _hypercall64_3(type, hcall, a1, a2, a3) \
+ ({ \
+ long res, tmp__; \
+ asm volatile ( \
+ "call hypercall_page + %c[offset]" \
+ : "=a" (res), "=D" (tmp__), "=S" (tmp__), "=d" (tmp__) \
+ : [offset] "i" (hcall * 32), \
+ "1" ((long)(a1)), "2" ((long)(a2)), "3" ((long)(a3)) \
+ : "memory" ); \
+ (type)res; \
+ })
+
+#define _hypercall64_4(type, hcall, a1, a2, a3, a4) \
+ ({ \
+ long res, tmp__; \
+ register long _a4 asm ("r10") = ((long)(a4)); \
+ asm volatile ( \
+ "call hypercall_page + %c[offset]" \
+ : "=a" (res), "=D" (tmp__), "=S" (tmp__), "=d" (tmp__), \
+ "=&r" (tmp__) \
+ : [offset] "i" (hcall * 32), \
+ "1" ((long)(a1)), "2" ((long)(a2)), "3" ((long)(a3)), \
+ "4" (_a4) \
+ : "memory" ); \
+ (type)res; \
+ })
+
+#endif /* CONFIG_XEN_GUEST */
+#endif /* __X86_XEN_HYPERCALL_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 23/62] x86/shutdown: Support for using SCHEDOP_{shutdown, reboot}
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (21 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 22/62] x86/guest: Hypercall support Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 24/62] x86/pvh: Retrieve memory map from Xen Wei Liu
` (38 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
v2:
1. Use sched_shutdown
2. Move header inclusion
---
docs/misc/xen-command-line.markdown | 3 +++
xen/arch/x86/shutdown.c | 34 ++++++++++++++++++++++++++++++----
xen/include/asm-x86/guest/hypercall.h | 32 ++++++++++++++++++++++++++++++++
3 files changed, 65 insertions(+), 4 deletions(-)
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index 781110d4b2..e5979bceee 100644
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -1478,6 +1478,9 @@ Specify the host reboot method.
'efi' instructs Xen to reboot using the EFI reboot call (in EFI mode by
default it will use that method first).
+`xen` instructs Xen to reboot using Xen's SCHEDOP hypercall (this is the default
+when running nested Xen)
+
### rmrr
> '= start<-end>=[s1]bdf1[,[s1]bdf2[,...]];start<-end>=[s2]bdf1[,[s2]bdf2[,...]]
diff --git a/xen/arch/x86/shutdown.c b/xen/arch/x86/shutdown.c
index a87aa60add..689f6f137d 100644
--- a/xen/arch/x86/shutdown.c
+++ b/xen/arch/x86/shutdown.c
@@ -25,6 +25,7 @@
#include <asm/mpspec.h>
#include <asm/tboot.h>
#include <asm/apic.h>
+#include <asm/guest.h>
enum reboot_type {
BOOT_INVALID,
@@ -34,6 +35,7 @@ enum reboot_type {
BOOT_CF9 = 'p',
BOOT_CF9_PWR = 'P',
BOOT_EFI = 'e',
+ BOOT_XEN = 'x',
};
static int reboot_mode;
@@ -49,6 +51,7 @@ static int reboot_mode;
* pci Use the so-called "PCI reset register", CF9
* Power Like 'pci' but for a full power-cyle reset
* efi Use the EFI reboot (if running under EFI)
+ * xen Use Xen SCHEDOP hypercall (if running under Xen as a guest)
*/
static enum reboot_type reboot_type = BOOT_INVALID;
@@ -75,6 +78,7 @@ static int __init set_reboot_type(const char *str)
case 'P':
case 'p':
case 't':
+ case 'x':
reboot_type = *str;
break;
default:
@@ -93,6 +97,13 @@ static int __init set_reboot_type(const char *str)
reboot_type = BOOT_INVALID;
}
+ if ( reboot_type == BOOT_XEN && !xen_guest )
+ {
+ printk("Xen reboot selected, but Xen hypervisor not detected\n"
+ "Falling back to default\n");
+ reboot_type = BOOT_INVALID;
+ }
+
return rc;
}
custom_param("reboot", set_reboot_type);
@@ -109,6 +120,10 @@ static inline void kb_wait(void)
static void noreturn __machine_halt(void *unused)
{
local_irq_disable();
+
+ if ( reboot_type == BOOT_XEN )
+ xen_hypercall_shutdown(SHUTDOWN_poweroff);
+
for ( ; ; )
halt();
}
@@ -129,10 +144,17 @@ void machine_halt(void)
static void default_reboot_type(void)
{
- if ( reboot_type == BOOT_INVALID )
- reboot_type = efi_enabled(EFI_RS) ? BOOT_EFI
- : acpi_disabled ? BOOT_KBD
- : BOOT_ACPI;
+ if ( reboot_type != BOOT_INVALID )
+ return;
+
+ if ( xen_guest )
+ reboot_type = BOOT_XEN;
+ else if ( efi_enabled(EFI_RS) )
+ reboot_type = BOOT_EFI;
+ else if ( acpi_disabled )
+ reboot_type = BOOT_KBD;
+ else
+ reboot_type = BOOT_ACPI;
}
static int __init override_reboot(struct dmi_system_id *d)
@@ -618,6 +640,10 @@ void machine_restart(unsigned int delay_millisecs)
}
reboot_type = BOOT_ACPI;
break;
+
+ case BOOT_XEN:
+ xen_hypercall_shutdown(SHUTDOWN_reboot);
+ break;
}
}
}
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index d959c3dd8a..a05041d30b 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -21,6 +21,11 @@
#ifdef CONFIG_XEN_GUEST
+#include <xen/types.h>
+
+#include <public/xen.h>
+#include <public/sched.h>
+
/*
* Hypercall primatives for 64bit
*
@@ -78,6 +83,33 @@
(type)res; \
})
+/*
+ * Primitive Hypercall wrappers
+ */
+static inline long xen_hypercall_sched_op(unsigned int cmd, void *arg)
+{
+ return _hypercall64_2(long, __HYPERVISOR_sched_op, cmd, arg);
+}
+
+/*
+ * Higher level hypercall helpers
+ */
+static inline long xen_hypercall_shutdown(unsigned int reason)
+{
+ struct sched_shutdown s = { .reason = reason };
+ return xen_hypercall_sched_op(SCHEDOP_shutdown, &s);
+}
+
+#else /* CONFIG_XEN_GUEST */
+
+#include <public/sched.h>
+
+static inline long xen_hypercall_shutdown(unsigned int reason)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+}
+
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_XEN_HYPERCALL_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 24/62] x86/pvh: Retrieve memory map from Xen
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (22 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 23/62] x86/shutdown: Support for using SCHEDOP_{shutdown, reboot} Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 25/62] xen/console: Introduce console=xen Wei Liu
` (37 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
v2: fixed clang build, dropped rb tag
---
xen/arch/x86/e820.c | 3 +--
xen/arch/x86/guest/pvh-boot.c | 20 ++++++++++++++++++++
xen/arch/x86/guest/xen.c | 3 +++
xen/arch/x86/setup.c | 7 ++++++-
xen/include/asm-x86/e820.h | 1 +
xen/include/asm-x86/guest/hypercall.h | 5 +++++
6 files changed, 36 insertions(+), 3 deletions(-)
diff --git a/xen/arch/x86/e820.c b/xen/arch/x86/e820.c
index 7c572bade2..b422a684ee 100644
--- a/xen/arch/x86/e820.c
+++ b/xen/arch/x86/e820.c
@@ -134,8 +134,7 @@ static struct change_member *change_point[2*E820MAX] __initdata;
static struct e820entry *overlap_list[E820MAX] __initdata;
static struct e820entry new_bios[E820MAX] __initdata;
-static int __init sanitize_e820_map(struct e820entry *biosmap,
- unsigned int *pnr_map)
+int __init sanitize_e820_map(struct e820entry *biosmap, unsigned int *pnr_map)
{
struct change_member *change_tmp;
unsigned long current_type, last_type;
diff --git a/xen/arch/x86/guest/pvh-boot.c b/xen/arch/x86/guest/pvh-boot.c
index 186e332657..be3122b16c 100644
--- a/xen/arch/x86/guest/pvh-boot.c
+++ b/xen/arch/x86/guest/pvh-boot.c
@@ -22,6 +22,7 @@
#include <xen/lib.h>
#include <xen/mm.h>
+#include <asm/e820.h>
#include <asm/guest.h>
#include <public/arch-x86/hvm/start_info.h>
@@ -70,10 +71,29 @@ static void __init convert_pvh_info(void)
}
}
+static void __init get_memory_map(void)
+{
+ struct xen_memory_map memmap = {
+ .nr_entries = E820MAX,
+ };
+
+ set_xen_guest_handle(memmap.buffer, e820_raw.map);
+ BUG_ON(xen_hypercall_memory_op(XENMEM_memory_map, &memmap));
+ e820_raw.nr_map = memmap.nr_entries;
+
+ /* :( Various toolstacks don't sort the memory map. */
+ sanitize_e820_map(e820_raw.map, &e820_raw.nr_map);
+}
+
multiboot_info_t *__init pvh_init(void)
{
convert_pvh_info();
+ probe_hypervisor();
+ ASSERT(xen_guest);
+
+ get_memory_map();
+
return &pvh_mbi;
}
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 10b90d0f61..c253ebd983 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -54,6 +54,9 @@ static void __init find_xen_leaves(void)
void __init probe_hypervisor(void)
{
+ if ( xen_guest )
+ return;
+
/* Too early to use cpu_has_hypervisor */
if ( !(cpuid_ecx(1) & cpufeat_mask(X86_FEATURE_HYPERVISOR)) )
return;
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 53ffccb9c5..9c3d7563f3 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -795,7 +795,12 @@ void __init noreturn __start_xen(unsigned long mbi_p)
if ( !(mbi->flags & MBI_MODULES) || (mbi->mods_count == 0) )
panic("dom0 kernel not specified. Check bootloader configuration.");
- if ( efi_enabled(EFI_LOADER) )
+ if ( pvh_boot )
+ {
+ /* pvh_init() already filled in e820_raw */
+ memmap_type = "PVH-e820";
+ }
+ else if ( efi_enabled(EFI_LOADER) )
{
set_pdx_range(xen_phys_start >> PAGE_SHIFT,
(xen_phys_start + BOOTSTRAP_MAP_BASE) >> PAGE_SHIFT);
diff --git a/xen/include/asm-x86/e820.h b/xen/include/asm-x86/e820.h
index 28defa8545..ee317b17aa 100644
--- a/xen/include/asm-x86/e820.h
+++ b/xen/include/asm-x86/e820.h
@@ -23,6 +23,7 @@ struct e820map {
struct e820entry map[E820MAX];
};
+extern int sanitize_e820_map(struct e820entry *biosmap, unsigned int *pnr_map);
extern int e820_all_mapped(u64 start, u64 end, unsigned type);
extern int reserve_e820_ram(struct e820map *e820, uint64_t s, uint64_t e);
extern int e820_change_range_type(
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index a05041d30b..e0b00f97fb 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -91,6 +91,11 @@ static inline long xen_hypercall_sched_op(unsigned int cmd, void *arg)
return _hypercall64_2(long, __HYPERVISOR_sched_op, cmd, arg);
}
+static inline long xen_hypercall_memory_op(unsigned int cmd, void *arg)
+{
+ return _hypercall64_2(long, __HYPERVISOR_memory_op, cmd, arg);
+}
+
/*
* Higher level hypercall helpers
*/
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 25/62] xen/console: Introduce console=xen
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (23 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 24/62] x86/pvh: Retrieve memory map from Xen Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 26/62] xen: introduce rangeset_claim_range Wei Liu
` (36 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
This specifies whether to use Xen specific console output. There are
two variants: one is the hypervisor console, the other is the magic
debug port 0xe9.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/drivers/char/console.c | 46 +++++++++++++++++++++++++++++++++++
xen/include/asm-x86/guest/hypercall.h | 13 ++++++++++
2 files changed, 59 insertions(+)
diff --git a/xen/drivers/char/console.c b/xen/drivers/char/console.c
index 19d0e74f17..d05ebf9f70 100644
--- a/xen/drivers/char/console.c
+++ b/xen/drivers/char/console.c
@@ -31,6 +31,10 @@
#include <xen/early_printk.h>
#include <xen/warning.h>
+#ifdef CONFIG_X86
+#include <asm/guest.h>
+#endif
+
/* console: comma-separated list of console outputs. */
static char __initdata opt_console[30] = OPT_CONSOLE_STR;
string_param("console", opt_console);
@@ -83,6 +87,10 @@ static uint32_t conringc, conringp;
static int __read_mostly sercon_handle = -1;
+#ifdef CONFIG_X86
+static bool __read_mostly opt_console_xen; /* console=xen */
+#endif
+
static DEFINE_SPINLOCK(console_lock);
/*
@@ -432,6 +440,16 @@ static void notify_dom0_con_ring(unsigned long unused)
static DECLARE_SOFTIRQ_TASKLET(notify_dom0_con_ring_tasklet,
notify_dom0_con_ring, 0);
+#ifdef CONFIG_X86
+static inline void xen_console_write_debug_port(const char *buf, size_t len)
+{
+ unsigned long tmp;
+ asm volatile ( "rep outsb;"
+ : "=&S" (tmp), "=&c" (tmp)
+ : "0" (buf), "1" (len), "d" (0xe9) );
+}
+#endif
+
static long guest_console_write(XEN_GUEST_HANDLE_PARAM(char) buffer, int count)
{
char kbuf[128];
@@ -458,6 +476,18 @@ static long guest_console_write(XEN_GUEST_HANDLE_PARAM(char) buffer, int count)
sercon_puts(kbuf);
video_puts(kbuf);
+#ifdef CONFIG_X86
+ if ( opt_console_xen )
+ {
+ size_t len = strlen(kbuf);
+
+ if ( xen_guest )
+ xen_hypercall_console_write(kbuf, len);
+ else
+ xen_console_write_debug_port(kbuf, len);
+ }
+#endif
+
if ( opt_console_to_ring )
{
conring_puts(kbuf);
@@ -567,6 +597,18 @@ static void __putstr(const char *str)
sercon_puts(str);
video_puts(str);
+#ifdef CONFIG_X86
+ if ( opt_console_xen )
+ {
+ size_t len = strlen(str);
+
+ if ( xen_guest )
+ xen_hypercall_console_write(str, len);
+ else
+ xen_console_write_debug_port(str, len);
+ }
+#endif
+
conring_puts(str);
if ( !console_locks_busted )
@@ -762,6 +804,10 @@ void __init console_init_preirq(void)
p++;
if ( !strncmp(p, "vga", 3) )
video_init();
+#ifdef CONFIG_X86
+ else if ( !strncmp(p, "xen", 3) )
+ opt_console_xen = true;
+#endif
else if ( !strncmp(p, "none", 4) )
continue;
else if ( (sh = serial_parse_handle(p)) >= 0 )
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index e0b00f97fb..9cd95d2b92 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -99,6 +99,13 @@ static inline long xen_hypercall_memory_op(unsigned int cmd, void *arg)
/*
* Higher level hypercall helpers
*/
+static inline void xen_hypercall_console_write(
+ const char *buf, unsigned int count)
+{
+ (void)_hypercall64_3(long, __HYPERVISOR_console_io,
+ CONSOLEIO_write, count, buf);
+}
+
static inline long xen_hypercall_shutdown(unsigned int reason)
{
struct sched_shutdown s = { .reason = reason };
@@ -109,6 +116,12 @@ static inline long xen_hypercall_shutdown(unsigned int reason)
#include <public/sched.h>
+static inline void xen_hypercall_console_write(
+ const char *buf, unsigned int count)
+{
+ ASSERT_UNREACHABLE();
+}
+
static inline long xen_hypercall_shutdown(unsigned int reason)
{
ASSERT_UNREACHABLE();
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 26/62] xen: introduce rangeset_claim_range
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (24 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 25/62] xen/console: Introduce console=xen Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 27/62] xen/pvshim: keep track of used PFN ranges Wei Liu
` (35 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Reserve a hole in a rangeset.
Signed-off-by: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
Changes since v1:
- Change function name.
- Use a local variable instead of *s.
- Add unlikely to the !prev case.
- Move the function prototype position in the header file.
---
xen/common/rangeset.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++
xen/include/xen/rangeset.h | 4 +++-
2 files changed, 55 insertions(+), 1 deletion(-)
diff --git a/xen/common/rangeset.c b/xen/common/rangeset.c
index 6c6293c15c..ade34f6a50 100644
--- a/xen/common/rangeset.c
+++ b/xen/common/rangeset.c
@@ -298,6 +298,58 @@ int rangeset_report_ranges(
return rc;
}
+int rangeset_claim_range(struct rangeset *r, unsigned long size,
+ unsigned long *s)
+{
+ struct range *prev, *next;
+ unsigned long start = 0;
+
+ write_lock(&r->lock);
+
+ for ( prev = NULL, next = first_range(r);
+ next;
+ prev = next, next = next_range(r, next) )
+ {
+ if ( (next->s - start) >= size )
+ goto insert;
+
+ if ( next->e == ~0UL )
+ goto out;
+
+ start = next->e + 1;
+ }
+
+ if ( (~0UL - start) + 1 >= size )
+ goto insert;
+
+ out:
+ write_unlock(&r->lock);
+ return -ENOSPC;
+
+ insert:
+ if ( unlikely(!prev) )
+ {
+ next = alloc_range(r);
+ if ( !next )
+ {
+ write_unlock(&r->lock);
+ return -ENOMEM;
+ }
+
+ next->s = start;
+ next->e = start + size - 1;
+ insert_range(r, prev, next);
+ }
+ else
+ prev->e += size;
+
+ write_unlock(&r->lock);
+
+ *s = start;
+
+ return 0;
+}
+
int rangeset_add_singleton(
struct rangeset *r, unsigned long s)
{
diff --git a/xen/include/xen/rangeset.h b/xen/include/xen/rangeset.h
index aa6408248b..1f83b1f44b 100644
--- a/xen/include/xen/rangeset.h
+++ b/xen/include/xen/rangeset.h
@@ -55,9 +55,11 @@ void rangeset_limit(
bool_t __must_check rangeset_is_empty(
const struct rangeset *r);
-/* Add/remove/query a numeric range. */
+/* Add/claim/remove/query a numeric range. */
int __must_check rangeset_add_range(
struct rangeset *r, unsigned long s, unsigned long e);
+int __must_check rangeset_claim_range(struct rangeset *r, unsigned long size,
+ unsigned long *s);
int __must_check rangeset_remove_range(
struct rangeset *r, unsigned long s, unsigned long e);
bool_t __must_check rangeset_contains_range(
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 27/62] xen/pvshim: keep track of used PFN ranges
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (25 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 26/62] xen: introduce rangeset_claim_range Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 28/62] x86/guest: map shared_info page Wei Liu
` (34 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Simple infrastructure to keep track of PFN space usage, so that we can
use unpopulated PFNs to map special pages like shared info and grant
table.
As rangeset depends on malloc being ready so hypervisor_setup is
introduced for things that can be initialised late in the process.
Note that the PFN is marked as reserved at least up to 4GiB (or more
if the guest has more memory). This is not a perfect solution but
avoids using the MMIO hole below 4GiB. Ideally the shim (L1) should
have a way to ask the underlying Xen (L0) which memory regions are
populated, unpopulated, or MMIO space.
Signed-off-by: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/arch/x86/guest/xen.c | 56 +++++++++++++++++++++++++++++++++++++++++
xen/arch/x86/setup.c | 3 +++
xen/include/asm-x86/guest/xen.h | 7 ++++++
3 files changed, 66 insertions(+)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index c253ebd983..abf53ebbc6 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -19,8 +19,12 @@
* Copyright (c) 2017 Citrix Systems Ltd.
*/
#include <xen/init.h>
+#include <xen/mm.h>
+#include <xen/pfn.h>
+#include <xen/rangeset.h>
#include <xen/types.h>
+#include <asm/e820.h>
#include <asm/guest.h>
#include <asm/msr.h>
#include <asm/processor.h>
@@ -31,6 +35,7 @@ bool __read_mostly xen_guest;
static __read_mostly uint32_t xen_cpuid_base;
extern char hypercall_page[];
+static struct rangeset *mem;
static void __init find_xen_leaves(void)
{
@@ -72,6 +77,57 @@ void __init probe_hypervisor(void)
xen_guest = true;
}
+static void __init init_memmap(void)
+{
+ unsigned int i;
+
+ mem = rangeset_new(NULL, "host memory map", 0);
+ if ( !mem )
+ panic("failed to allocate PFN usage rangeset");
+
+ /*
+ * Mark up to the last memory page (or 4GiB) as RAM. This is done because
+ * Xen doesn't know the position of possible MMIO holes, so at least try to
+ * avoid the know MMIO hole below 4GiB. Note that this is subject to future
+ * discussion and improvements.
+ */
+ if ( rangeset_add_range(mem, 0, max_t(unsigned long, max_page - 1,
+ PFN_DOWN(GB(4) - 1))) )
+ panic("unable to add RAM to in-use PFN rangeset");
+
+ for ( i = 0; i < e820.nr_map; i++ )
+ {
+ struct e820entry *e = &e820.map[i];
+
+ if ( rangeset_add_range(mem, PFN_DOWN(e->addr),
+ PFN_UP(e->addr + e->size - 1)) )
+ panic("unable to add range [%#lx, %#lx] to in-use PFN rangeset",
+ PFN_DOWN(e->addr), PFN_UP(e->addr + e->size - 1));
+ }
+}
+
+void __init hypervisor_setup(void)
+{
+ init_memmap();
+}
+
+int hypervisor_alloc_unused_page(mfn_t *mfn)
+{
+ unsigned long m;
+ int rc;
+
+ rc = rangeset_claim_range(mem, 1, &m);
+ if ( !rc )
+ *mfn = _mfn(m);
+
+ return rc;
+}
+
+int hypervisor_free_unused_page(mfn_t mfn)
+{
+ return rangeset_remove_range(mem, mfn_x(mfn), mfn_x(mfn));
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 9c3d7563f3..32ec741591 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -1472,6 +1472,9 @@ void __init noreturn __start_xen(unsigned long mbi_p)
max_cpus = nr_cpu_ids;
}
+ if ( xen_guest )
+ hypervisor_setup();
+
/* Low mappings were only needed for some BIOS table parsing. */
zap_low_mappings();
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index 97a7c8d531..427837797b 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -26,12 +26,19 @@
extern bool xen_guest;
void probe_hypervisor(void);
+void hypervisor_setup(void);
+int hypervisor_alloc_unused_page(mfn_t *mfn);
+int hypervisor_free_unused_page(mfn_t mfn);
#else
#define xen_guest 0
static inline void probe_hypervisor(void) {};
+static inline void hypervisor_setup(void)
+{
+ ASSERT_UNREACHABLE();
+}
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_GUEST_XEN_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 28/62] x86/guest: map shared_info page
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (26 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 27/62] xen/pvshim: keep track of used PFN ranges Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 29/62] xen/guest: fetch vCPU ID from Xen Wei Liu
` (33 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Use an unpopulated PFN in order to map it.
Signed-off-by: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
Changes since v1:
- Use an unpopulated PFN to map the shared_info page.
- Mask all event channels.
- Report XENMEM_add_to_physmap error code in case of failure.
---
xen/arch/x86/guest/xen.c | 27 +++++++++++++++++++++++++++
xen/include/asm-x86/fixmap.h | 3 +++
xen/include/asm-x86/guest/xen.h | 5 +++++
3 files changed, 35 insertions(+)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index abf53ebbc6..f62f93af16 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -77,6 +77,31 @@ void __init probe_hypervisor(void)
xen_guest = true;
}
+static void map_shared_info(void)
+{
+ mfn_t mfn;
+ struct xen_add_to_physmap xatp = {
+ .domid = DOMID_SELF,
+ .space = XENMAPSPACE_shared_info,
+ };
+ unsigned int i;
+ unsigned long rc;
+
+ if ( hypervisor_alloc_unused_page(&mfn) )
+ panic("unable to reserve shared info memory page");
+
+ xatp.gpfn = mfn_x(mfn);
+ rc = xen_hypercall_memory_op(XENMEM_add_to_physmap, &xatp);
+ if ( rc )
+ panic("failed to map shared_info page: %ld", rc);
+
+ set_fixmap(FIX_XEN_SHARED_INFO, mfn_x(mfn) << PAGE_SHIFT);
+
+ /* Mask all upcalls */
+ for ( i = 0; i < ARRAY_SIZE(XEN_shared_info->evtchn_mask); i++ )
+ write_atomic(&XEN_shared_info->evtchn_mask[i], ~0ul);
+}
+
static void __init init_memmap(void)
{
unsigned int i;
@@ -109,6 +134,8 @@ static void __init init_memmap(void)
void __init hypervisor_setup(void)
{
init_memmap();
+
+ map_shared_info();
}
int hypervisor_alloc_unused_page(mfn_t *mfn)
diff --git a/xen/include/asm-x86/fixmap.h b/xen/include/asm-x86/fixmap.h
index 51b0e7e945..ded4ddf21b 100644
--- a/xen/include/asm-x86/fixmap.h
+++ b/xen/include/asm-x86/fixmap.h
@@ -45,6 +45,9 @@ enum fixed_addresses {
FIX_COM_BEGIN,
FIX_COM_END,
FIX_EHCI_DBGP,
+#ifdef CONFIG_XEN_GUEST
+ FIX_XEN_SHARED_INFO,
+#endif /* CONFIG_XEN_GUEST */
/* Everything else should go further down. */
FIX_APIC_BASE,
FIX_IO_APIC_BASE_0,
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index 427837797b..f25ad4241b 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -21,6 +21,11 @@
#include <xen/types.h>
+#include <asm/e820.h>
+#include <asm/fixmap.h>
+
+#define XEN_shared_info ((struct shared_info *)fix_to_virt(FIX_XEN_SHARED_INFO))
+
#ifdef CONFIG_XEN_GUEST
extern bool xen_guest;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 29/62] xen/guest: fetch vCPU ID from Xen
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (27 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 28/62] x86/guest: map shared_info page Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 30/62] x86/guest: map per-cpu vcpu_info area Wei Liu
` (32 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
If available.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
[ wei: fix non-shim build ]
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/arch/x86/guest/xen.c | 23 +++++++++++++++++++++++
xen/arch/x86/smpboot.c | 4 ++++
xen/include/asm-x86/guest/xen.h | 7 +++++++
3 files changed, 34 insertions(+)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index f62f93af16..de8cfc6e36 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -37,6 +37,8 @@ static __read_mostly uint32_t xen_cpuid_base;
extern char hypercall_page[];
static struct rangeset *mem;
+DEFINE_PER_CPU(unsigned int, vcpu_id);
+
static void __init find_xen_leaves(void)
{
uint32_t eax, ebx, ecx, edx, base;
@@ -102,6 +104,20 @@ static void map_shared_info(void)
write_atomic(&XEN_shared_info->evtchn_mask[i], ~0ul);
}
+static void set_vcpu_id(void)
+{
+ uint32_t eax, ebx, ecx, edx;
+
+ ASSERT(xen_cpuid_base);
+
+ /* Fetch vcpu id from cpuid. */
+ cpuid(xen_cpuid_base + 4, &eax, &ebx, &ecx, &edx);
+ if ( eax & XEN_HVM_CPUID_VCPU_ID_PRESENT )
+ this_cpu(vcpu_id) = ebx;
+ else
+ this_cpu(vcpu_id) = smp_processor_id();
+}
+
static void __init init_memmap(void)
{
unsigned int i;
@@ -136,6 +152,13 @@ void __init hypervisor_setup(void)
init_memmap();
map_shared_info();
+
+ set_vcpu_id();
+}
+
+void hypervisor_ap_setup(void)
+{
+ set_vcpu_id();
}
int hypervisor_alloc_unused_page(mfn_t *mfn)
diff --git a/xen/arch/x86/smpboot.c b/xen/arch/x86/smpboot.c
index 7b97ff86cb..945b6d594c 100644
--- a/xen/arch/x86/smpboot.c
+++ b/xen/arch/x86/smpboot.c
@@ -38,6 +38,7 @@
#include <asm/desc.h>
#include <asm/div64.h>
#include <asm/flushtlb.h>
+#include <asm/guest.h>
#include <asm/msr.h>
#include <asm/mtrr.h>
#include <asm/time.h>
@@ -372,6 +373,9 @@ void start_secondary(void *unused)
cpumask_set_cpu(cpu, &cpu_online_map);
unlock_vector_lock();
+ if ( xen_guest )
+ hypervisor_ap_setup();
+
/* We can take interrupts now: we're officially "up". */
local_irq_enable();
mtrr_ap_init();
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index f25ad4241b..db35a9e628 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -32,9 +32,12 @@ extern bool xen_guest;
void probe_hypervisor(void);
void hypervisor_setup(void);
+void hypervisor_ap_setup(void);
int hypervisor_alloc_unused_page(mfn_t *mfn);
int hypervisor_free_unused_page(mfn_t mfn);
+DECLARE_PER_CPU(unsigned int, vcpu_id);
+
#else
#define xen_guest 0
@@ -44,6 +47,10 @@ static inline void hypervisor_setup(void)
{
ASSERT_UNREACHABLE();
}
+static inline void hypervisor_ap_setup(void)
+{
+ ASSERT_UNREACHABLE();
+}
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_GUEST_XEN_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 30/62] x86/guest: map per-cpu vcpu_info area.
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (28 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 29/62] xen/guest: fetch vCPU ID from Xen Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 31/62] x86: xen pv clock time source Wei Liu
` (31 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Mapping the per-vcpu vcpu_info area is required in order to use more
than XEN_LEGACY_MAX_VCPUS.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
Changes since v1:
- Make vcpu_info_mapped static.
- Add a BUG_ON in case VCPUOP_register_vcpu_info fails.
- Remove one indentation level in hypervisor_setup.
- Make xen_hypercall_vcpu_op return int.
---
xen/arch/x86/guest/xen.c | 57 +++++++++++++++++++++++++++++++++++
xen/include/asm-x86/guest/hypercall.h | 8 +++++
xen/include/asm-x86/guest/xen.h | 1 +
3 files changed, 66 insertions(+)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index de8cfc6e36..60626ec21c 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -39,6 +39,10 @@ static struct rangeset *mem;
DEFINE_PER_CPU(unsigned int, vcpu_id);
+static struct vcpu_info *vcpu_info;
+static unsigned long vcpu_info_mapped[BITS_TO_LONGS(NR_CPUS)];
+DEFINE_PER_CPU(struct vcpu_info *, vcpu_info);
+
static void __init find_xen_leaves(void)
{
uint32_t eax, ebx, ecx, edx, base;
@@ -104,6 +108,41 @@ static void map_shared_info(void)
write_atomic(&XEN_shared_info->evtchn_mask[i], ~0ul);
}
+static int map_vcpuinfo(void)
+{
+ unsigned int vcpu = this_cpu(vcpu_id);
+ struct vcpu_register_vcpu_info info;
+ int rc;
+
+ if ( !vcpu_info )
+ {
+ this_cpu(vcpu_info) = &XEN_shared_info->vcpu_info[vcpu];
+ return 0;
+ }
+
+ if ( test_bit(vcpu, vcpu_info_mapped) )
+ {
+ this_cpu(vcpu_info) = &vcpu_info[vcpu];
+ return 0;
+ }
+
+ info.mfn = virt_to_mfn(&vcpu_info[vcpu]);
+ info.offset = (unsigned long)&vcpu_info[vcpu] & ~PAGE_MASK;
+ rc = xen_hypercall_vcpu_op(VCPUOP_register_vcpu_info, vcpu, &info);
+ if ( rc )
+ {
+ BUG_ON(vcpu >= XEN_LEGACY_MAX_VCPUS);
+ this_cpu(vcpu_info) = &XEN_shared_info->vcpu_info[vcpu];
+ }
+ else
+ {
+ this_cpu(vcpu_info) = &vcpu_info[vcpu];
+ set_bit(vcpu, vcpu_info_mapped);
+ }
+
+ return rc;
+}
+
static void set_vcpu_id(void)
{
uint32_t eax, ebx, ecx, edx;
@@ -154,11 +193,29 @@ void __init hypervisor_setup(void)
map_shared_info();
set_vcpu_id();
+ vcpu_info = xzalloc_array(struct vcpu_info, nr_cpu_ids);
+ if ( map_vcpuinfo() )
+ {
+ xfree(vcpu_info);
+ vcpu_info = NULL;
+ }
+ if ( !vcpu_info && nr_cpu_ids > XEN_LEGACY_MAX_VCPUS )
+ {
+ unsigned int i;
+
+ for ( i = XEN_LEGACY_MAX_VCPUS; i < nr_cpu_ids; i++ )
+ __cpumask_clear_cpu(i, &cpu_present_map);
+ nr_cpu_ids = XEN_LEGACY_MAX_VCPUS;
+ printk(XENLOG_WARNING
+ "unable to map vCPU info, limiting vCPUs to: %u\n",
+ XEN_LEGACY_MAX_VCPUS);
+ }
}
void hypervisor_ap_setup(void)
{
set_vcpu_id();
+ map_vcpuinfo();
}
int hypervisor_alloc_unused_page(mfn_t *mfn)
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index 9cd95d2b92..dbc57a566e 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -26,6 +26,8 @@
#include <public/xen.h>
#include <public/sched.h>
+#include <public/vcpu.h>
+
/*
* Hypercall primatives for 64bit
*
@@ -96,6 +98,12 @@ static inline long xen_hypercall_memory_op(unsigned int cmd, void *arg)
return _hypercall64_2(long, __HYPERVISOR_memory_op, cmd, arg);
}
+static inline int xen_hypercall_vcpu_op(unsigned int cmd, unsigned int vcpu,
+ void *arg)
+{
+ return _hypercall64_3(long, __HYPERVISOR_vcpu_op, cmd, vcpu, arg);
+}
+
/*
* Higher level hypercall helpers
*/
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index db35a9e628..b3e684f756 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -37,6 +37,7 @@ int hypervisor_alloc_unused_page(mfn_t *mfn);
int hypervisor_free_unused_page(mfn_t mfn);
DECLARE_PER_CPU(unsigned int, vcpu_id);
+DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
#else
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 31/62] x86: xen pv clock time source
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (29 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 30/62] x86/guest: map per-cpu vcpu_info area Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 20:45 ` Joao Martins
2018-01-12 11:28 ` [PATCH v2 32/62] x86: APIC timer calibration when running as a guest Wei Liu
` (30 subsequent siblings)
61 siblings, 1 reply; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
It is a variant of TSC clock source.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Use the mapped vcpu_info.
---
xen/arch/x86/time.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 89 insertions(+)
diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
index 3b654d7b7d..4cbd068d17 100644
--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -29,6 +29,7 @@
#include <asm/mpspec.h>
#include <asm/processor.h>
#include <asm/fixmap.h>
+#include <asm/guest.h>
#include <asm/mc146818rtc.h>
#include <asm/div64.h>
#include <asm/acpi.h>
@@ -525,6 +526,91 @@ static struct platform_timesource __initdata plt_tsc =
.init = init_tsc,
};
+#ifdef CONFIG_XEN_GUEST
+/************************************************************
+ * PLATFORM TIMER 5: XEN PV CLOCK SOURCE
+ *
+ * Xen clock source is a variant of TSC source.
+ */
+
+static uint64_t xen_timer_cpu_frequency(void)
+{
+ struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
+ uint64_t freq;
+
+ freq = 1000000000ULL << 32;
+ do_div(freq, info->tsc_to_system_mul);
+ if ( info->tsc_shift < 0 )
+ freq <<= -info->tsc_shift;
+ else
+ freq >>= info->tsc_shift;
+
+ return freq;
+}
+
+static int64_t __init init_xen_timer(struct platform_timesource *pts)
+{
+ if ( !xen_guest )
+ return 0;
+
+ pts->frequency = xen_timer_cpu_frequency();
+
+ return pts->frequency;
+}
+
+static always_inline uint64_t read_cycle(const struct vcpu_time_info *info,
+ uint64_t tsc)
+{
+ uint64_t delta = tsc - info->tsc_timestamp;
+ struct time_scale ts = {
+ .shift = info->tsc_shift,
+ .mul_frac = info->tsc_to_system_mul,
+ };
+ uint64_t offset = scale_delta(delta, &ts);
+
+ return info->system_time + offset;
+}
+
+static uint64_t read_xen_timer(void)
+{
+ struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
+ uint32_t version;
+ uint64_t ret;
+ uint64_t last;
+ static uint64_t last_value;
+
+ do {
+ version = info->version & ~1;
+ /* Make sure version is read before the data */
+ smp_rmb();
+
+ ret = read_cycle(info, rdtsc_ordered());
+ /* Ignore fancy flags for now */
+
+ /* Make sure version is reread after the data */
+ smp_rmb();
+ } while ( unlikely(version != info->version) );
+
+ /* Maintain a monotonic global value */
+ do {
+ last = read_atomic(&last_value);
+ if ( ret < last )
+ return last;
+ } while ( unlikely(cmpxchg(&last_value, last, ret) != last) );
+
+ return ret;
+}
+
+static struct platform_timesource __initdata plt_xen_timer =
+{
+ .id = "xen",
+ .name = "XEN PV CLOCK",
+ .read_counter = read_xen_timer,
+ .init = init_xen_timer,
+ .counter_bits = 63,
+};
+#endif
+
/************************************************************
* GENERIC PLATFORM TIMER INFRASTRUCTURE
*/
@@ -672,6 +758,9 @@ static s64 __init try_platform_timer(struct platform_timesource *pts)
static u64 __init init_platform_timer(void)
{
static struct platform_timesource * __initdata plt_timers[] = {
+#ifdef CONFIG_XEN_GUEST
+ &plt_xen_timer,
+#endif
&plt_hpet, &plt_pmtimer, &plt_pit
};
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v2 31/62] x86: xen pv clock time source
2018-01-12 11:28 ` [PATCH v2 31/62] x86: xen pv clock time source Wei Liu
@ 2018-01-12 20:45 ` Joao Martins
2018-01-16 12:22 ` Wei Liu
0 siblings, 1 reply; 66+ messages in thread
From: Joao Martins @ 2018-01-12 20:45 UTC (permalink / raw)
To: Wei Liu; +Cc: Xen Devel
On 01/12/2018 11:28 AM, Wei Liu wrote:
> It is a variant of TSC clock source.
>
> Signed-off-by: Wei Liu <wei.liu2@citrix.com>
> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
> ---
> Changes since v1:
> - Use the mapped vcpu_info.
> ---
> xen/arch/x86/time.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 89 insertions(+)
>
> diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
> index 3b654d7b7d..4cbd068d17 100644
> --- a/xen/arch/x86/time.c
> +++ b/xen/arch/x86/time.c
> @@ -29,6 +29,7 @@
> #include <asm/mpspec.h>
> #include <asm/processor.h>
> #include <asm/fixmap.h>
> +#include <asm/guest.h>
> #include <asm/mc146818rtc.h>
> #include <asm/div64.h>
> #include <asm/acpi.h>
> @@ -525,6 +526,91 @@ static struct platform_timesource __initdata plt_tsc =
> .init = init_tsc,
> };
>
> +#ifdef CONFIG_XEN_GUEST
> +/************************************************************
> + * PLATFORM TIMER 5: XEN PV CLOCK SOURCE
> + *
> + * Xen clock source is a variant of TSC source.
> + */
> +
> +static uint64_t xen_timer_cpu_frequency(void)
> +{
> + struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
> + uint64_t freq;
> +
> + freq = 1000000000ULL << 32;
> + do_div(freq, info->tsc_to_system_mul);
> + if ( info->tsc_shift < 0 )
> + freq <<= -info->tsc_shift;
> + else
> + freq >>= info->tsc_shift;
> +
> + return freq;
> +}
> +
> +static int64_t __init init_xen_timer(struct platform_timesource *pts)
> +{
> + if ( !xen_guest )
> + return 0;
> +
> + pts->frequency = xen_timer_cpu_frequency();
> +
> + return pts->frequency;
> +}
> +
Guests that don't have TSC_RELIABLE set in verify_tsc_reliability() you end up
evaluating this condition to true:
if ( boot_cpu_has(X86_FEATURE_CONSTANT_TSC) &&
!boot_cpu_has(X86_FEATURE_TSC_RELIABLE) )
time_calibration_rendezvous_fn = time_calibration_tsc_rendezvous;
And that's possibly not necessary to keep sync the TSC in sync with pvclock?
Meaning the default rendezvous function (time_calibration_std_rendezvous) might
be enough.
> +static always_inline uint64_t read_cycle(const struct vcpu_time_info *info,
> + uint64_t tsc)
> +{
> + uint64_t delta = tsc - info->tsc_timestamp;
> + struct time_scale ts = {
> + .shift = info->tsc_shift,
> + .mul_frac = info->tsc_to_system_mul,
> + };
> + uint64_t offset = scale_delta(delta, &ts);
> +
> + return info->system_time + offset;
> +}
> +
> +static uint64_t read_xen_timer(void)
> +{
> + struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
> + uint32_t version;
> + uint64_t ret;
> + uint64_t last;
> + static uint64_t last_value;
> +
> + do {
> + version = info->version & ~1;
> + /* Make sure version is read before the data */
> + smp_rmb();
> +
> + ret = read_cycle(info, rdtsc_ordered());
> + /* Ignore fancy flags for now */
> +
Heh :)
When the dust settles around the security issues, perhaps this patch could help
(I can formally submit it after more proper testing, but should give an
expectation on the changes).
-- >8 --
Subject: x86/time: mark pvclock as stable if supported
diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
index c90524de9c..3aaa72b6bf 100644
--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -65,8 +65,12 @@ struct platform_timesource {
s64 (*init)(struct platform_timesource *);
void (*resume)(struct platform_timesource *);
int counter_bits;
+ unsigned int flags;
};
+/* Clocksource counter monotonically increases across CPUs */
+#define CLOCKSOURCE_F_STABLE (1 << 0)
+
static DEFINE_PER_CPU(struct cpu_time, cpu_time);
/* Calibrate all CPUs to platform timer every EPOCH. */
@@ -533,9 +537,8 @@ static struct platform_timesource __initdata plt_tsc =
* Xen clock source is a variant of TSC source.
*/
-static uint64_t xen_timer_cpu_frequency(void)
+static uint64_t xen_timer_cpu_frequency(struct vcpu_time_info *info)
{
- struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
uint64_t freq;
freq = 1000000000ULL << 32;
@@ -550,10 +553,20 @@ static uint64_t xen_timer_cpu_frequency(void)
static int64_t __init init_xen_timer(struct platform_timesource *pts)
{
+ struct vcpu_time_info *info;
+
if ( !xen_guest )
return 0;
- pts->frequency = xen_timer_cpu_frequency();
+ info = &this_cpu(vcpu_info)->time;
+
+ pts->frequency = xen_timer_cpu_frequency(info);
+
+ /* Xen pvclock monotonically increases across CPUs */
+ if ( info->flags & XEN_PVCLOCK_TSC_STABLE_BIT ) {
+ pts->flags |= CLOCKSOURCE_F_STABLE;
+ printk("PV CLOCK marked as stable.\n");
+ }
return pts->frequency;
}
@@ -585,12 +598,14 @@ static uint64_t read_xen_timer(void)
smp_rmb();
ret = read_cycle(info, rdtsc_ordered());
- /* Ignore fancy flags for now */
/* Make sure version is reread after the data */
smp_rmb();
} while ( unlikely(version != info->version) );
+ if ( info->flags & XEN_PVCLOCK_TSC_STABLE_BIT )
+ return ret;
+
/* Maintain a monotonic global value */
do {
last = read_atomic(&last_value);
@@ -632,6 +647,8 @@ static u64 plt_stamp64; /* 64-bit platform counter
stamp */
static u64 plt_stamp; /* hardware-width platform counter stamp */
static struct timer plt_overflow_timer;
+static bool clocksource_is_stable(void);
+
static s_time_t __read_platform_stime(u64 platform_time)
{
u64 diff = platform_time - platform_timer_stamp;
@@ -1077,7 +1094,7 @@ static void __update_vcpu_system_time(struct vcpu *v, int
force)
* pvclock read to check whether they can resort solely on this tuple
* or if it further requires monotonicity checks with other vcpus.
*/
- if ( clocksource_is_tsc() )
+ if ( clocksource_is_stable() )
_u.flags |= XEN_PVCLOCK_TSC_STABLE_BIT;
if ( is_hvm_domain(d) )
@@ -1564,7 +1581,7 @@ static void time_calibration(void *unused)
.semaphore = ATOMIC_INIT(0)
};
- if ( clocksource_is_tsc() )
+ if ( clocksource_is_stable() )
{
local_irq_disable();
r.master_stime = read_platform_stime(&r.master_tsc_stamp);
@@ -1740,6 +1757,9 @@ static int __init verify_tsc_reliability(void)
*/
on_selected_cpus(&cpu_online_map, reset_percpu_time, NULL, 1);
+ /* TSC clocksource monotonically increases across CPUs. */
+ plt_tsc.flags |= CLOCKSOURCE_F_STABLE;
+
/*
* We won't do CPU Hotplug and TSC clocksource is being used which
* means we have a reliable TSC, plus we don't sync with any other
@@ -1770,7 +1790,12 @@ static int __init verify_tsc_reliability(void)
*/
if ( boot_cpu_has(X86_FEATURE_CONSTANT_TSC) &&
!boot_cpu_has(X86_FEATURE_TSC_RELIABLE) )
- time_calibration_rendezvous_fn = time_calibration_tsc_rendezvous;
+ {
+ if ( clocksource_is_stable() )
+ time_calibration_rendezvous_fn = time_calibration_nop_rendezvous;
+ else if ( plt_src.read_counter != read_xen_timer )
+ time_calibration_rendezvous_fn = time_calibration_tsc_rendezvous;
+ }
return 0;
}
@@ -2057,6 +2082,23 @@ bool clocksource_is_tsc(void)
return plt_src.read_counter == read_tsc;
}
+static bool clocksource_is_stable(void)
+{
+ struct vcpu_time_info *info;
+
+ if ( !(plt_src.flags & CLOCKSOURCE_F_STABLE) )
+ return false;
+
+ if ( plt_src.read_counter == read_tsc )
+ return true;
+
+ info = &this_cpu(vcpu_info)->time;
+ if ( !(info->flags & XEN_PVCLOCK_TSC_STABLE_BIT) )
+ return false;
+
+ return true;
+}
+
int host_tsc_is_safe(void)
{
return boot_cpu_has(X86_FEATURE_TSC_RELIABLE);
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* Re: [PATCH v2 31/62] x86: xen pv clock time source
2018-01-12 20:45 ` Joao Martins
@ 2018-01-16 12:22 ` Wei Liu
0 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-16 12:22 UTC (permalink / raw)
To: Joao Martins; +Cc: Xen Devel, Wei Liu
On Fri, Jan 12, 2018 at 08:45:17PM +0000, Joao Martins wrote:
> On 01/12/2018 11:28 AM, Wei Liu wrote:
> > It is a variant of TSC clock source.
> >
> > Signed-off-by: Wei Liu <wei.liu2@citrix.com>
> > Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> > Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
> > ---
> > Changes since v1:
> > - Use the mapped vcpu_info.
> > ---
> > xen/arch/x86/time.c | 89 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 89 insertions(+)
> >
> > diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
> > index 3b654d7b7d..4cbd068d17 100644
> > --- a/xen/arch/x86/time.c
> > +++ b/xen/arch/x86/time.c
> > @@ -29,6 +29,7 @@
> > #include <asm/mpspec.h>
> > #include <asm/processor.h>
> > #include <asm/fixmap.h>
> > +#include <asm/guest.h>
> > #include <asm/mc146818rtc.h>
> > #include <asm/div64.h>
> > #include <asm/acpi.h>
> > @@ -525,6 +526,91 @@ static struct platform_timesource __initdata plt_tsc =
> > .init = init_tsc,
> > };
> >
> > +#ifdef CONFIG_XEN_GUEST
> > +/************************************************************
> > + * PLATFORM TIMER 5: XEN PV CLOCK SOURCE
> > + *
> > + * Xen clock source is a variant of TSC source.
> > + */
> > +
> > +static uint64_t xen_timer_cpu_frequency(void)
> > +{
> > + struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
> > + uint64_t freq;
> > +
> > + freq = 1000000000ULL << 32;
> > + do_div(freq, info->tsc_to_system_mul);
> > + if ( info->tsc_shift < 0 )
> > + freq <<= -info->tsc_shift;
> > + else
> > + freq >>= info->tsc_shift;
> > +
> > + return freq;
> > +}
> > +
> > +static int64_t __init init_xen_timer(struct platform_timesource *pts)
> > +{
> > + if ( !xen_guest )
> > + return 0;
> > +
> > + pts->frequency = xen_timer_cpu_frequency();
> > +
> > + return pts->frequency;
> > +}
> > +
> Guests that don't have TSC_RELIABLE set in verify_tsc_reliability() you end up
> evaluating this condition to true:
>
> if ( boot_cpu_has(X86_FEATURE_CONSTANT_TSC) &&
> !boot_cpu_has(X86_FEATURE_TSC_RELIABLE) )
> time_calibration_rendezvous_fn = time_calibration_tsc_rendezvous;
>
> And that's possibly not necessary to keep sync the TSC in sync with pvclock?
>
> Meaning the default rendezvous function (time_calibration_std_rendezvous) might
> be enough.
Yes, it is possible that the default function is enough. But we haven't
tested that yet.
>
> > +static always_inline uint64_t read_cycle(const struct vcpu_time_info *info,
> > + uint64_t tsc)
> > +{
> > + uint64_t delta = tsc - info->tsc_timestamp;
> > + struct time_scale ts = {
> > + .shift = info->tsc_shift,
> > + .mul_frac = info->tsc_to_system_mul,
> > + };
> > + uint64_t offset = scale_delta(delta, &ts);
> > +
> > + return info->system_time + offset;
> > +}
> > +
> > +static uint64_t read_xen_timer(void)
> > +{
> > + struct vcpu_time_info *info = &this_cpu(vcpu_info)->time;
> > + uint32_t version;
> > + uint64_t ret;
> > + uint64_t last;
> > + static uint64_t last_value;
> > +
> > + do {
> > + version = info->version & ~1;
> > + /* Make sure version is read before the data */
> > + smp_rmb();
> > +
> > + ret = read_cycle(info, rdtsc_ordered());
> > + /* Ignore fancy flags for now */
> > +
>
> Heh :)
>
> When the dust settles around the security issues, perhaps this patch could help
> (I can formally submit it after more proper testing, but should give an
> expectation on the changes).
>
I skimmed through it. It looks like it is mostly an optimisation
for this patch. That is, properly detecting the stable bit and using it
where appropriate. We can definitely improve in that regard. But as you
said, let's handle the security issues at hand first.
Wei.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply [flat|nested] 66+ messages in thread
* [PATCH v2 32/62] x86: APIC timer calibration when running as a guest
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (30 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 31/62] x86: xen pv clock time source Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 33/62] x86: read wallclock from Xen when running in pvh mode Wei Liu
` (29 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
The timer calibration currently depends on PIT. Introduce a variant
to wait for a tick's worth of time to elapse when running as a PVH
guest.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
xen/arch/x86/apic.c | 38 ++++++++++++++++++++++++++++++--------
1 file changed, 30 insertions(+), 8 deletions(-)
diff --git a/xen/arch/x86/apic.c b/xen/arch/x86/apic.c
index ed59440c45..5039173827 100644
--- a/xen/arch/x86/apic.c
+++ b/xen/arch/x86/apic.c
@@ -36,6 +36,8 @@
#include <mach_apic.h>
#include <io_ports.h>
#include <xen/kexec.h>
+#include <asm/guest.h>
+#include <asm/time.h>
static bool __read_mostly tdt_enabled;
static bool __initdata tdt_enable = true;
@@ -1091,6 +1093,20 @@ static void setup_APIC_timer(void)
local_irq_restore(flags);
}
+static void wait_tick_pvh(void)
+{
+ u64 lapse_ns = 1000000000ULL / HZ;
+ s_time_t start, curr_time;
+
+ start = NOW();
+
+ /* Won't wrap around */
+ do {
+ cpu_relax();
+ curr_time = NOW();
+ } while ( curr_time - start < lapse_ns );
+}
+
/*
* In this function we calibrate APIC bus clocks to the external
* timer. Unfortunately we cannot use jiffies and the timer irq
@@ -1123,12 +1139,15 @@ static int __init calibrate_APIC_clock(void)
*/
__setup_APIC_LVTT(1000000000);
- /*
- * The timer chip counts down to zero. Let's wait
- * for a wraparound to start exact measurement:
- * (the current tick might have been already half done)
- */
- wait_8254_wraparound();
+ if ( !xen_guest )
+ /*
+ * The timer chip counts down to zero. Let's wait
+ * for a wraparound to start exact measurement:
+ * (the current tick might have been already half done)
+ */
+ wait_8254_wraparound();
+ else
+ wait_tick_pvh();
/*
* We wrapped around just now. Let's start:
@@ -1137,10 +1156,13 @@ static int __init calibrate_APIC_clock(void)
tt1 = apic_read(APIC_TMCCT);
/*
- * Let's wait LOOPS wraprounds:
+ * Let's wait LOOPS ticks:
*/
for (i = 0; i < LOOPS; i++)
- wait_8254_wraparound();
+ if ( !xen_guest )
+ wait_8254_wraparound();
+ else
+ wait_tick_pvh();
tt2 = apic_read(APIC_TMCCT);
t2 = rdtsc_ordered();
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 33/62] x86: read wallclock from Xen when running in pvh mode
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (31 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 32/62] x86: APIC timer calibration when running as a guest Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 34/62] x86: don't swallow the first command line item in guest mode Wei Liu
` (28 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
---
xen/arch/x86/time.c | 32 ++++++++++++++++++++++++++++----
1 file changed, 28 insertions(+), 4 deletions(-)
diff --git a/xen/arch/x86/time.c b/xen/arch/x86/time.c
index 4cbd068d17..c90524de9c 100644
--- a/xen/arch/x86/time.c
+++ b/xen/arch/x86/time.c
@@ -964,6 +964,30 @@ static unsigned long get_cmos_time(void)
return mktime(rtc.year, rtc.mon, rtc.day, rtc.hour, rtc.min, rtc.sec);
}
+static unsigned long get_wallclock_time(void)
+{
+#ifdef CONFIG_XEN_GUEST
+ if ( xen_guest )
+ {
+ struct shared_info *sh_info = XEN_shared_info;
+ uint32_t wc_version;
+ uint64_t wc_sec;
+
+ do {
+ wc_version = sh_info->wc_version & ~1;
+ smp_rmb();
+
+ wc_sec = sh_info->wc_sec;
+ smp_rmb();
+ } while ( wc_version != sh_info->wc_version );
+
+ return wc_sec + read_xen_timer() / 1000000000;
+ }
+#endif
+
+ return get_cmos_time();
+}
+
/***************************************************************************
* System Time
***************************************************************************/
@@ -1759,8 +1783,8 @@ int __init init_xen_time(void)
open_softirq(TIME_CALIBRATE_SOFTIRQ, local_time_calibration);
- /* NB. get_cmos_time() can take over one second to execute. */
- do_settime(get_cmos_time(), 0, NOW());
+ /* NB. get_wallclock_time() can take over one second to execute. */
+ do_settime(get_wallclock_time(), 0, NOW());
/* Finish platform timer initialization. */
try_platform_timer_tail(false);
@@ -1870,7 +1894,7 @@ int time_suspend(void)
{
if ( smp_processor_id() == 0 )
{
- cmos_utc_offset = -get_cmos_time();
+ cmos_utc_offset = -get_wallclock_time();
cmos_utc_offset += get_sec();
kill_timer(&calibration_timer);
@@ -1897,7 +1921,7 @@ int time_resume(void)
set_timer(&calibration_timer, NOW() + EPOCH);
- do_settime(get_cmos_time() + cmos_utc_offset, 0, NOW());
+ do_settime(get_wallclock_time() + cmos_utc_offset, 0, NOW());
update_vcpu_system_time(current);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 34/62] x86: don't swallow the first command line item in guest mode
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (32 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 33/62] x86: read wallclock from Xen when running in pvh mode Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 35/62] x86/guest: setup event channel upcall vector Wei Liu
` (27 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/arch/x86/setup.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 32ec741591..0c9b55ff71 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -632,8 +632,8 @@ static char * __init cmdline_cook(char *p, const char *loader_name)
while ( *p == ' ' )
p++;
- /* GRUB2 does not include image name as first item on command line. */
- if ( loader_is_grub2(loader_name) )
+ /* GRUB2 and PVH don't not include image name as first item on command line. */
+ if ( xen_guest || loader_is_grub2(loader_name) )
return p;
/* Strip image name plus whitespace. */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 35/62] x86/guest: setup event channel upcall vector
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (33 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 34/62] x86: don't swallow the first command line item in guest mode Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 36/62] x86/guest: add PV console code Wei Liu
` (26 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
And a dummy event channel upcall handler.
Note that with the current code the underlying Xen (L0) must support
HVMOP_set_evtchn_upcall_vector or else event channel setup is going to
fail. This limitation can be lifted by implementing more event channel
interrupt injection methods as a backup.
Register callback_irq to trick toolstack to think the domain is
enlightened.
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/arch/x86/guest/xen.c | 41 +++++++++++++++++++++++++++++++++++
xen/include/asm-x86/guest/hypercall.h | 17 +++++++++++++++
2 files changed, 58 insertions(+)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 60626ec21c..59871170c8 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -24,6 +24,7 @@
#include <xen/rangeset.h>
#include <xen/types.h>
+#include <asm/apic.h>
#include <asm/e820.h>
#include <asm/guest.h>
#include <asm/msr.h>
@@ -186,6 +187,43 @@ static void __init init_memmap(void)
}
}
+static void xen_evtchn_upcall(struct cpu_user_regs *regs)
+{
+ struct vcpu_info *vcpu_info = this_cpu(vcpu_info);
+
+ vcpu_info->evtchn_upcall_pending = 0;
+ write_atomic(&vcpu_info->evtchn_pending_sel, 0);
+
+ ack_APIC_irq();
+}
+
+static void init_evtchn(void)
+{
+ static uint8_t evtchn_upcall_vector;
+ int rc;
+
+ if ( !evtchn_upcall_vector )
+ alloc_direct_apic_vector(&evtchn_upcall_vector, xen_evtchn_upcall);
+
+ ASSERT(evtchn_upcall_vector);
+
+ rc = xen_hypercall_set_evtchn_upcall_vector(this_cpu(vcpu_id),
+ evtchn_upcall_vector);
+ if ( rc )
+ panic("Unable to set evtchn upcall vector: %d", rc);
+
+ /* Trick toolstack to think we are enlightened */
+ {
+ struct xen_hvm_param a = {
+ .domid = DOMID_SELF,
+ .index = HVM_PARAM_CALLBACK_IRQ,
+ .value = 1,
+ };
+
+ BUG_ON(xen_hypercall_hvm_op(HVMOP_set_param, &a));
+ }
+}
+
void __init hypervisor_setup(void)
{
init_memmap();
@@ -210,12 +248,15 @@ void __init hypervisor_setup(void)
"unable to map vCPU info, limiting vCPUs to: %u\n",
XEN_LEGACY_MAX_VCPUS);
}
+
+ init_evtchn();
}
void hypervisor_ap_setup(void)
{
set_vcpu_id();
map_vcpuinfo();
+ init_evtchn();
}
int hypervisor_alloc_unused_page(mfn_t *mfn)
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index dbc57a566e..b36a1cc189 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -25,6 +25,7 @@
#include <public/xen.h>
#include <public/sched.h>
+#include <public/hvm/hvm_op.h>
#include <public/vcpu.h>
@@ -104,6 +105,11 @@ static inline int xen_hypercall_vcpu_op(unsigned int cmd, unsigned int vcpu,
return _hypercall64_3(long, __HYPERVISOR_vcpu_op, cmd, vcpu, arg);
}
+static inline long xen_hypercall_hvm_op(unsigned int op, void *arg)
+{
+ return _hypercall64_2(long, __HYPERVISOR_hvm_op, op, arg);
+}
+
/*
* Higher level hypercall helpers
*/
@@ -120,6 +126,17 @@ static inline long xen_hypercall_shutdown(unsigned int reason)
return xen_hypercall_sched_op(SCHEDOP_shutdown, &s);
}
+static inline long xen_hypercall_set_evtchn_upcall_vector(
+ unsigned int cpu, unsigned int vector)
+{
+ struct xen_hvm_evtchn_upcall_vector a = {
+ .vcpu = cpu,
+ .vector = vector,
+ };
+
+ return xen_hypercall_hvm_op(HVMOP_set_evtchn_upcall_vector, &a);
+}
+
#else /* CONFIG_XEN_GUEST */
#include <public/sched.h>
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 36/62] x86/guest: add PV console code
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (34 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 35/62] x86/guest: setup event channel upcall vector Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 37/62] x86/guest: use PV console for Xen/Dom0 I/O Wei Liu
` (25 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
xen/drivers/char/Makefile | 1 +
xen/drivers/char/xen_pv_console.c | 205 ++++++++++++++++++++++++++++++++++
xen/include/asm-x86/fixmap.h | 1 +
xen/include/asm-x86/guest/hypercall.h | 33 ++++++
xen/include/xen/pv_console.h | 32 ++++++
5 files changed, 272 insertions(+)
create mode 100644 xen/drivers/char/xen_pv_console.c
create mode 100644 xen/include/xen/pv_console.h
diff --git a/xen/drivers/char/Makefile b/xen/drivers/char/Makefile
index aa169d7961..9d48d0f2dc 100644
--- a/xen/drivers/char/Makefile
+++ b/xen/drivers/char/Makefile
@@ -8,3 +8,4 @@ obj-$(CONFIG_HAS_SCIF) += scif-uart.o
obj-$(CONFIG_HAS_EHCI) += ehci-dbgp.o
obj-$(CONFIG_ARM) += arm-uart.o
obj-y += serial.o
+obj-$(CONFIG_XEN_GUEST) += xen_pv_console.o
diff --git a/xen/drivers/char/xen_pv_console.c b/xen/drivers/char/xen_pv_console.c
new file mode 100644
index 0000000000..f5aca4c69e
--- /dev/null
+++ b/xen/drivers/char/xen_pv_console.c
@@ -0,0 +1,205 @@
+/******************************************************************************
+ * drivers/char/xen_pv_console.c
+ *
+ * A frontend driver for Xen's PV console.
+ * Can be used when Xen is running on top of Xen in pv-in-pvh mode.
+ * (Linux's name for this is hvc console)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#include <xen/lib.h>
+#include <xen/hypercall.h>
+#include <xen/pv_console.h>
+
+#include <asm/fixmap.h>
+#include <asm/guest.h>
+
+#include <public/io/console.h>
+
+static struct xencons_interface *cons_ring;
+static evtchn_port_t cons_evtchn;
+static serial_rx_fn cons_rx_handler;
+static DEFINE_SPINLOCK(tx_lock);
+
+void __init pv_console_init(void)
+{
+ long r;
+ uint64_t raw_pfn = 0, raw_evtchn = 0;
+
+ if ( !xen_guest )
+ {
+ printk("PV console init failed: xen_guest mode is not active!\n");
+ return;
+ }
+
+ r = xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_PFN, &raw_pfn);
+ if ( r < 0 )
+ goto error;
+
+ r = xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_EVTCHN, &raw_evtchn);
+ if ( r < 0 )
+ goto error;
+
+ set_fixmap(FIX_PV_CONSOLE, raw_pfn << PAGE_SHIFT);
+ cons_ring = (struct xencons_interface *)fix_to_virt(FIX_PV_CONSOLE);
+ cons_evtchn = raw_evtchn;
+
+ printk("Initialised PV console at 0x%p with pfn %#lx and evtchn %#x\n",
+ cons_ring, raw_pfn, cons_evtchn);
+ return;
+
+ error:
+ printk("Couldn't initialise PV console\n");
+}
+
+void __init pv_console_set_rx_handler(serial_rx_fn fn)
+{
+ cons_rx_handler = fn;
+}
+
+void __init pv_console_init_postirq(void)
+{
+ if ( !cons_ring )
+ return;
+
+ xen_hypercall_evtchn_unmask(cons_evtchn);
+}
+
+static void notify_daemon(void)
+{
+ xen_hypercall_evtchn_send(cons_evtchn);
+}
+
+size_t pv_console_rx(struct cpu_user_regs *regs)
+{
+ char c;
+ XENCONS_RING_IDX cons, prod;
+ size_t recv = 0;
+
+ if ( !cons_ring )
+ return 0;
+
+ /* TODO: move this somewhere */
+ if ( !test_bit(cons_evtchn, XEN_shared_info->evtchn_pending) )
+ return 0;
+
+ prod = ACCESS_ONCE(cons_ring->in_prod);
+ cons = cons_ring->in_cons;
+
+ /*
+ * Latch pointers before accessing the ring. Included compiler barrier also
+ * ensures that pointers are really read only once into local variables.
+ */
+ smp_rmb();
+
+ ASSERT((prod - cons) <= sizeof(cons_ring->in));
+
+ while ( cons != prod )
+ {
+ c = cons_ring->in[MASK_XENCONS_IDX(cons++, cons_ring->in)];
+ if ( cons_rx_handler )
+ cons_rx_handler(c, regs);
+ recv++;
+ }
+
+ /* No need for a mem barrier because every character was already consumed */
+ barrier();
+ ACCESS_ONCE(cons_ring->in_cons) = cons;
+ notify_daemon();
+
+ clear_bit(cons_evtchn, XEN_shared_info->evtchn_pending);
+
+ return recv;
+}
+
+static size_t pv_ring_puts(const char *buf)
+{
+ XENCONS_RING_IDX cons, prod;
+ size_t sent = 0, avail;
+ bool put_r = false;
+
+ while ( buf[sent] != '\0' || put_r )
+ {
+ cons = ACCESS_ONCE(cons_ring->out_cons);
+ prod = cons_ring->out_prod;
+
+ /*
+ * Latch pointers before accessing the ring. Included compiler barrier
+ * ensures that pointers are really read only once into local variables.
+ */
+ smp_rmb();
+
+ ASSERT((prod - cons) <= sizeof(cons_ring->out));
+ avail = sizeof(cons_ring->out) - (prod - cons);
+
+ if ( avail == 0 )
+ {
+ /* Wait for xenconsoled to consume our output */
+ xen_hypercall_sched_op(SCHEDOP_yield, NULL);
+ continue;
+ }
+
+ while ( avail && (buf[sent] != '\0' || put_r) )
+ {
+ if ( put_r )
+ {
+ cons_ring->out[MASK_XENCONS_IDX(prod++, cons_ring->out)] = '\r';
+ put_r = false;
+ }
+ else
+ {
+ cons_ring->out[MASK_XENCONS_IDX(prod++, cons_ring->out)] =
+ buf[sent];
+
+ /* Send '\r' for every '\n' */
+ if ( buf[sent] == '\n' )
+ put_r = true;
+ sent++;
+ }
+ avail--;
+ }
+
+ /* Write to the ring before updating the pointer */
+ smp_wmb();
+ ACCESS_ONCE(cons_ring->out_prod) = prod;
+ notify_daemon();
+ }
+
+ return sent;
+}
+
+void pv_console_puts(const char *buf)
+{
+ unsigned long flags;
+
+ if ( !cons_ring )
+ return;
+
+ spin_lock_irqsave(&tx_lock, flags);
+ pv_ring_puts(buf);
+ spin_unlock_irqrestore(&tx_lock, flags);
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/asm-x86/fixmap.h b/xen/include/asm-x86/fixmap.h
index ded4ddf21b..16ccaa2c77 100644
--- a/xen/include/asm-x86/fixmap.h
+++ b/xen/include/asm-x86/fixmap.h
@@ -46,6 +46,7 @@ enum fixed_addresses {
FIX_COM_END,
FIX_EHCI_DBGP,
#ifdef CONFIG_XEN_GUEST
+ FIX_PV_CONSOLE,
FIX_XEN_SHARED_INFO,
#endif /* CONFIG_XEN_GUEST */
/* Everything else should go further down. */
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index b36a1cc189..81a955d479 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -105,6 +105,11 @@ static inline int xen_hypercall_vcpu_op(unsigned int cmd, unsigned int vcpu,
return _hypercall64_3(long, __HYPERVISOR_vcpu_op, cmd, vcpu, arg);
}
+static inline long xen_hypercall_event_channel_op(unsigned int cmd, void *arg)
+{
+ return _hypercall64_2(long, __HYPERVISOR_event_channel_op, cmd, arg);
+}
+
static inline long xen_hypercall_hvm_op(unsigned int op, void *arg)
{
return _hypercall64_2(long, __HYPERVISOR_hvm_op, op, arg);
@@ -126,6 +131,34 @@ static inline long xen_hypercall_shutdown(unsigned int reason)
return xen_hypercall_sched_op(SCHEDOP_shutdown, &s);
}
+static inline long xen_hypercall_evtchn_send(evtchn_port_t port)
+{
+ struct evtchn_send send = { .port = port };
+
+ return xen_hypercall_event_channel_op(EVTCHNOP_send, &send);
+}
+
+static inline long xen_hypercall_evtchn_unmask(evtchn_port_t port)
+{
+ struct evtchn_unmask unmask = { .port = port };
+
+ return xen_hypercall_event_channel_op(EVTCHNOP_unmask, &unmask);
+}
+
+static inline long xen_hypercall_hvm_get_param(uint32_t index, uint64_t *value)
+{
+ struct xen_hvm_param xhv = {
+ .domid = DOMID_SELF,
+ .index = index,
+ };
+ long ret = xen_hypercall_hvm_op(HVMOP_get_param, &xhv);
+
+ if ( ret == 0 )
+ *value = xhv.value;
+
+ return ret;
+}
+
static inline long xen_hypercall_set_evtchn_upcall_vector(
unsigned int cpu, unsigned int vector)
{
diff --git a/xen/include/xen/pv_console.h b/xen/include/xen/pv_console.h
new file mode 100644
index 0000000000..e578b56620
--- /dev/null
+++ b/xen/include/xen/pv_console.h
@@ -0,0 +1,32 @@
+#ifndef __XEN_PV_CONSOLE_H__
+#define __XEN_PV_CONSOLE_H__
+
+#include <xen/serial.h>
+
+#ifdef CONFIG_XEN_GUEST
+
+void pv_console_init(void);
+void pv_console_set_rx_handler(serial_rx_fn fn);
+void pv_console_init_postirq(void);
+void pv_console_puts(const char *buf);
+size_t pv_console_rx(struct cpu_user_regs *regs);
+
+#else
+
+static inline void pv_console_init(void) {}
+static inline void pv_console_set_rx_handler(serial_rx_fn fn) { }
+static inline void pv_console_init_postirq(void) { }
+static inline void pv_console_puts(const char *buf) { }
+static inline size_t pv_console_rx(struct cpu_user_regs *regs) { return 0; }
+
+#endif /* !CONFIG_XEN_GUEST */
+#endif /* __XEN_PV_CONSOLE_H__ */
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 37/62] x86/guest: use PV console for Xen/Dom0 I/O
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (35 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 36/62] x86/guest: add PV console code Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 38/62] x86/shim: Kconfig and command line options Wei Liu
` (24 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
docs/misc/xen-command-line.markdown | 5 ++++-
xen/arch/x86/guest/xen.c | 3 +++
xen/drivers/char/console.c | 16 ++++++++++++++++
3 files changed, 23 insertions(+), 1 deletion(-)
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index e5979bceee..da006dd4f7 100644
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -365,7 +365,7 @@ The following are examples of correct specifications:
Specify the size of the console ring buffer.
### console
-> `= List of [ vga | com1[H,L] | com2[H,L] | dbgp | none ]`
+> `= List of [ vga | com1[H,L] | com2[H,L] | pv | dbgp | none ]`
> Default: `console=com1,vga`
@@ -381,6 +381,9 @@ the converse; transmitted and received characters will have their MSB
cleared. This allows a single port to be shared by two subsystems
(e.g. console and debugger).
+`pv` indicates that Xen should use Xen's PV console. This option is
+only available when used together with `pv-in-pvh`.
+
`dbgp` indicates that Xen should use a USB debug port.
`none` indicates that Xen should not use a console. This option only
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 59871170c8..d4968b47aa 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -23,6 +23,7 @@
#include <xen/pfn.h>
#include <xen/rangeset.h>
#include <xen/types.h>
+#include <xen/pv_console.h>
#include <asm/apic.h>
#include <asm/e820.h>
@@ -194,6 +195,8 @@ static void xen_evtchn_upcall(struct cpu_user_regs *regs)
vcpu_info->evtchn_upcall_pending = 0;
write_atomic(&vcpu_info->evtchn_pending_sel, 0);
+ pv_console_rx(regs);
+
ack_APIC_irq();
}
diff --git a/xen/drivers/char/console.c b/xen/drivers/char/console.c
index d05ebf9f70..8acd358395 100644
--- a/xen/drivers/char/console.c
+++ b/xen/drivers/char/console.c
@@ -32,6 +32,7 @@
#include <xen/warning.h>
#ifdef CONFIG_X86
+#include <xen/pv_console.h>
#include <asm/guest.h>
#endif
@@ -344,6 +345,11 @@ static void sercon_puts(const char *s)
(*serial_steal_fn)(s);
else
serial_puts(sercon_handle, s);
+
+#ifdef CONFIG_X86
+ /* Copy all serial output into PV console */
+ pv_console_puts(s);
+#endif
}
static void dump_console_ring_key(unsigned char key)
@@ -805,6 +811,8 @@ void __init console_init_preirq(void)
if ( !strncmp(p, "vga", 3) )
video_init();
#ifdef CONFIG_X86
+ else if ( !strncmp(p, "pv", 2) )
+ pv_console_init();
else if ( !strncmp(p, "xen", 3) )
opt_console_xen = true;
#endif
@@ -828,6 +836,10 @@ void __init console_init_preirq(void)
serial_set_rx_handler(sercon_handle, serial_rx);
+#ifdef CONFIG_X86
+ pv_console_set_rx_handler(serial_rx);
+#endif
+
/* HELLO WORLD --- start-of-day banner text. */
spin_lock(&console_lock);
__putstr(xen_banner());
@@ -880,6 +892,10 @@ void __init console_init_postirq(void)
{
serial_init_postirq();
+#ifdef CONFIG_X86
+ pv_console_init_postirq();
+#endif
+
if ( conring != _conring )
return;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 38/62] x86/shim: Kconfig and command line options
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (36 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 37/62] x86/guest: use PV console for Xen/Dom0 I/O Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 39/62] tools/firmware: Build and install xen-shim Wei Liu
` (23 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
docs/misc/xen-command-line.markdown | 11 ++++++++++
xen/arch/x86/Kconfig | 22 +++++++++++++++++++
xen/arch/x86/pv/Makefile | 1 +
xen/arch/x86/pv/shim.c | 39 ++++++++++++++++++++++++++++++++++
xen/include/asm-x86/guest.h | 1 +
xen/include/asm-x86/pv/shim.h | 42 +++++++++++++++++++++++++++++++++++++
6 files changed, 116 insertions(+)
create mode 100644 xen/arch/x86/pv/shim.c
create mode 100644 xen/include/asm-x86/pv/shim.h
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index da006dd4f7..3a1a9c1fba 100644
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -1445,6 +1445,17 @@ do; there may be other custom operating systems which do. If you're
certain you don't plan on having PV guests which use this feature,
turning it off can reduce the attack surface.
+### pv-shim (x86)
+> `= <boolean>`
+
+> Default: `false`
+
+This option is intended for use by a toolstack, when choosing to run a PV
+guest compatibly inside an HVM container.
+
+In this mode, the kernel and initrd passed as modules to the hypervisor are
+constructed into a plain unprivileged PV domain.
+
### rcu-idle-timer-period-ms
> `= <integer>`
diff --git a/xen/arch/x86/Kconfig b/xen/arch/x86/Kconfig
index c0b0bcdcb3..4953533f16 100644
--- a/xen/arch/x86/Kconfig
+++ b/xen/arch/x86/Kconfig
@@ -133,6 +133,28 @@ config PVH_GUEST
---help---
Support booting using the PVH ABI.
+ If unsure, say N.
+
+config PV_SHIM
+ def_bool n
+ prompt "PV Shim"
+ depends on PV && XEN_GUEST
+ ---help---
+ Build Xen with a mode which acts as a shim to allow PV guest to run
+ in an HVM/PVH container. This mode can only be enabled with command
+ line option.
+
+ If unsure, say N.
+
+config PV_SHIM_EXCLUSIVE
+ def_bool n
+ prompt "PV Shim Exclusive"
+ depends on PV_SHIM
+ ---help---
+ Build Xen in a way which unconditionally assumes PV_SHIM mode. This
+ option is only intended for use when building a dedicated PV Shim
+ firmware, and will not function correctly in other scenarios.
+
If unsure, say N.
endmenu
diff --git a/xen/arch/x86/pv/Makefile b/xen/arch/x86/pv/Makefile
index bac2792aa2..65bca04175 100644
--- a/xen/arch/x86/pv/Makefile
+++ b/xen/arch/x86/pv/Makefile
@@ -11,6 +11,7 @@ obj-y += iret.o
obj-y += misc-hypercalls.o
obj-y += mm.o
obj-y += ro-page-fault.o
+obj-$(CONFIG_PV_SHIM) += shim.o
obj-y += traps.o
obj-bin-y += dom0_build.init.o
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
new file mode 100644
index 0000000000..4d037355db
--- /dev/null
+++ b/xen/arch/x86/pv/shim.c
@@ -0,0 +1,39 @@
+/******************************************************************************
+ * arch/x86/pv/shim.c
+ *
+ * Functionaltiy for PV Shim mode
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+#include <xen/init.h>
+#include <xen/types.h>
+
+#include <asm/apic.h>
+
+#ifndef CONFIG_PV_SHIM_EXCLUSIVE
+bool pv_shim;
+boolean_param("pv-shim", pv_shim);
+#endif
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/asm-x86/guest.h b/xen/include/asm-x86/guest.h
index 5abdb8c433..a38c6b5b3f 100644
--- a/xen/include/asm-x86/guest.h
+++ b/xen/include/asm-x86/guest.h
@@ -22,6 +22,7 @@
#include <asm/guest/hypercall.h>
#include <asm/guest/pvh-boot.h>
#include <asm/guest/xen.h>
+#include <asm/pv/shim.h>
#endif /* __X86_GUEST_H__ */
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
new file mode 100644
index 0000000000..1468cfd498
--- /dev/null
+++ b/xen/include/asm-x86/pv/shim.h
@@ -0,0 +1,42 @@
+/******************************************************************************
+ * asm-x86/guest/shim.h
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms and conditions of the GNU General Public
+ * License, version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#ifndef __X86_PV_SHIM_H__
+#define __X86_PV_SHIM_H__
+
+#include <xen/types.h>
+
+#if defined(CONFIG_PV_SHIM_EXCLUSIVE)
+# define pv_shim 1
+#elif defined(CONFIG_PV_SHIM)
+extern bool pv_shim;
+#else
+# define pv_shim 0
+#endif /* CONFIG_PV_SHIM{,_EXCLUSIVE} */
+
+#endif /* __X86_PV_SHIM_H__ */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 39/62] tools/firmware: Build and install xen-shim
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (37 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 38/62] x86/shim: Kconfig and command line options Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 40/62] xen/x86: make VGA support selectable Wei Liu
` (22 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Andrew Cooper <andrew.cooper3@citrix.com>
Link a minimum set of files to build the shim. The linkfarm rune can
handle creation and deletion of files. Introduce build-shim and
install-shim targets in xen/Makefile.
We can do better by properly generate the dependency from the list of
files but that's an improvement for later.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2: Introduce a top-level build-shim target. Split the xen-shim build
with normal build.
---
.gitignore | 4 ++
tools/firmware/Makefile | 9 ++++
tools/firmware/xen-dir/Makefile | 59 ++++++++++++++++++++++++++
tools/firmware/xen-dir/shim.config | 87 ++++++++++++++++++++++++++++++++++++++
xen/Makefile | 16 +++++--
5 files changed, 172 insertions(+), 3 deletions(-)
create mode 100644 tools/firmware/xen-dir/Makefile
create mode 100644 tools/firmware/xen-dir/shim.config
diff --git a/.gitignore b/.gitignore
index 8da67daf31..f6cc61a701 100644
--- a/.gitignore
+++ b/.gitignore
@@ -155,6 +155,10 @@ tools/firmware/rombios/rombios[^/]*.s
tools/firmware/rombios/32bit/32bitbios_flat.h
tools/firmware/vgabios/vbetables-gen
tools/firmware/vgabios/vbetables.h
+tools/firmware/xen-dir/*.old
+tools/firmware/xen-dir/linkfarm.stamp*
+tools/firmware/xen-dir/xen-root
+tools/firmware/xen-dir/xen-shim
tools/flask/utils/flask-getenforce
tools/flask/utils/flask-get-bool
tools/flask/utils/flask-loadpolicy
diff --git a/tools/firmware/Makefile b/tools/firmware/Makefile
index 868b506920..9387cc0878 100644
--- a/tools/firmware/Makefile
+++ b/tools/firmware/Makefile
@@ -1,6 +1,8 @@
XEN_ROOT = $(CURDIR)/../..
include $(XEN_ROOT)/tools/Rules.mk
+CONFIG_PV_SHIM := y
+
# hvmloader is a 32-bit protected mode binary.
TARGET := hvmloader/hvmloader
INST_DIR := $(DESTDIR)$(XENFIRMWAREDIR)
@@ -11,6 +13,7 @@ SUBDIRS-$(CONFIG_SEABIOS) += seabios-dir
SUBDIRS-$(CONFIG_ROMBIOS) += rombios
SUBDIRS-$(CONFIG_ROMBIOS) += vgabios
SUBDIRS-$(CONFIG_ROMBIOS) += etherboot
+SUBDIRS-$(CONFIG_PV_SHIM) += xen-dir
SUBDIRS-y += hvmloader
LD32BIT-$(CONFIG_FreeBSD) := LD32BIT_FLAG=-melf_i386_fbsd
@@ -48,6 +51,9 @@ endif
ifeq ($(CONFIG_OVMF),y)
$(INSTALL_DATA) ovmf-dir/ovmf.bin $(INST_DIR)/ovmf.bin
endif
+ifeq ($(CONFIG_PV_SHIM),y)
+ $(INSTALL_DATA) xen-dir/xen-shim $(INST_DIR)/xen-shim
+endif
.PHONY: uninstall
uninstall:
@@ -58,6 +64,9 @@ endif
ifeq ($(CONFIG_OVMF),y)
rm -f $(INST_DIR)/ovmf.bin
endif
+ifeq ($(CONFIG_PV_SHIM),y)
+ rm -f $(INST_DIR)/xen-shim
+endif
.PHONY: clean
clean: subdirs-clean
diff --git a/tools/firmware/xen-dir/Makefile b/tools/firmware/xen-dir/Makefile
new file mode 100644
index 0000000000..adf6c31e8d
--- /dev/null
+++ b/tools/firmware/xen-dir/Makefile
@@ -0,0 +1,59 @@
+XEN_ROOT = $(CURDIR)/../../..
+
+all: xen-shim
+
+.PHONY: FORCE
+FORCE:
+
+D=xen-root
+
+# Minimun set of files / directories go get Xen to build
+LINK_DIRS=config xen
+LINK_FILES=Config.mk
+
+DEP_DIRS=$(foreach i, $(LINK_DIRS), $(XEN_ROOT)/$(i))
+DEP_FILES=$(foreach i, $(LINK_FILES), $(XEN_ROOT)/$(i))
+
+linkfarm.stamp: $(DEP_DIRS) $(DEP_FILES) FORCE
+ mkdir -p $(D)
+ set -e
+ rm -f linkfarm.stamp.tmp
+ $(foreach d, $(LINK_DIRS), \
+ (mkdir -p $(D)/$(d); \
+ cd $(D)/$(d); \
+ find $(XEN_ROOT)/$(d)/ -type d -printf "./%P\n" | xargs mkdir -p);)
+ $(foreach d, $(LINK_DIRS), \
+ (cd $(XEN_ROOT); \
+ find $(d) ! -type l -type f \
+ $(addprefix ! -path , '*.[oda1]' '*.d[12]')) \
+ >> linkfarm.stamp.tmp ; )
+ $(foreach f, $(LINK_FILES), \
+ echo $(f) >> linkfarm.stamp.tmp ;)
+ cmp -s linkfarm.stamp.tmp linkfarm.stamp && \
+ rm linkfarm.stamp.tmp || { \
+ mv linkfarm.stamp.tmp linkfarm.stamp; \
+ cat linkfarm.stamp | while read f; \
+ do rm -f "$(D)/$$f"; ln -s "$(XEN_ROOT)/$$f" "$(D)/$$f"; done \
+ }
+
+# Copy enough of the tree to build the shim hypervisor
+$(D): linkfarm.stamp
+ $(MAKE) -C $(D)/xen distclean
+
+.PHONY: shim-%config
+shim-%config: $(D) FORCE
+ $(MAKE) -C $(D)/xen $*config \
+ XEN_CONFIG_EXPERT=y \
+ KCONFIG_CONFIG=$(CURDIR)/shim.config
+
+xen-shim: $(D) shim-olddefconfig
+ $(MAKE) -C $(D)/xen install-shim \
+ XEN_CONFIG_EXPERT=y \
+ KCONFIG_CONFIG=$(CURDIR)/shim.config \
+ DESTDIR=$(CURDIR)
+
+.PHONY: distclean clean
+distclean clean:
+ rm -f xen-shim *.old
+ rm -rf $(D)
+ rm -f linkfarm.stamp*
diff --git a/tools/firmware/xen-dir/shim.config b/tools/firmware/xen-dir/shim.config
new file mode 100644
index 0000000000..227a12fb4c
--- /dev/null
+++ b/tools/firmware/xen-dir/shim.config
@@ -0,0 +1,87 @@
+#
+# Automatically generated file; DO NOT EDIT.
+# Xen/x86 4.11-unstable Configuration
+#
+CONFIG_X86_64=y
+CONFIG_X86=y
+CONFIG_ARCH_DEFCONFIG="arch/x86/configs/x86_64_defconfig"
+
+#
+# Architecture Features
+#
+CONFIG_NR_CPUS=32
+CONFIG_PV=y
+CONFIG_PV_LINEAR_PT=y
+CONFIG_HVM=y
+# CONFIG_SHADOW_PAGING is not set
+# CONFIG_BIGMEM is not set
+# CONFIG_HVM_FEP is not set
+# CONFIG_TBOOT is not set
+CONFIG_XEN_GUEST=y
+CONFIG_PVH_GUEST=y
+CONFIG_PV_SHIM=y
+CONFIG_PV_SHIM_EXCLUSIVE=y
+
+#
+# Common Features
+#
+CONFIG_COMPAT=y
+CONFIG_CORE_PARKING=y
+CONFIG_HAS_ALTERNATIVE=y
+CONFIG_HAS_EX_TABLE=y
+CONFIG_HAS_MEM_ACCESS=y
+CONFIG_HAS_MEM_PAGING=y
+CONFIG_HAS_MEM_SHARING=y
+CONFIG_HAS_PDX=y
+CONFIG_HAS_UBSAN=y
+CONFIG_HAS_KEXEC=y
+CONFIG_HAS_GDBSX=y
+CONFIG_HAS_IOPORTS=y
+# CONFIG_KEXEC is not set
+# CONFIG_TMEM is not set
+# CONFIG_XENOPROF is not set
+# CONFIG_XSM is not set
+
+#
+# Schedulers
+#
+CONFIG_SCHED_CREDIT=y
+# CONFIG_SCHED_CREDIT2 is not set
+# CONFIG_SCHED_RTDS is not set
+# CONFIG_SCHED_ARINC653 is not set
+CONFIG_SCHED_NULL=y
+# CONFIG_SCHED_CREDIT_DEFAULT is not set
+CONFIG_SCHED_NULL_DEFAULT=y
+CONFIG_SCHED_DEFAULT="null"
+# CONFIG_LIVEPATCH is not set
+# CONFIG_SUPPRESS_DUPLICATE_SYMBOL_WARNINGS is not set
+CONFIG_CMDLINE=""
+
+#
+# Device Drivers
+#
+CONFIG_ACPI=y
+CONFIG_ACPI_LEGACY_TABLES_LOOKUP=y
+CONFIG_NUMA=y
+CONFIG_HAS_NS16550=y
+CONFIG_HAS_EHCI=y
+CONFIG_HAS_CPUFREQ=y
+CONFIG_HAS_PASSTHROUGH=y
+CONFIG_HAS_PCI=y
+CONFIG_VIDEO=y
+CONFIG_VGA=y
+CONFIG_DEFCONFIG_LIST="$ARCH_DEFCONFIG"
+CONFIG_ARCH_SUPPORTS_INT128=y
+
+#
+# Debugging Options
+#
+# CONFIG_DEBUG is not set
+# CONFIG_CRASH_DEBUG is not set
+# CONFIG_FRAME_POINTER is not set
+# CONFIG_GCOV is not set
+# CONFIG_LOCK_PROFILE is not set
+# CONFIG_PERF_COUNTERS is not set
+# CONFIG_VERBOSE_DEBUG is not set
+# CONFIG_SCRUB_DEBUG is not set
+# CONFIG_UBSAN is not set
diff --git a/xen/Makefile b/xen/Makefile
index 044e7c82a3..027c5adfdd 100644
--- a/xen/Makefile
+++ b/xen/Makefile
@@ -37,10 +37,10 @@ default: build
.PHONY: dist
dist: install
-build install:: include/config/auto.conf
+build install build-shim:: include/config/auto.conf
-.PHONY: build install uninstall clean distclean cscope TAGS tags MAP gtags tests
-build install uninstall debug clean distclean cscope TAGS tags MAP gtags tests::
+.PHONY: build install uninstall clean distclean cscope TAGS tags MAP gtags tests install-shim build-shim
+build install uninstall debug clean distclean cscope TAGS tags MAP gtags tests install-shim build-shim::
ifneq ($(XEN_TARGET_ARCH),x86_32)
$(MAKE) -f Rules.mk _$@
else
@@ -80,6 +80,13 @@ _install: $(TARGET)$(CONFIG_XEN_INSTALL_SUFFIX)
fi; \
fi
+.PHONY: _build-shim
+_build-shim: $(TARGET)-shim
+
+.PHONY: _install-shim
+_install-shim: build-shim
+ $(INSTALL_DATA) $(TARGET)-shim $(DESTDIR)
+
.PHONY: _tests
_tests:
$(MAKE) -f $(BASEDIR)/Rules.mk -C test tests
@@ -144,6 +151,9 @@ $(TARGET): delete-unfresh-files
$(MAKE) -f $(BASEDIR)/Rules.mk include/asm-$(TARGET_ARCH)/asm-offsets.h
$(MAKE) -f $(BASEDIR)/Rules.mk -C arch/$(TARGET_ARCH) $(TARGET)
+$(TARGET)-shim: $(TARGET)
+ $(MAKE) -f $(BASEDIR)/Rules.mk -C arch/$(TARGET_ARCH) $(TARGET)-shim
+
# drivers/char/console.o contains static banner/compile info. Blow it away.
# Don't refresh these files during e.g., 'sudo make install'
.PHONY: delete-unfresh-files
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 40/62] xen/x86: make VGA support selectable
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (38 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 39/62] tools/firmware: Build and install xen-shim Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 41/62] xen/pvh: do not mark the low 1MB as IO mem Wei Liu
` (21 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Through a Kconfig option. Enable it by default, and disable it for the
PV-in-PVH shim.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
tools/firmware/xen-dir/shim.config | 2 --
xen/arch/x86/Kconfig | 1 -
xen/arch/x86/boot/build32.mk | 1 +
xen/arch/x86/boot/cmdline.c | 5 ++++-
xen/arch/x86/boot/trampoline.S | 7 +++++++
xen/arch/x86/efi/efi-boot.h | 4 ++++
xen/arch/x86/platform_hypercall.c | 2 ++
xen/arch/x86/pv/dom0_build.c | 2 ++
xen/arch/x86/setup.c | 6 ++++++
xen/drivers/video/Kconfig | 8 +++++++-
xen/include/asm-x86/setup.h | 6 ++++++
11 files changed, 39 insertions(+), 5 deletions(-)
diff --git a/tools/firmware/xen-dir/shim.config b/tools/firmware/xen-dir/shim.config
index 227a12fb4c..d5bd516632 100644
--- a/tools/firmware/xen-dir/shim.config
+++ b/tools/firmware/xen-dir/shim.config
@@ -68,8 +68,6 @@ CONFIG_HAS_EHCI=y
CONFIG_HAS_CPUFREQ=y
CONFIG_HAS_PASSTHROUGH=y
CONFIG_HAS_PCI=y
-CONFIG_VIDEO=y
-CONFIG_VGA=y
CONFIG_DEFCONFIG_LIST="$ARCH_DEFCONFIG"
CONFIG_ARCH_SUPPORTS_INT128=y
diff --git a/xen/arch/x86/Kconfig b/xen/arch/x86/Kconfig
index 4953533f16..f621e799ed 100644
--- a/xen/arch/x86/Kconfig
+++ b/xen/arch/x86/Kconfig
@@ -24,7 +24,6 @@ config X86
select HAS_PDX
select HAS_UBSAN
select NUMA
- select VGA
config ARCH_DEFCONFIG
string
diff --git a/xen/arch/x86/boot/build32.mk b/xen/arch/x86/boot/build32.mk
index f7e8ebe67d..48c7407c00 100644
--- a/xen/arch/x86/boot/build32.mk
+++ b/xen/arch/x86/boot/build32.mk
@@ -5,6 +5,7 @@ include $(XEN_ROOT)/Config.mk
$(call cc-options-add,CFLAGS,CC,$(EMBEDDED_EXTRA_CFLAGS))
CFLAGS += -Werror -fno-asynchronous-unwind-tables -fno-builtin -g0 -msoft-float
+CFLAGS += -I$(XEN_ROOT)/xen/include
CFLAGS := $(filter-out -flto,$(CFLAGS))
# NB. awk invocation is a portable alternative to 'head -n -1'
diff --git a/xen/arch/x86/boot/cmdline.c b/xen/arch/x86/boot/cmdline.c
index 06aa064e72..51b0659a04 100644
--- a/xen/arch/x86/boot/cmdline.c
+++ b/xen/arch/x86/boot/cmdline.c
@@ -30,6 +30,7 @@ asm (
" jmp cmdline_parse_early \n"
);
+#include <xen/kconfig.h>
#include "defs.h"
#include "video.h"
@@ -336,5 +337,7 @@ void __stdcall cmdline_parse_early(const char *cmdline, early_boot_opts_t *ebo)
ebo->skip_realmode = skip_realmode(cmdline);
ebo->opt_edd = edd_parse(cmdline);
ebo->opt_edid = edid_parse(cmdline);
- vga_parse(cmdline, ebo);
+
+ if ( IS_ENABLED(CONFIG_VIDEO) )
+ vga_parse(cmdline, ebo);
}
diff --git a/xen/arch/x86/boot/trampoline.S b/xen/arch/x86/boot/trampoline.S
index 4d640f3fcd..a17a90df5e 100644
--- a/xen/arch/x86/boot/trampoline.S
+++ b/xen/arch/x86/boot/trampoline.S
@@ -219,7 +219,9 @@ trampoline_boot_cpu_entry:
*/
call get_memory_map
call get_edd
+#ifdef CONFIG_VIDEO
call video
+#endif
mov $0x0200,%ax
int $0x16
@@ -267,10 +269,13 @@ opt_edid:
.byte 0 /* EDID parsing option (force/no/default). */
/* Padding. */
.byte 0
+
+#ifdef CONFIG_VIDEO
GLOBAL(boot_vid_mode)
.word VIDEO_80x25 /* If we don't run at all, assume basic video mode 3 at 80x25. */
vesa_size:
.word 0,0,0 /* width x depth x height */
+#endif
GLOBAL(kbd_shift_flags)
.byte 0
@@ -279,4 +284,6 @@ rm_idt: .word 256*4-1, 0, 0
#include "mem.S"
#include "edd.S"
+#ifdef CONFIG_VIDEO
#include "video.S"
+#endif
diff --git a/xen/arch/x86/efi/efi-boot.h b/xen/arch/x86/efi/efi-boot.h
index d30f688a5a..5789d2cb70 100644
--- a/xen/arch/x86/efi/efi-boot.h
+++ b/xen/arch/x86/efi/efi-boot.h
@@ -479,16 +479,19 @@ static void __init efi_arch_edd(void)
static void __init efi_arch_console_init(UINTN cols, UINTN rows)
{
+#ifdef CONFIG_VIDEO
vga_console_info.video_type = XEN_VGATYPE_TEXT_MODE_3;
vga_console_info.u.text_mode_3.columns = cols;
vga_console_info.u.text_mode_3.rows = rows;
vga_console_info.u.text_mode_3.font_height = 16;
+#endif
}
static void __init efi_arch_video_init(EFI_GRAPHICS_OUTPUT_PROTOCOL *gop,
UINTN info_size,
EFI_GRAPHICS_OUTPUT_MODE_INFORMATION *mode_info)
{
+#ifdef CONFIG_VIDEO
int bpp = 0;
switch ( mode_info->PixelFormat )
@@ -550,6 +553,7 @@ static void __init efi_arch_video_init(EFI_GRAPHICS_OUTPUT_PROTOCOL *gop,
vga_console_info.u.vesa_lfb.lfb_size =
(gop->Mode->FrameBufferSize + 0xffff) >> 16;
}
+#endif
}
static void __init efi_arch_memory_setup(void)
diff --git a/xen/arch/x86/platform_hypercall.c b/xen/arch/x86/platform_hypercall.c
index ebc2f394ee..ea18c3215a 100644
--- a/xen/arch/x86/platform_hypercall.c
+++ b/xen/arch/x86/platform_hypercall.c
@@ -388,6 +388,7 @@ ret_t do_platform_op(XEN_GUEST_HANDLE_PARAM(xen_platform_op_t) u_xenpf_op)
}
case XEN_FW_VBEDDC_INFO:
ret = -ESRCH;
+#ifdef CONFIG_VIDEO
if ( op->u.firmware_info.index != 0 )
break;
if ( *(u32 *)bootsym(boot_edid_info) == 0x13131313 )
@@ -406,6 +407,7 @@ ret_t do_platform_op(XEN_GUEST_HANDLE_PARAM(xen_platform_op_t) u_xenpf_op)
copy_to_compat(op->u.firmware_info.u.vbeddc_info.edid,
bootsym(boot_edid_info), 128) )
ret = -EFAULT;
+#endif
break;
case XEN_FW_EFI_INFO:
ret = efi_get_info(op->u.firmware_info.index,
diff --git a/xen/arch/x86/pv/dom0_build.c b/xen/arch/x86/pv/dom0_build.c
index 8fd40b5254..50a423e35d 100644
--- a/xen/arch/x86/pv/dom0_build.c
+++ b/xen/arch/x86/pv/dom0_build.c
@@ -827,11 +827,13 @@ int __init dom0_construct_pv(struct domain *d,
if ( cmdline != NULL )
strlcpy((char *)si->cmd_line, cmdline, sizeof(si->cmd_line));
+#ifdef CONFIG_VIDEO
if ( fill_console_start_info((void *)(si + 1)) )
{
si->console.dom0.info_off = sizeof(struct start_info);
si->console.dom0.info_size = sizeof(struct dom0_vga_console_info);
}
+#endif
if ( is_pv_32bit_domain(d) )
xlat_start_info(si, XLAT_start_info_console_dom0);
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 0c9b55ff71..b2abcaa278 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -456,6 +456,7 @@ static void __init setup_max_pdx(unsigned long top_page)
/* A temporary copy of the e820 map that we can mess with during bootstrap. */
static struct e820map __initdata boot_e820;
+#ifdef CONFIG_VIDEO
struct boot_video_info {
u8 orig_x; /* 0x00 */
u8 orig_y; /* 0x01 */
@@ -486,9 +487,11 @@ struct boot_video_info {
u16 vesa_attrib; /* 0x28 */
};
extern struct boot_video_info boot_vid_info;
+#endif
static void __init parse_video_info(void)
{
+#ifdef CONFIG_VIDEO
struct boot_video_info *bvi = &bootsym(boot_vid_info);
/* vga_console_info is filled directly on EFI platform. */
@@ -524,6 +527,7 @@ static void __init parse_video_info(void)
vga_console_info.u.vesa_lfb.gbl_caps = bvi->capabilities;
vga_console_info.u.vesa_lfb.mode_attrs = bvi->vesa_attrib;
}
+#endif
}
static void __init kexec_reserve_area(struct e820map *e820)
@@ -741,6 +745,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
printk("Xen image load base address: %#lx\n", xen_phys_start);
+#ifdef CONFIG_VIDEO
printk("Video information:\n");
/* Print VGA display mode information. */
@@ -784,6 +789,7 @@ void __init noreturn __start_xen(unsigned long mbi_p)
printk("of reasons unknown\n");
}
}
+#endif
printk("Disc information:\n");
printk(" Found %d MBR signatures\n",
diff --git a/xen/drivers/video/Kconfig b/xen/drivers/video/Kconfig
index 0ffbbd9a88..52e8ce6c15 100644
--- a/xen/drivers/video/Kconfig
+++ b/xen/drivers/video/Kconfig
@@ -3,8 +3,14 @@ config VIDEO
bool
config VGA
- bool
+ bool "VGA support" if !PV_SHIM_EXCLUSIVE
select VIDEO
+ depends on X86
+ default y if !PV_SHIM_EXCLUSIVE
+ ---help---
+ Enable VGA output for the Xen hypervisor.
+
+ If unsure, say Y.
config HAS_ARM_HDLCD
bool
diff --git a/xen/include/asm-x86/setup.h b/xen/include/asm-x86/setup.h
index 05d647d02e..19232afa01 100644
--- a/xen/include/asm-x86/setup.h
+++ b/xen/include/asm-x86/setup.h
@@ -31,8 +31,14 @@ void arch_init_memory(void);
void subarch_init_memory(void);
void init_IRQ(void);
+
+#ifdef CONFIG_VIDEO
void vesa_init(void);
void vesa_mtrr_init(void);
+#else
+static inline void vesa_init(void) {};
+static inline void vesa_mtrr_init(void) {};
+#endif
int construct_dom0(
struct domain *d,
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 41/62] xen/pvh: do not mark the low 1MB as IO mem
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (39 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 40/62] xen/x86: make VGA support selectable Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 42/62] sched/null: skip vCPUs on the waitqueue that are blocked Wei Liu
` (20 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
On PVH there's nothing special on the low 1MB.
This is an optional patch that doesn't affect the functionality of the
shim.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
xen/arch/x86/mm.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index f73fee225e..355e6747bb 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -122,6 +122,7 @@
#include <asm/fixmap.h>
#include <asm/io_apic.h>
#include <asm/pci.h>
+#include <asm/guest.h>
#include <asm/hvm/grant_table.h>
#include <asm/pv/grant_table.h>
@@ -288,8 +289,12 @@ void __init arch_init_memory(void)
dom_cow = domain_create(DOMID_COW, DOMCRF_dummy, 0, NULL);
BUG_ON(IS_ERR(dom_cow));
- /* First 1MB of RAM is historically marked as I/O. */
- for ( i = 0; i < 0x100; i++ )
+ /*
+ * First 1MB of RAM is historically marked as I/O. If we booted PVH,
+ * reclaim the space. Irrespective, leave MFN 0 as special for the sake
+ * of 0 being a very common default value.
+ */
+ for ( i = 0; i < (pvh_boot ? 1 : 0x100); i++ )
share_xen_page_with_guest(mfn_to_page(_mfn(i)),
dom_io, XENSHARE_writable);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 42/62] sched/null: skip vCPUs on the waitqueue that are blocked
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (40 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 41/62] xen/pvh: do not mark the low 1MB as IO mem Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 43/62] xen/pvshim: skip Dom0-only domain builder parts Wei Liu
` (19 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Avoid scheduling vCPUs that are down, there's no point in assigning
them to a pCPU because they are not going to run anyway.
Since down vCPUs are not assigned to pCPUs after this change, force a
rescheduling when a vCPU is brought up if it's on the waitqueue. Also
when scheduling try to pick a vCPU from the runqueue if the pCPU is
running idle.
There's no current way to prevent a user from adding more vcpus to a
pool than there are pcpus (if nothing else, by creating a new VM in a
given pool), or from taking pcpus from a pool in which #vcpus >=
#pcpus.
The null scheduler deals with this by having a queue of "unassigned"
vcpus that are waiting for a free pcpu. When a pcpu becomes
available, it will do the assignment. When a pcpu that has a vcpu is
assigned is removed from the pool, that vcpu is assigned to a
different pcpu if one is available; if not, it is put on the list.
In the case of shim mode, this also seems to happen whenever curvcpus
< maxvcpus: The L1 hypervisor (shim) only sees curvcpus cpus on which
to schedule L2 vcpus, but the L2 guest has maxvcpus vcpus to schedule,
of which (maxvcpus-curvcpus) are marked down. In this case, it also
seems that the null scheduler sometimes schedules a down vcpu when
there are up vcpus on the list; meaning that the up vcpus are never
scheduled.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Cc: George Dunlap <george.dunlap@eu.citrix.com>
Cc: Dario Faggioli <raistlin@linux.it>
---
Changes since v1:
- Force a rescheduling when a vCPU is brought up.
- Try to pick a vCPU from the runqueue if running the idle vCPU.
- Add George Dunlap description of the problem to the commit log.
---
xen/common/sched_null.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/xen/common/sched_null.c b/xen/common/sched_null.c
index b4a24baf8e..bacfb31cb3 100644
--- a/xen/common/sched_null.c
+++ b/xen/common/sched_null.c
@@ -574,6 +574,8 @@ static void null_vcpu_wake(const struct scheduler *ops, struct vcpu *v)
{
/* Not exactly "on runq", but close enough for reusing the counter */
SCHED_STAT_CRANK(vcpu_wake_onrunq);
+ /* Force a rescheduling in case some CPU is idle can pick this vCPU */
+ cpumask_raise_softirq(&cpu_online_map, SCHEDULE_SOFTIRQ);
return;
}
@@ -761,9 +763,10 @@ static struct task_slice null_schedule(const struct scheduler *ops,
/*
* We may be new in the cpupool, or just coming back online. In which
* case, there may be vCPUs in the waitqueue that we can assign to us
- * and run.
+ * and run. Also check whether this CPU is running idle, in which case try
+ * to pick a vCPU from the waitqueue.
*/
- if ( unlikely(ret.task == NULL) )
+ if ( unlikely(ret.task == NULL || ret.task == idle_vcpu[cpu]) )
{
spin_lock(&prv->waitq_lock);
@@ -781,6 +784,10 @@ static struct task_slice null_schedule(const struct scheduler *ops,
{
list_for_each_entry( wvc, &prv->waitq, waitq_elem )
{
+ if ( test_bit(_VPF_down, &wvc->vcpu->pause_flags) )
+ /* Skip vCPUs that are down. */
+ continue;
+
if ( bs == BALANCE_SOFT_AFFINITY &&
!has_soft_affinity(wvc->vcpu, wvc->vcpu->cpu_hard_affinity) )
continue;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 43/62] xen/pvshim: skip Dom0-only domain builder parts
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (41 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 42/62] sched/null: skip vCPUs on the waitqueue that are blocked Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 44/62] xen: mark xenstore/console pages as RAM Wei Liu
` (18 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Do not allow access to any iomem or ioport by the shim, and also
remove the check for Dom0 kernel support.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
---
xen/arch/x86/dom0_build.c | 4 ++++
xen/arch/x86/pv/dom0_build.c | 3 ++-
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 75b2bb3bee..907acad060 100644
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -13,6 +13,7 @@
#include <xen/softirq.h>
#include <asm/dom0_build.h>
+#include <asm/guest.h>
#include <asm/hpet.h>
#include <asm/io_apic.h>
#include <asm/p2m.h>
@@ -385,6 +386,9 @@ int __init dom0_setup_permissions(struct domain *d)
unsigned int i;
int rc;
+ if ( pv_shim )
+ return 0;
+
/* The hardware domain is initially permitted full I/O capabilities. */
rc = ioports_permit_access(d, 0, 0xFFFF);
rc |= iomem_permit_access(d, 0UL, (1UL << (paddr_bits - PAGE_SHIFT)) - 1);
diff --git a/xen/arch/x86/pv/dom0_build.c b/xen/arch/x86/pv/dom0_build.c
index 50a423e35d..6118970d44 100644
--- a/xen/arch/x86/pv/dom0_build.c
+++ b/xen/arch/x86/pv/dom0_build.c
@@ -17,6 +17,7 @@
#include <asm/bzimage.h>
#include <asm/dom0_build.h>
+#include <asm/guest.h>
#include <asm/page.h>
#include <asm/pv/mm.h>
#include <asm/setup.h>
@@ -372,7 +373,7 @@ int __init dom0_construct_pv(struct domain *d,
if ( parms.elf_notes[XEN_ELFNOTE_SUPPORTED_FEATURES].type != XEN_ENT_NONE )
{
- if ( !test_bit(XENFEAT_dom0, parms.f_supported) )
+ if ( !pv_shim && !test_bit(XENFEAT_dom0, parms.f_supported) )
{
printk("Kernel does not support Dom0 operation\n");
rc = -EINVAL;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 44/62] xen: mark xenstore/console pages as RAM
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (42 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 43/62] xen/pvshim: skip Dom0-only domain builder parts Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 45/62] xen/pvshim: modify Dom0 builder in order to build a DomU Wei Liu
` (17 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
This si required so that later they can be shared with the guest if
Xen is running in shim mode.
Also prevent them from being used by Xen by marking them as bad pages
in init_boot_pages.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
Changes since v1:
- Remove adding the pages to dom_io, there's no need since they are
already marked as bad pages.
- Use a static global array to store the memory address of this
special pages, so Xen avoids having to call
xen_hypercall_hvm_get_param twice.
---
xen/arch/x86/e820.c | 4 ++++
xen/arch/x86/guest/xen.c | 43 +++++++++++++++++++++++++++++++++++++++
xen/common/page_alloc.c | 15 ++++++++++++++
xen/drivers/char/xen_pv_console.c | 4 ++++
xen/include/asm-x86/guest/xen.h | 14 +++++++++++++
5 files changed, 80 insertions(+)
diff --git a/xen/arch/x86/e820.c b/xen/arch/x86/e820.c
index b422a684ee..590ea985ef 100644
--- a/xen/arch/x86/e820.c
+++ b/xen/arch/x86/e820.c
@@ -9,6 +9,7 @@
#include <asm/processor.h>
#include <asm/mtrr.h>
#include <asm/msr.h>
+#include <asm/guest.h>
/*
* opt_mem: Limit maximum address of physical RAM.
@@ -699,6 +700,9 @@ unsigned long __init init_e820(const char *str, struct e820map *raw)
machine_specific_memory_setup(raw);
+ if ( xen_guest )
+ hypervisor_fixup_e820(&e820);
+
printk("%s RAM map:\n", str);
print_e820_memory_map(e820.map, e820.nr_map);
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index d4968b47aa..27a6c47753 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -32,12 +32,14 @@
#include <asm/processor.h>
#include <public/arch-x86/cpuid.h>
+#include <public/hvm/params.h>
bool __read_mostly xen_guest;
static __read_mostly uint32_t xen_cpuid_base;
extern char hypercall_page[];
static struct rangeset *mem;
+static unsigned long __initdata reserved_pages[2];
DEFINE_PER_CPU(unsigned int, vcpu_id);
@@ -279,6 +281,47 @@ int hypervisor_free_unused_page(mfn_t mfn)
return rangeset_remove_range(mem, mfn_x(mfn), mfn_x(mfn));
}
+static void __init mark_pfn_as_ram(struct e820map *e820, uint64_t pfn)
+{
+ if ( !e820_add_range(e820, pfn << PAGE_SHIFT,
+ (pfn << PAGE_SHIFT) + PAGE_SIZE, E820_RAM) )
+ if ( !e820_change_range_type(e820, pfn << PAGE_SHIFT,
+ (pfn << PAGE_SHIFT) + PAGE_SIZE,
+ E820_RESERVED, E820_RAM) )
+ panic("Unable to add/change memory type of pfn %#lx to RAM", pfn);
+}
+
+void __init hypervisor_fixup_e820(struct e820map *e820)
+{
+ uint64_t pfn = 0;
+ unsigned int i = 0;
+ long rc;
+
+ ASSERT(xen_guest);
+
+#define MARK_PARAM_RAM(p) ({ \
+ rc = xen_hypercall_hvm_get_param(p, &pfn); \
+ if ( rc ) \
+ panic("Unable to get " #p); \
+ mark_pfn_as_ram(e820, pfn); \
+ ASSERT(i < ARRAY_SIZE(reserved_pages)); \
+ reserved_pages[i++] = pfn << PAGE_SHIFT; \
+})
+ MARK_PARAM_RAM(HVM_PARAM_STORE_PFN);
+ if ( !pv_console )
+ MARK_PARAM_RAM(HVM_PARAM_CONSOLE_PFN);
+#undef MARK_PARAM_RAM
+}
+
+const unsigned long *__init hypervisor_reserved_pages(unsigned int *size)
+{
+ ASSERT(xen_guest);
+
+ *size = ARRAY_SIZE(reserved_pages);
+
+ return reserved_pages;
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/common/page_alloc.c b/xen/common/page_alloc.c
index c0c2d82906..4de8988bea 100644
--- a/xen/common/page_alloc.c
+++ b/xen/common/page_alloc.c
@@ -143,6 +143,7 @@
#include <asm/numa.h>
#include <asm/flushtlb.h>
#ifdef CONFIG_X86
+#include <asm/guest.h>
#include <asm/p2m.h>
#include <asm/setup.h> /* for highmem_start only */
#else
@@ -303,6 +304,20 @@ void __init init_boot_pages(paddr_t ps, paddr_t pe)
badpage++;
}
}
+
+ if ( xen_guest )
+ {
+ badpage = hypervisor_reserved_pages(&array_size);
+ if ( badpage )
+ {
+ for ( i = 0; i < array_size; i++ )
+ {
+ bootmem_region_zap(*badpage >> PAGE_SHIFT,
+ (*badpage >> PAGE_SHIFT) + 1);
+ badpage++;
+ }
+ }
+ }
#endif
/* Check new pages against the bad-page list. */
diff --git a/xen/drivers/char/xen_pv_console.c b/xen/drivers/char/xen_pv_console.c
index f5aca4c69e..d4f0532101 100644
--- a/xen/drivers/char/xen_pv_console.c
+++ b/xen/drivers/char/xen_pv_console.c
@@ -35,6 +35,8 @@ static evtchn_port_t cons_evtchn;
static serial_rx_fn cons_rx_handler;
static DEFINE_SPINLOCK(tx_lock);
+bool pv_console;
+
void __init pv_console_init(void)
{
long r;
@@ -60,6 +62,8 @@ void __init pv_console_init(void)
printk("Initialised PV console at 0x%p with pfn %#lx and evtchn %#x\n",
cons_ring, raw_pfn, cons_evtchn);
+ pv_console = true;
+
return;
error:
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index b3e684f756..62255fda8b 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -29,12 +29,15 @@
#ifdef CONFIG_XEN_GUEST
extern bool xen_guest;
+extern bool pv_console;
void probe_hypervisor(void);
void hypervisor_setup(void);
void hypervisor_ap_setup(void);
int hypervisor_alloc_unused_page(mfn_t *mfn);
int hypervisor_free_unused_page(mfn_t mfn);
+void hypervisor_fixup_e820(struct e820map *e820);
+const unsigned long *hypervisor_reserved_pages(unsigned int *size);
DECLARE_PER_CPU(unsigned int, vcpu_id);
DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
@@ -42,6 +45,7 @@ DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
#else
#define xen_guest 0
+#define pv_console 0
static inline void probe_hypervisor(void) {};
static inline void hypervisor_setup(void)
@@ -53,6 +57,16 @@ static inline void hypervisor_ap_setup(void)
ASSERT_UNREACHABLE();
}
+static inline void hypervisor_fixup_e820(struct e820map *e820)
+{
+ ASSERT_UNREACHABLE();
+}
+static inline const unsigned long *hypervisor_reserved_pages(unsigned int *size)
+{
+ ASSERT_UNREACHABLE();
+ return NULL;
+};
+
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_GUEST_XEN_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 45/62] xen/pvshim: modify Dom0 builder in order to build a DomU
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (43 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 44/62] xen: mark xenstore/console pages as RAM Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 46/62] xen/pvshim: set correct domid value Wei Liu
` (16 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
According to the PV ABI the initial virtual memory regions should
contain the xenstore and console pages after the start_info. Also set
the correct values in the start_info for DomU operation.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Modify the position of the __init attribute in dom0_update_physmap.
- Move the addition of sizeof(struct dom0_vga_console_info) to
vstartinfo_end with an existing if branch.
- Add a TODO item for fill_console_start_info in the !CONFIG_VIDEO
case.
- s/replace_va/replace_va_mapping/.
- Remove call to free_domheap_pages in replace_va_mapping.
put_page_and_type should already take care of freeing the page.
- Use PFN_DOWN in SET_AND_MAP_PARAM macro.
- Parenthesize va in SET_AND_MAP_PARAM macro when required.
---
xen/arch/x86/pv/dom0_build.c | 48 +++++++++++++++++++++++-------
xen/arch/x86/pv/shim.c | 63 ++++++++++++++++++++++++++++++++++++++++
xen/include/asm-x86/dom0_build.h | 4 +++
xen/include/asm-x86/pv/shim.h | 21 ++++++++++++++
4 files changed, 126 insertions(+), 10 deletions(-)
diff --git a/xen/arch/x86/pv/dom0_build.c b/xen/arch/x86/pv/dom0_build.c
index 6118970d44..52576795cc 100644
--- a/xen/arch/x86/pv/dom0_build.c
+++ b/xen/arch/x86/pv/dom0_build.c
@@ -31,9 +31,8 @@
#define L3_PROT (BASE_PROT|_PAGE_DIRTY)
#define L4_PROT (BASE_PROT|_PAGE_DIRTY)
-static __init void dom0_update_physmap(struct domain *d, unsigned long pfn,
- unsigned long mfn,
- unsigned long vphysmap_s)
+void __init dom0_update_physmap(struct domain *d, unsigned long pfn,
+ unsigned long mfn, unsigned long vphysmap_s)
{
if ( !is_pv_32bit_domain(d) )
((unsigned long *)vphysmap_s)[pfn] = mfn;
@@ -315,6 +314,10 @@ int __init dom0_construct_pv(struct domain *d,
unsigned long vphysmap_end;
unsigned long vstartinfo_start;
unsigned long vstartinfo_end;
+ unsigned long vxenstore_start = 0;
+ unsigned long vxenstore_end = 0;
+ unsigned long vconsole_start = 0;
+ unsigned long vconsole_end = 0;
unsigned long vstack_start;
unsigned long vstack_end;
unsigned long vpt_start;
@@ -440,11 +443,22 @@ int __init dom0_construct_pv(struct domain *d,
if ( parms.p2m_base != UNSET_ADDR )
vphysmap_end = vphysmap_start;
vstartinfo_start = round_pgup(vphysmap_end);
- vstartinfo_end = (vstartinfo_start +
- sizeof(struct start_info) +
- sizeof(struct dom0_vga_console_info));
+ vstartinfo_end = vstartinfo_start + sizeof(struct start_info);
+
+ if ( pv_shim )
+ {
+ vxenstore_start = round_pgup(vstartinfo_end);
+ vxenstore_end = vxenstore_start + PAGE_SIZE;
+ vconsole_start = vxenstore_end;
+ vconsole_end = vconsole_start + PAGE_SIZE;
+ vpt_start = vconsole_end;
+ }
+ else
+ {
+ vpt_start = round_pgup(vstartinfo_end);
+ vstartinfo_end += sizeof(struct dom0_vga_console_info);
+ }
- vpt_start = round_pgup(vstartinfo_end);
for ( nr_pt_pages = 2; ; nr_pt_pages++ )
{
vpt_end = vpt_start + (nr_pt_pages * PAGE_SIZE);
@@ -537,6 +551,8 @@ int __init dom0_construct_pv(struct domain *d,
" Init. ramdisk: %p->%p\n"
" Phys-Mach map: %p->%p\n"
" Start info: %p->%p\n"
+ " Xenstore ring: %p->%p\n"
+ " Console ring: %p->%p\n"
" Page tables: %p->%p\n"
" Boot stack: %p->%p\n"
" TOTAL: %p->%p\n",
@@ -544,6 +560,8 @@ int __init dom0_construct_pv(struct domain *d,
_p(vinitrd_start), _p(vinitrd_end),
_p(vphysmap_start), _p(vphysmap_end),
_p(vstartinfo_start), _p(vstartinfo_end),
+ _p(vxenstore_start), _p(vxenstore_end),
+ _p(vconsole_start), _p(vconsole_end),
_p(vpt_start), _p(vpt_end),
_p(vstack_start), _p(vstack_end),
_p(v_start), _p(v_end));
@@ -737,7 +755,8 @@ int __init dom0_construct_pv(struct domain *d,
si->shared_info = virt_to_maddr(d->shared_info);
- si->flags = SIF_PRIVILEGED | SIF_INITDOMAIN;
+ if ( !pv_shim )
+ si->flags = SIF_PRIVILEGED | SIF_INITDOMAIN;
if ( !vinitrd_start && initrd_len )
si->flags |= SIF_MOD_START_PFN;
si->flags |= (xen_processor_pmbits << 8) & SIF_PM_MASK;
@@ -829,15 +848,24 @@ int __init dom0_construct_pv(struct domain *d,
strlcpy((char *)si->cmd_line, cmdline, sizeof(si->cmd_line));
#ifdef CONFIG_VIDEO
- if ( fill_console_start_info((void *)(si + 1)) )
+ if ( !pv_shim && fill_console_start_info((void *)(si + 1)) )
{
si->console.dom0.info_off = sizeof(struct start_info);
si->console.dom0.info_size = sizeof(struct dom0_vga_console_info);
}
#endif
+ /*
+ * TODO: provide an empty stub for fill_console_start_info in the
+ * !CONFIG_VIDEO case so the logic here can be simplified.
+ */
+ if ( pv_shim )
+ pv_shim_setup_dom(d, l4start, v_start, vxenstore_start, vconsole_start,
+ vphysmap_start, si);
+
if ( is_pv_32bit_domain(d) )
- xlat_start_info(si, XLAT_start_info_console_dom0);
+ xlat_start_info(si, pv_shim ? XLAT_start_info_console_domU
+ : XLAT_start_info_console_dom0);
/* Return to idle domain's page tables. */
mapcache_override_current(NULL);
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 4d037355db..75365b0697 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -18,16 +18,79 @@
*
* Copyright (c) 2017 Citrix Systems Ltd.
*/
+#include <xen/hypercall.h>
#include <xen/init.h>
#include <xen/types.h>
#include <asm/apic.h>
+#include <asm/dom0_build.h>
+#include <asm/guest.h>
+#include <asm/pv/mm.h>
#ifndef CONFIG_PV_SHIM_EXCLUSIVE
bool pv_shim;
boolean_param("pv-shim", pv_shim);
#endif
+#define L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_USER| \
+ _PAGE_GUEST_KERNEL)
+#define COMPAT_L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED)
+
+static void __init replace_va_mapping(struct domain *d, l4_pgentry_t *l4start,
+ unsigned long va, unsigned long mfn)
+{
+ struct page_info *page;
+ l4_pgentry_t *pl4e;
+ l3_pgentry_t *pl3e;
+ l2_pgentry_t *pl2e;
+ l1_pgentry_t *pl1e;
+
+ pl4e = l4start + l4_table_offset(va);
+ pl3e = l4e_to_l3e(*pl4e);
+ pl3e += l3_table_offset(va);
+ pl2e = l3e_to_l2e(*pl3e);
+ pl2e += l2_table_offset(va);
+ pl1e = l2e_to_l1e(*pl2e);
+ pl1e += l1_table_offset(va);
+
+ page = mfn_to_page(l1e_get_pfn(*pl1e));
+ put_page_and_type(page);
+
+ *pl1e = l1e_from_pfn(mfn, (!is_pv_32bit_domain(d) ? L1_PROT
+ : COMPAT_L1_PROT));
+}
+
+void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
+ unsigned long va_start, unsigned long store_va,
+ unsigned long console_va, unsigned long vphysmap,
+ start_info_t *si)
+{
+ uint64_t param = 0;
+ long rc;
+
+#define SET_AND_MAP_PARAM(p, si, va) ({ \
+ rc = xen_hypercall_hvm_get_param(p, ¶m); \
+ if ( rc ) \
+ panic("Unable to get " #p "\n"); \
+ (si) = param; \
+ if ( va ) \
+ { \
+ share_xen_page_with_guest(mfn_to_page(param), d, XENSHARE_writable); \
+ replace_va_mapping(d, l4start, va, param); \
+ dom0_update_physmap(d, PFN_DOWN((va) - va_start), param, vphysmap); \
+ } \
+})
+ SET_AND_MAP_PARAM(HVM_PARAM_STORE_PFN, si->store_mfn, store_va);
+ SET_AND_MAP_PARAM(HVM_PARAM_STORE_EVTCHN, si->store_evtchn, 0);
+ if ( !pv_console )
+ {
+ SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_PFN, si->console.domU.mfn,
+ console_va);
+ SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_EVTCHN, si->console.domU.evtchn, 0);
+ }
+#undef SET_AND_MAP_PARAM
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/include/asm-x86/dom0_build.h b/xen/include/asm-x86/dom0_build.h
index 25d124f8d4..33a5483739 100644
--- a/xen/include/asm-x86/dom0_build.h
+++ b/xen/include/asm-x86/dom0_build.h
@@ -1,6 +1,7 @@
#ifndef _DOM0_BUILD_H_
#define _DOM0_BUILD_H_
+#include <xen/libelf.h>
#include <xen/sched.h>
#include <asm/setup.h>
@@ -27,6 +28,9 @@ int dom0_construct_pvh(struct domain *d, const module_t *image,
unsigned long dom0_paging_pages(const struct domain *d,
unsigned long nr_pages);
+void dom0_update_physmap(struct domain *d, unsigned long pfn,
+ unsigned long mfn, unsigned long vphysmap_s);
+
#endif /* _DOM0_BUILD_H_ */
/*
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index 1468cfd498..b0c361cba1 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -29,6 +29,27 @@ extern bool pv_shim;
# define pv_shim 0
#endif /* CONFIG_PV_SHIM{,_EXCLUSIVE} */
+#ifdef CONFIG_PV_SHIM
+
+void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
+ unsigned long va_start, unsigned long store_va,
+ unsigned long console_va, unsigned long vphysmap,
+ start_info_t *si);
+
+#else
+
+static inline void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
+ unsigned long va_start,
+ unsigned long store_va,
+ unsigned long console_va,
+ unsigned long vphysmap,
+ start_info_t *si)
+{
+ ASSERT_UNREACHABLE();
+}
+
+#endif
+
#endif /* __X86_PV_SHIM_H__ */
/*
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 46/62] xen/pvshim: set correct domid value
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (44 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 45/62] xen/pvshim: modify Dom0 builder in order to build a DomU Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 47/62] xen/pvshim: forward evtchn ops between L0 Xen and L2 DomU Wei Liu
` (15 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
If domid is not provided by L0 set domid to 1 by default. Note that L0
not provinding the domid can cause trouble if the guest tries to use
it's domid instead of DOMID_SELF when performing hypercalls that are
forwarded to the L0 hypervisor.
Since the domain created is no longer the hardware domain add a hook
to the domain shutdown path in order to forward shutdown operations to
the L0 hypervisor.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
---
Changes since v1:
- s/get_dom0_domid/get_initial_domain_id/.
- Add a comment regarding why dom0 needs to be global.
- Fix compilation of xen/common/domain.c on ARM.
---
xen/arch/x86/dom0_build.c | 2 +-
xen/arch/x86/guest/xen.c | 5 +++++
xen/arch/x86/pv/shim.c | 21 +++++++++++++++++++++
xen/arch/x86/setup.c | 16 +++++++++++-----
xen/common/domain.c | 12 ++++++++++++
xen/include/asm-x86/guest/xen.h | 6 ++++++
xen/include/asm-x86/pv/shim.h | 10 ++++++++++
7 files changed, 66 insertions(+), 6 deletions(-)
diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 907acad060..8600123516 100644
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -472,7 +472,7 @@ int __init construct_dom0(struct domain *d, const module_t *image,
int rc;
/* Sanity! */
- BUG_ON(d->domain_id != 0);
+ BUG_ON(!pv_shim && d->domain_id != 0);
BUG_ON(d->vcpu[0] == NULL);
BUG_ON(d->vcpu[0]->is_initialised);
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 27a6c47753..aff16a0e35 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -322,6 +322,11 @@ const unsigned long *__init hypervisor_reserved_pages(unsigned int *size)
return reserved_pages;
}
+uint32_t hypervisor_cpuid_base(void)
+{
+ return xen_cpuid_base;
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 75365b0697..78351c9ee0 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -20,6 +20,7 @@
*/
#include <xen/hypercall.h>
#include <xen/init.h>
+#include <xen/shutdown.h>
#include <xen/types.h>
#include <asm/apic.h>
@@ -27,6 +28,8 @@
#include <asm/guest.h>
#include <asm/pv/mm.h>
+#include <public/arch-x86/cpuid.h>
+
#ifndef CONFIG_PV_SHIM_EXCLUSIVE
bool pv_shim;
boolean_param("pv-shim", pv_shim);
@@ -91,6 +94,24 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
#undef SET_AND_MAP_PARAM
}
+void pv_shim_shutdown(uint8_t reason)
+{
+ /* XXX: handle suspend */
+ xen_hypercall_shutdown(reason);
+}
+
+domid_t get_initial_domain_id(void)
+{
+ uint32_t eax, ebx, ecx, edx;
+
+ if ( !pv_shim )
+ return 0;
+
+ cpuid(hypervisor_cpuid_base() + 4, &eax, &ebx, &ecx, &edx);
+
+ return (eax & XEN_HVM_CPUID_DOMID_PRESENT) ? ecx : 1;
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index b2abcaa278..0f4282c42f 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -104,6 +104,12 @@ unsigned long __read_mostly mmu_cr4_features = XEN_MINIMAL_CR4;
#define SMEP_HVM_ONLY (-1)
static s8 __initdata opt_smep = 1;
+/*
+ * Initial domain place holder. Needs to be global so it can be created in
+ * __start_xen and unpaused in init_done.
+ */
+static struct domain *__initdata dom0;
+
static int __init parse_smep_param(const char *s)
{
if ( !*s )
@@ -576,11 +582,11 @@ static void noinline init_done(void)
system_state = SYS_STATE_active;
+ domain_unpause_by_systemcontroller(dom0);
+
/* MUST be done prior to removing .init data. */
unregister_init_virtual_region();
- domain_unpause_by_systemcontroller(hardware_domain);
-
/* Zero the .init code and data. */
for ( va = __init_begin; va < _p(__init_end); va += PAGE_SIZE )
clear_page(va);
@@ -659,7 +665,6 @@ void __init noreturn __start_xen(unsigned long mbi_p)
unsigned long nr_pages, raw_max_page, modules_headroom, *module_map;
int i, j, e820_warn = 0, bytes = 0;
bool acpi_boot_table_init_done = false, relocated = false;
- struct domain *dom0;
struct ns16550_defaults ns16550 = {
.data_bits = 8,
.parity = 'n',
@@ -1617,11 +1622,12 @@ void __init noreturn __start_xen(unsigned long mbi_p)
}
/* Create initial domain 0. */
- dom0 = domain_create(0, domcr_flags, 0, &config);
+ dom0 = domain_create(get_initial_domain_id(), domcr_flags, 0, &config);
if ( IS_ERR(dom0) || (alloc_dom0_vcpu0(dom0) == NULL) )
panic("Error creating domain 0");
- dom0->is_privileged = 1;
+ if ( !pv_shim )
+ dom0->is_privileged = 1;
dom0->target = NULL;
/* Grab the DOM0 command line. */
diff --git a/xen/common/domain.c b/xen/common/domain.c
index 7af8d12512..ef566f7f45 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -43,6 +43,10 @@
#include <xen/tmem.h>
#include <asm/setup.h>
+#ifdef CONFIG_X86
+#include <asm/guest.h>
+#endif
+
/* Linux config option: propageted to domain0 */
/* xen_processor_pmbits: xen control Cx, Px, ... */
unsigned int xen_processor_pmbits = XEN_PROCESSOR_PM_PX;
@@ -697,6 +701,14 @@ void domain_shutdown(struct domain *d, u8 reason)
{
struct vcpu *v;
+#ifdef CONFIG_X86
+ if ( pv_shim )
+ {
+ pv_shim_shutdown(reason);
+ return;
+ }
+#endif
+
spin_lock(&d->shutdown_lock);
if ( d->shutdown_code == SHUTDOWN_CODE_INVALID )
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index 62255fda8b..ac48dcbe44 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -38,6 +38,7 @@ int hypervisor_alloc_unused_page(mfn_t *mfn);
int hypervisor_free_unused_page(mfn_t mfn);
void hypervisor_fixup_e820(struct e820map *e820);
const unsigned long *hypervisor_reserved_pages(unsigned int *size);
+uint32_t hypervisor_cpuid_base(void);
DECLARE_PER_CPU(unsigned int, vcpu_id);
DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
@@ -66,6 +67,11 @@ static inline const unsigned long *hypervisor_reserved_pages(unsigned int *size)
ASSERT_UNREACHABLE();
return NULL;
};
+static inline uint32_t hypervisor_cpuid_base(void)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+};
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_GUEST_XEN_H__ */
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index b0c361cba1..ff7c050dc6 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -35,6 +35,8 @@ void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
unsigned long va_start, unsigned long store_va,
unsigned long console_va, unsigned long vphysmap,
start_info_t *si);
+void pv_shim_shutdown(uint8_t reason);
+domid_t get_initial_domain_id(void);
#else
@@ -47,6 +49,14 @@ static inline void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
{
ASSERT_UNREACHABLE();
}
+static inline void pv_shim_shutdown(uint8_t reason)
+{
+ ASSERT_UNREACHABLE();
+}
+static inline domid_t get_initial_domain_id(void)
+{
+ return 0;
+}
#endif
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 47/62] xen/pvshim: forward evtchn ops between L0 Xen and L2 DomU
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (45 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 46/62] xen/pvshim: set correct domid value Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 48/62] xen/pvshim: add grant table operations Wei Liu
` (14 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Note that the unmask and the virq operations are handled by the shim
itself, and that FIFO event channels are not exposed to the guest.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Anthony Liguori <aliguori@amazon.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
---
Changes since v1:
- Use find_first_set_bit instead of ffsl.
- Indent macro one more level.
- Have a single evtchn_close struct in pv_shim_event_channel_op.
- Add blank lines between switch cases.
- Use -EOPNOTSUPP in order to signal lack of FIFO or PIRQ support.
- Switch evtchn_bind_virq parameter to evtchn_port_t and use 0 signal
allocation needed.
- Switch evtchn helpers return type to int instead of long.
- Re-write event channel hypercall table handler instead of adding
hooks.
- Remove the pv_domain variable and instead use a static variable in
shim code.
---
xen/arch/x86/compat.c | 4 +-
xen/arch/x86/guest/xen.c | 25 +++-
xen/arch/x86/pv/hypercall.c | 17 +++
xen/arch/x86/pv/shim.c | 263 ++++++++++++++++++++++++++++++++++++++
xen/common/event_channel.c | 99 ++++++++------
xen/drivers/char/xen_pv_console.c | 11 +-
xen/include/asm-x86/hypercall.h | 3 +
xen/include/asm-x86/pv/shim.h | 5 +
xen/include/xen/event.h | 15 +++
xen/include/xen/pv_console.h | 6 +
10 files changed, 402 insertions(+), 46 deletions(-)
diff --git a/xen/arch/x86/compat.c b/xen/arch/x86/compat.c
index f417cd5034..9d376a4589 100644
--- a/xen/arch/x86/compat.c
+++ b/xen/arch/x86/compat.c
@@ -69,8 +69,8 @@ long do_event_channel_op_compat(XEN_GUEST_HANDLE_PARAM(evtchn_op_t) uop)
case EVTCHNOP_bind_ipi:
case EVTCHNOP_bind_vcpu:
case EVTCHNOP_unmask:
- return do_event_channel_op(op.cmd,
- guest_handle_from_ptr(&uop.p->u, void));
+ return pv_get_hypercall_handler(__HYPERVISOR_event_channel_op, false)
+ (op.cmd, (unsigned long)&uop.p->u, 0, 0, 0, 0);
default:
return -ENOSYS;
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index aff16a0e35..57b297ad47 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -18,6 +18,7 @@
*
* Copyright (c) 2017 Citrix Systems Ltd.
*/
+#include <xen/event.h>
#include <xen/init.h>
#include <xen/mm.h>
#include <xen/pfn.h>
@@ -193,11 +194,31 @@ static void __init init_memmap(void)
static void xen_evtchn_upcall(struct cpu_user_regs *regs)
{
struct vcpu_info *vcpu_info = this_cpu(vcpu_info);
+ unsigned long pending;
vcpu_info->evtchn_upcall_pending = 0;
- write_atomic(&vcpu_info->evtchn_pending_sel, 0);
+ pending = xchg(&vcpu_info->evtchn_pending_sel, 0);
- pv_console_rx(regs);
+ while ( pending )
+ {
+ unsigned int l1 = find_first_set_bit(pending);
+ unsigned long evtchn = xchg(&XEN_shared_info->evtchn_pending[l1], 0);
+
+ __clear_bit(l1, &pending);
+ evtchn &= ~XEN_shared_info->evtchn_mask[l1];
+ while ( evtchn )
+ {
+ unsigned int port = find_first_set_bit(evtchn);
+
+ __clear_bit(port, &evtchn);
+ port += l1 * BITS_PER_LONG;
+
+ if ( pv_console && port == pv_console_evtchn() )
+ pv_console_rx(regs);
+ else if ( pv_shim )
+ pv_shim_inject_evtchn(port);
+ }
+ }
ack_APIC_irq();
}
diff --git a/xen/arch/x86/pv/hypercall.c b/xen/arch/x86/pv/hypercall.c
index f79f7eef62..3b72d6a44d 100644
--- a/xen/arch/x86/pv/hypercall.c
+++ b/xen/arch/x86/pv/hypercall.c
@@ -320,6 +320,23 @@ void hypercall_page_initialise_ring1_kernel(void *hypercall_page)
*(u16 *)(p+ 6) = (HYPERCALL_VECTOR << 8) | 0xcd; /* int $xx */
}
+void __init pv_hypercall_table_replace(unsigned int hypercall,
+ hypercall_fn_t * native,
+ hypercall_fn_t *compat)
+{
+#define HANDLER_POINTER(f) \
+ ((unsigned long *)__va(__pa(&pv_hypercall_table[hypercall].f)))
+ write_atomic(HANDLER_POINTER(native), (unsigned long)native);
+ write_atomic(HANDLER_POINTER(compat), (unsigned long)compat);
+#undef HANDLER_POINTER
+}
+
+hypercall_fn_t *pv_get_hypercall_handler(unsigned int hypercall, bool compat)
+{
+ return compat ? pv_hypercall_table[hypercall].compat
+ : pv_hypercall_table[hypercall].native;
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 78351c9ee0..36f3a366d3 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -18,6 +18,8 @@
*
* Copyright (c) 2017 Citrix Systems Ltd.
*/
+#include <xen/event.h>
+#include <xen/guest_access.h>
#include <xen/hypercall.h>
#include <xen/init.h>
#include <xen/shutdown.h>
@@ -35,6 +37,10 @@ bool pv_shim;
boolean_param("pv-shim", pv_shim);
#endif
+static struct domain *guest;
+
+static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg);
+
#define L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_USER| \
_PAGE_GUEST_KERNEL)
#define COMPAT_L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED)
@@ -63,6 +69,27 @@ static void __init replace_va_mapping(struct domain *d, l4_pgentry_t *l4start,
: COMPAT_L1_PROT));
}
+static void evtchn_reserve(struct domain *d, unsigned int port)
+{
+ ASSERT(port_is_valid(d, port));
+ evtchn_from_port(d, port)->state = ECS_RESERVED;
+ BUG_ON(xen_hypercall_evtchn_unmask(port));
+}
+
+static bool evtchn_handled(struct domain *d, unsigned int port)
+{
+ ASSERT(port_is_valid(d, port));
+ /* The shim manages VIRQs, the rest is forwarded to L0. */
+ return evtchn_from_port(d, port)->state == ECS_VIRQ;
+}
+
+static void evtchn_assign_vcpu(struct domain *d, unsigned int port,
+ unsigned int vcpu)
+{
+ ASSERT(port_is_valid(d, port));
+ evtchn_from_port(d, port)->notify_vcpu_id = vcpu;
+}
+
void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
unsigned long va_start, unsigned long store_va,
unsigned long console_va, unsigned long vphysmap,
@@ -82,6 +109,11 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
replace_va_mapping(d, l4start, va, param); \
dom0_update_physmap(d, PFN_DOWN((va) - va_start), param, vphysmap); \
} \
+ else \
+ { \
+ BUG_ON(evtchn_allocate_port(d, param)); \
+ evtchn_reserve(d, param); \
+ } \
})
SET_AND_MAP_PARAM(HVM_PARAM_STORE_PFN, si->store_mfn, store_va);
SET_AND_MAP_PARAM(HVM_PARAM_STORE_EVTCHN, si->store_evtchn, 0);
@@ -92,6 +124,10 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_EVTCHN, si->console.domU.evtchn, 0);
}
#undef SET_AND_MAP_PARAM
+ pv_hypercall_table_replace(__HYPERVISOR_event_channel_op,
+ (hypercall_fn_t *)pv_shim_event_channel_op,
+ (hypercall_fn_t *)pv_shim_event_channel_op);
+ guest = d;
}
void pv_shim_shutdown(uint8_t reason)
@@ -100,6 +136,233 @@ void pv_shim_shutdown(uint8_t reason)
xen_hypercall_shutdown(reason);
}
+static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
+{
+ struct domain *d = current->domain;
+ struct evtchn_close close;
+ long rc;
+
+ switch ( cmd )
+ {
+#define EVTCHN_FORWARD(cmd, port_field) \
+ case EVTCHNOP_##cmd: { \
+ struct evtchn_##cmd op; \
+ \
+ if ( copy_from_guest(&op, arg, 1) != 0 ) \
+ return -EFAULT; \
+ \
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_##cmd, &op); \
+ if ( rc ) \
+ break; \
+ \
+ spin_lock(&d->event_lock); \
+ rc = evtchn_allocate_port(d, op.port_field); \
+ if ( rc ) \
+ { \
+ close.port = op.port_field; \
+ BUG_ON(xen_hypercall_event_channel_op(EVTCHNOP_close, &close)); \
+ } \
+ else \
+ evtchn_reserve(d, op.port_field); \
+ spin_unlock(&d->event_lock); \
+ \
+ if ( !rc && __copy_to_guest(arg, &op, 1) ) \
+ rc = -EFAULT; \
+ \
+ break; \
+ }
+
+ EVTCHN_FORWARD(alloc_unbound, port)
+ EVTCHN_FORWARD(bind_interdomain, local_port)
+#undef EVTCHN_FORWARD
+
+ case EVTCHNOP_bind_virq: {
+ struct evtchn_bind_virq virq;
+ struct evtchn_alloc_unbound alloc = {
+ .dom = DOMID_SELF,
+ .remote_dom = DOMID_SELF,
+ };
+
+ if ( copy_from_guest(&virq, arg, 1) != 0 )
+ return -EFAULT;
+ /*
+ * The event channel space is actually controlled by L0 Xen, so
+ * allocate a port from L0 and then force the VIRQ to be bound to that
+ * specific port.
+ *
+ * This is only required for VIRQ because the rest of the event channel
+ * operations are handled directly by L0.
+ */
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_alloc_unbound, &alloc);
+ if ( rc )
+ break;
+
+ /* Force L1 to use the event channel port allocated on L0. */
+ rc = evtchn_bind_virq(&virq, alloc.port);
+ if ( rc )
+ {
+ close.port = alloc.port;
+ BUG_ON(xen_hypercall_event_channel_op(EVTCHNOP_close, &close));
+ }
+
+ if ( !rc && __copy_to_guest(arg, &virq, 1) )
+ rc = -EFAULT;
+
+ break;
+ }
+
+ case EVTCHNOP_status: {
+ struct evtchn_status status;
+
+ if ( copy_from_guest(&status, arg, 1) != 0 )
+ return -EFAULT;
+
+ /*
+ * NB: if the event channel is not handled by the shim, just forward
+ * the status request to L0, even if the port is not valid.
+ */
+ if ( port_is_valid(d, status.port) && evtchn_handled(d, status.port) )
+ rc = evtchn_status(&status);
+ else
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_status, &status);
+
+ break;
+ }
+
+ case EVTCHNOP_bind_vcpu: {
+ struct evtchn_bind_vcpu vcpu;
+
+ if ( copy_from_guest(&vcpu, arg, 1) != 0 )
+ return -EFAULT;
+
+ if ( !port_is_valid(d, vcpu.port) )
+ return -EINVAL;
+
+ if ( evtchn_handled(d, vcpu.port) )
+ rc = evtchn_bind_vcpu(vcpu.port, vcpu.vcpu);
+ else
+ {
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_bind_vcpu, &vcpu);
+ if ( !rc )
+ evtchn_assign_vcpu(d, vcpu.port, vcpu.vcpu);
+ }
+
+ break;
+ }
+
+ case EVTCHNOP_close: {
+ if ( copy_from_guest(&close, arg, 1) != 0 )
+ return -EFAULT;
+
+ if ( !port_is_valid(d, close.port) )
+ return -EINVAL;
+
+ set_bit(close.port, XEN_shared_info->evtchn_mask);
+
+ if ( evtchn_handled(d, close.port) )
+ {
+ rc = evtchn_close(d, close.port, true);
+ if ( rc )
+ break;
+ }
+ else
+ evtchn_free(d, evtchn_from_port(d, close.port));
+
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_close, &close);
+ if ( rc )
+ /*
+ * If the port cannot be closed on the L0 mark it as reserved
+ * in the shim to avoid re-using it.
+ */
+ evtchn_reserve(d, close.port);
+
+ break;
+ }
+
+ case EVTCHNOP_bind_ipi: {
+ struct evtchn_bind_ipi ipi;
+
+ if ( copy_from_guest(&ipi, arg, 1) != 0 )
+ return -EFAULT;
+
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_bind_ipi, &ipi);
+ if ( rc )
+ break;
+
+ spin_lock(&d->event_lock);
+ rc = evtchn_allocate_port(d, ipi.port);
+ if ( rc )
+ {
+ spin_unlock(&d->event_lock);
+
+ close.port = ipi.port;
+ BUG_ON(xen_hypercall_event_channel_op(EVTCHNOP_close, &close));
+ break;
+ }
+
+ evtchn_assign_vcpu(d, ipi.port, ipi.vcpu);
+ evtchn_reserve(d, ipi.port);
+ spin_unlock(&d->event_lock);
+
+ if ( __copy_to_guest(arg, &ipi, 1) )
+ rc = -EFAULT;
+
+ break;
+ }
+
+ case EVTCHNOP_unmask: {
+ struct evtchn_unmask unmask;
+
+ if ( copy_from_guest(&unmask, arg, 1) != 0 )
+ return -EFAULT;
+
+ /* Unmask is handled in L1 */
+ rc = evtchn_unmask(unmask.port);
+
+ break;
+ }
+
+ case EVTCHNOP_send: {
+ struct evtchn_send send;
+
+ if ( copy_from_guest(&send, arg, 1) != 0 )
+ return -EFAULT;
+
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_send, &send);
+
+ break;
+ }
+
+ case EVTCHNOP_reset: {
+ struct evtchn_reset reset;
+
+ if ( copy_from_guest(&reset, arg, 1) != 0 )
+ return -EFAULT;
+
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_reset, &reset);
+
+ break;
+ }
+
+ default:
+ /* No FIFO or PIRQ support for now */
+ rc = -EOPNOTSUPP;
+ break;
+ }
+
+ return rc;
+}
+
+void pv_shim_inject_evtchn(unsigned int port)
+{
+ if ( port_is_valid(guest, port) )
+ {
+ struct evtchn *chn = evtchn_from_port(guest, port);
+
+ evtchn_port_set_pending(guest, chn->notify_vcpu_id, chn);
+ }
+}
+
domid_t get_initial_domain_id(void)
{
uint32_t eax, ebx, ecx, edx;
diff --git a/xen/common/event_channel.c b/xen/common/event_channel.c
index c69f9db6db..be834c5c78 100644
--- a/xen/common/event_channel.c
+++ b/xen/common/event_channel.c
@@ -156,46 +156,62 @@ static void free_evtchn_bucket(struct domain *d, struct evtchn *bucket)
xfree(bucket);
}
+int evtchn_allocate_port(struct domain *d, evtchn_port_t port)
+{
+ if ( port > d->max_evtchn_port || port >= d->max_evtchns )
+ return -ENOSPC;
+
+ if ( port_is_valid(d, port) )
+ {
+ if ( evtchn_from_port(d, port)->state != ECS_FREE ||
+ evtchn_port_is_busy(d, port) )
+ return -EBUSY;
+ }
+ else
+ {
+ struct evtchn *chn;
+ struct evtchn **grp;
+
+ if ( !group_from_port(d, port) )
+ {
+ grp = xzalloc_array(struct evtchn *, BUCKETS_PER_GROUP);
+ if ( !grp )
+ return -ENOMEM;
+ group_from_port(d, port) = grp;
+ }
+
+ chn = alloc_evtchn_bucket(d, port);
+ if ( !chn )
+ return -ENOMEM;
+ bucket_from_port(d, port) = chn;
+
+ write_atomic(&d->valid_evtchns, d->valid_evtchns + EVTCHNS_PER_BUCKET);
+ }
+
+ return 0;
+}
+
static int get_free_port(struct domain *d)
{
- struct evtchn *chn;
- struct evtchn **grp;
int port;
if ( d->is_dying )
return -EINVAL;
- for ( port = 0; port_is_valid(d, port); port++ )
+ for ( port = 0; port <= d->max_evtchn_port; port++ )
{
- if ( port > d->max_evtchn_port )
- return -ENOSPC;
- if ( evtchn_from_port(d, port)->state == ECS_FREE
- && !evtchn_port_is_busy(d, port) )
- return port;
- }
+ int rc = evtchn_allocate_port(d, port);
- if ( port == d->max_evtchns || port > d->max_evtchn_port )
- return -ENOSPC;
+ if ( rc == -EBUSY )
+ continue;
- if ( !group_from_port(d, port) )
- {
- grp = xzalloc_array(struct evtchn *, BUCKETS_PER_GROUP);
- if ( !grp )
- return -ENOMEM;
- group_from_port(d, port) = grp;
+ return port;
}
- chn = alloc_evtchn_bucket(d, port);
- if ( !chn )
- return -ENOMEM;
- bucket_from_port(d, port) = chn;
-
- write_atomic(&d->valid_evtchns, d->valid_evtchns + EVTCHNS_PER_BUCKET);
-
- return port;
+ return -ENOSPC;
}
-static void free_evtchn(struct domain *d, struct evtchn *chn)
+void evtchn_free(struct domain *d, struct evtchn *chn)
{
/* Clear pending event to avoid unexpected behavior on re-bind. */
evtchn_port_clear_pending(d, chn);
@@ -345,13 +361,13 @@ static long evtchn_bind_interdomain(evtchn_bind_interdomain_t *bind)
}
-static long evtchn_bind_virq(evtchn_bind_virq_t *bind)
+int evtchn_bind_virq(evtchn_bind_virq_t *bind, evtchn_port_t port)
{
struct evtchn *chn;
struct vcpu *v;
struct domain *d = current->domain;
- int port, virq = bind->virq, vcpu = bind->vcpu;
- long rc = 0;
+ int virq = bind->virq, vcpu = bind->vcpu;
+ int rc = 0;
if ( (virq < 0) || (virq >= ARRAY_SIZE(v->virq_to_evtchn)) )
return -EINVAL;
@@ -368,8 +384,19 @@ static long evtchn_bind_virq(evtchn_bind_virq_t *bind)
if ( v->virq_to_evtchn[virq] != 0 )
ERROR_EXIT(-EEXIST);
- if ( (port = get_free_port(d)) < 0 )
- ERROR_EXIT(port);
+ if ( port != 0 )
+ {
+ if ( (rc = evtchn_allocate_port(d, port)) != 0 )
+ ERROR_EXIT(rc);
+ }
+ else
+ {
+ int alloc_port = get_free_port(d);
+
+ if ( alloc_port < 0 )
+ ERROR_EXIT(alloc_port);
+ port = alloc_port;
+ }
chn = evtchn_from_port(d, port);
@@ -511,7 +538,7 @@ static long evtchn_bind_pirq(evtchn_bind_pirq_t *bind)
}
-static long evtchn_close(struct domain *d1, int port1, bool_t guest)
+int evtchn_close(struct domain *d1, int port1, bool guest)
{
struct domain *d2 = NULL;
struct vcpu *v;
@@ -619,7 +646,7 @@ static long evtchn_close(struct domain *d1, int port1, bool_t guest)
double_evtchn_lock(chn1, chn2);
- free_evtchn(d1, chn1);
+ evtchn_free(d1, chn1);
chn2->state = ECS_UNBOUND;
chn2->u.unbound.remote_domid = d1->domain_id;
@@ -633,7 +660,7 @@ static long evtchn_close(struct domain *d1, int port1, bool_t guest)
}
spin_lock(&chn1->lock);
- free_evtchn(d1, chn1);
+ evtchn_free(d1, chn1);
spin_unlock(&chn1->lock);
out:
@@ -839,7 +866,7 @@ static void clear_global_virq_handlers(struct domain *d)
}
}
-static long evtchn_status(evtchn_status_t *status)
+int evtchn_status(evtchn_status_t *status)
{
struct domain *d;
domid_t dom = status->dom;
@@ -1056,7 +1083,7 @@ long do_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
struct evtchn_bind_virq bind_virq;
if ( copy_from_guest(&bind_virq, arg, 1) != 0 )
return -EFAULT;
- rc = evtchn_bind_virq(&bind_virq);
+ rc = evtchn_bind_virq(&bind_virq, 0);
if ( !rc && __copy_to_guest(arg, &bind_virq, 1) )
rc = -EFAULT; /* Cleaning up here would be a mess! */
break;
diff --git a/xen/drivers/char/xen_pv_console.c b/xen/drivers/char/xen_pv_console.c
index d4f0532101..948343303e 100644
--- a/xen/drivers/char/xen_pv_console.c
+++ b/xen/drivers/char/xen_pv_console.c
@@ -88,6 +88,11 @@ static void notify_daemon(void)
xen_hypercall_evtchn_send(cons_evtchn);
}
+evtchn_port_t pv_console_evtchn(void)
+{
+ return cons_evtchn;
+}
+
size_t pv_console_rx(struct cpu_user_regs *regs)
{
char c;
@@ -97,10 +102,6 @@ size_t pv_console_rx(struct cpu_user_regs *regs)
if ( !cons_ring )
return 0;
- /* TODO: move this somewhere */
- if ( !test_bit(cons_evtchn, XEN_shared_info->evtchn_pending) )
- return 0;
-
prod = ACCESS_ONCE(cons_ring->in_prod);
cons = cons_ring->in_cons;
@@ -125,8 +126,6 @@ size_t pv_console_rx(struct cpu_user_regs *regs)
ACCESS_ONCE(cons_ring->in_cons) = cons;
notify_daemon();
- clear_bit(cons_evtchn, XEN_shared_info->evtchn_pending);
-
return recv;
}
diff --git a/xen/include/asm-x86/hypercall.h b/xen/include/asm-x86/hypercall.h
index 3eb4a8db89..b9f3ecf9a3 100644
--- a/xen/include/asm-x86/hypercall.h
+++ b/xen/include/asm-x86/hypercall.h
@@ -28,6 +28,9 @@ extern const hypercall_args_t hypercall_args_table[NR_hypercalls];
void pv_hypercall(struct cpu_user_regs *regs);
void hypercall_page_initialise_ring3_kernel(void *hypercall_page);
void hypercall_page_initialise_ring1_kernel(void *hypercall_page);
+void pv_hypercall_table_replace(unsigned int hypercall, hypercall_fn_t * native,
+ hypercall_fn_t *compat);
+hypercall_fn_t *pv_get_hypercall_handler(unsigned int hypercall, bool compat);
/*
* Both do_mmuext_op() and do_mmu_update():
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index ff7c050dc6..ab656fd854 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -36,6 +36,7 @@ void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
unsigned long console_va, unsigned long vphysmap,
start_info_t *si);
void pv_shim_shutdown(uint8_t reason);
+void pv_shim_inject_evtchn(unsigned int port);
domid_t get_initial_domain_id(void);
#else
@@ -53,6 +54,10 @@ static inline void pv_shim_shutdown(uint8_t reason)
{
ASSERT_UNREACHABLE();
}
+static inline void pv_shim_inject_evtchn(unsigned int port)
+{
+ ASSERT_UNREACHABLE();
+}
static inline domid_t get_initial_domain_id(void)
{
return 0;
diff --git a/xen/include/xen/event.h b/xen/include/xen/event.h
index 87915ead69..ebb879e88d 100644
--- a/xen/include/xen/event.h
+++ b/xen/include/xen/event.h
@@ -48,6 +48,21 @@ int evtchn_send(struct domain *d, unsigned int lport);
/* Bind a local event-channel port to the specified VCPU. */
long evtchn_bind_vcpu(unsigned int port, unsigned int vcpu_id);
+/* Bind a VIRQ. */
+int evtchn_bind_virq(evtchn_bind_virq_t *bind, evtchn_port_t port);
+
+/* Get the status of an event channel port. */
+int evtchn_status(evtchn_status_t *status);
+
+/* Close an event channel. */
+int evtchn_close(struct domain *d1, int port1, bool guest);
+
+/* Free an event channel. */
+void evtchn_free(struct domain *d, struct evtchn *chn);
+
+/* Allocate a specific event channel port. */
+int evtchn_allocate_port(struct domain *d, unsigned int port);
+
/* Unmask a local event-channel port. */
int evtchn_unmask(unsigned int port);
diff --git a/xen/include/xen/pv_console.h b/xen/include/xen/pv_console.h
index e578b56620..cb92539666 100644
--- a/xen/include/xen/pv_console.h
+++ b/xen/include/xen/pv_console.h
@@ -10,6 +10,7 @@ void pv_console_set_rx_handler(serial_rx_fn fn);
void pv_console_init_postirq(void);
void pv_console_puts(const char *buf);
size_t pv_console_rx(struct cpu_user_regs *regs);
+evtchn_port_t pv_console_evtchn(void);
#else
@@ -18,6 +19,11 @@ static inline void pv_console_set_rx_handler(serial_rx_fn fn) { }
static inline void pv_console_init_postirq(void) { }
static inline void pv_console_puts(const char *buf) { }
static inline size_t pv_console_rx(struct cpu_user_regs *regs) { return 0; }
+evtchn_port_t pv_console_evtchn(void)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+}
#endif /* !CONFIG_XEN_GUEST */
#endif /* __XEN_PV_CONSOLE_H__ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 48/62] xen/pvshim: add grant table operations
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (46 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 47/62] xen/pvshim: forward evtchn ops between L0 Xen and L2 DomU Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 49/62] x86/pv-shim: shadow PV console's page for L2 DomU Wei Liu
` (13 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Anthony Liguori <aliguori@amazon.com>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
Changes since v1:
- Use __ of copy_to_guest.
- Return EOPNOTSUPP for not implemented grant table hypercalls.
- Forward user provided buffer in GNTTABOP_query_size.
- Rewrite grant table hypercall handler.
---
xen/arch/x86/pv/shim.c | 164 ++++++++++++++++++++++++++++++++++
xen/include/asm-x86/guest/hypercall.h | 6 ++
2 files changed, 170 insertions(+)
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 36f3a366d3..eb8b146785 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -22,6 +22,7 @@
#include <xen/guest_access.h>
#include <xen/hypercall.h>
#include <xen/init.h>
+#include <xen/iocap.h>
#include <xen/shutdown.h>
#include <xen/types.h>
@@ -32,6 +33,8 @@
#include <public/arch-x86/cpuid.h>
+#include <compat/grant_table.h>
+
#ifndef CONFIG_PV_SHIM_EXCLUSIVE
bool pv_shim;
boolean_param("pv-shim", pv_shim);
@@ -39,7 +42,14 @@ boolean_param("pv-shim", pv_shim);
static struct domain *guest;
+static unsigned int nr_grant_list;
+static unsigned long *grant_frames;
+static DEFINE_SPINLOCK(grant_lock);
+
static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg);
+static long pv_shim_grant_table_op(unsigned int cmd,
+ XEN_GUEST_HANDLE_PARAM(void) uop,
+ unsigned int count);
#define L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_USER| \
_PAGE_GUEST_KERNEL)
@@ -127,6 +137,9 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
pv_hypercall_table_replace(__HYPERVISOR_event_channel_op,
(hypercall_fn_t *)pv_shim_event_channel_op,
(hypercall_fn_t *)pv_shim_event_channel_op);
+ pv_hypercall_table_replace(__HYPERVISOR_grant_table_op,
+ (hypercall_fn_t *)pv_shim_grant_table_op,
+ (hypercall_fn_t *)pv_shim_grant_table_op);
guest = d;
}
@@ -363,6 +376,157 @@ void pv_shim_inject_evtchn(unsigned int port)
}
}
+static long pv_shim_grant_table_op(unsigned int cmd,
+ XEN_GUEST_HANDLE_PARAM(void) uop,
+ unsigned int count)
+{
+ struct domain *d = current->domain;
+ long rc = 0;
+
+ if ( count != 1 )
+ return -EINVAL;
+
+ switch ( cmd )
+ {
+ case GNTTABOP_setup_table:
+ {
+ bool compat = is_pv_32bit_domain(d);
+ struct gnttab_setup_table nat;
+ struct compat_gnttab_setup_table cmp;
+ unsigned int i;
+
+ if ( unlikely(compat ? copy_from_guest(&cmp, uop, 1)
+ : copy_from_guest(&nat, uop, 1)) ||
+ unlikely(compat ? !compat_handle_okay(cmp.frame_list,
+ cmp.nr_frames)
+ : !guest_handle_okay(nat.frame_list,
+ nat.nr_frames)) )
+ {
+ rc = -EFAULT;
+ break;
+ }
+ if ( compat )
+#define XLAT_gnttab_setup_table_HNDL_frame_list(d, s)
+ XLAT_gnttab_setup_table(&nat, &cmp);
+#undef XLAT_gnttab_setup_table_HNDL_frame_list
+
+ nat.status = GNTST_okay;
+
+ spin_lock(&grant_lock);
+ if ( !nr_grant_list )
+ {
+ struct gnttab_query_size query_size = {
+ .dom = DOMID_SELF,
+ };
+
+ rc = xen_hypercall_grant_table_op(GNTTABOP_query_size,
+ &query_size, 1);
+ if ( rc )
+ {
+ spin_unlock(&grant_lock);
+ break;
+ }
+
+ ASSERT(!grant_frames);
+ grant_frames = xzalloc_array(unsigned long,
+ query_size.max_nr_frames);
+ if ( !grant_frames )
+ {
+ spin_unlock(&grant_lock);
+ rc = -ENOMEM;
+ break;
+ }
+
+ nr_grant_list = query_size.max_nr_frames;
+ }
+
+ if ( nat.nr_frames > nr_grant_list )
+ {
+ spin_unlock(&grant_lock);
+ rc = -EINVAL;
+ break;
+ }
+
+ for ( i = 0; i < nat.nr_frames; i++ )
+ {
+ if ( !grant_frames[i] )
+ {
+ struct xen_add_to_physmap xatp = {
+ .domid = DOMID_SELF,
+ .idx = i,
+ .space = XENMAPSPACE_grant_table,
+ };
+ mfn_t mfn;
+
+ rc = hypervisor_alloc_unused_page(&mfn);
+ if ( rc )
+ {
+ gprintk(XENLOG_ERR,
+ "unable to get memory for grant table\n");
+ break;
+ }
+
+ xatp.gpfn = mfn_x(mfn);
+ rc = xen_hypercall_memory_op(XENMEM_add_to_physmap, &xatp);
+ if ( rc )
+ {
+ hypervisor_free_unused_page(mfn);
+ break;
+ }
+
+ BUG_ON(iomem_permit_access(d, mfn_x(mfn), mfn_x(mfn)));
+ grant_frames[i] = mfn_x(mfn);
+ }
+
+ ASSERT(grant_frames[i]);
+ if ( compat )
+ {
+ compat_pfn_t pfn = grant_frames[i];
+
+ if ( __copy_to_compat_offset(cmp.frame_list, i, &pfn, 1) )
+ {
+ nat.status = GNTST_bad_virt_addr;
+ rc = -EFAULT;
+ break;
+ }
+ }
+ else if ( __copy_to_guest_offset(nat.frame_list, i,
+ &grant_frames[i], 1) )
+ {
+ nat.status = GNTST_bad_virt_addr;
+ rc = -EFAULT;
+ break;
+ }
+ }
+ spin_unlock(&grant_lock);
+
+ if ( compat )
+#define XLAT_gnttab_setup_table_HNDL_frame_list(d, s)
+ XLAT_gnttab_setup_table(&cmp, &nat);
+#undef XLAT_gnttab_setup_table_HNDL_frame_list
+
+ if ( unlikely(compat ? __copy_to_guest(uop, &cmp, 1)
+ : __copy_to_guest(uop, &nat, 1)) )
+ {
+ rc = -EFAULT;
+ break;
+ }
+
+ break;
+ }
+
+ case GNTTABOP_query_size:
+ rc = xen_hypercall_grant_table_op(GNTTABOP_query_size, uop.p, count);
+ break;
+
+ default:
+ rc = -EOPNOTSUPP;
+ break;
+ }
+
+ return rc;
+}
+
domid_t get_initial_domain_id(void)
{
uint32_t eax, ebx, ecx, edx;
diff --git a/xen/include/asm-x86/guest/hypercall.h b/xen/include/asm-x86/guest/hypercall.h
index 81a955d479..e9e626b474 100644
--- a/xen/include/asm-x86/guest/hypercall.h
+++ b/xen/include/asm-x86/guest/hypercall.h
@@ -110,6 +110,12 @@ static inline long xen_hypercall_event_channel_op(unsigned int cmd, void *arg)
return _hypercall64_2(long, __HYPERVISOR_event_channel_op, cmd, arg);
}
+static inline long xen_hypercall_grant_table_op(unsigned int cmd, void *arg,
+ unsigned int count)
+{
+ return _hypercall64_3(long, __HYPERVISOR_grant_table_op, cmd, arg, count);
+}
+
static inline long xen_hypercall_hvm_op(unsigned int op, void *arg)
{
return _hypercall64_2(long, __HYPERVISOR_hvm_op, op, arg);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 49/62] x86/pv-shim: shadow PV console's page for L2 DomU
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (47 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 48/62] xen/pvshim: add grant table operations Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 50/62] xen/pvshim: add migration support Wei Liu
` (12 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
[remove notify_guest helper and directly use pv_shim_inject_evtchn]
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
Changes since v1:
- Use pv_shim_inject_evtchn.
---
xen/arch/x86/pv/shim.c | 31 ++++++++--
xen/drivers/char/Makefile | 1 +
xen/drivers/char/console.c | 6 ++
xen/drivers/char/consoled.c | 148 ++++++++++++++++++++++++++++++++++++++++++++
xen/include/xen/consoled.h | 27 ++++++++
5 files changed, 209 insertions(+), 4 deletions(-)
create mode 100644 xen/drivers/char/consoled.c
create mode 100644 xen/include/xen/consoled.h
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index eb8b146785..986f9da58a 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -25,6 +25,8 @@
#include <xen/iocap.h>
#include <xen/shutdown.h>
#include <xen/types.h>
+#include <xen/consoled.h>
+#include <xen/pv_console.h>
#include <asm/apic.h>
#include <asm/dom0_build.h>
@@ -127,13 +129,28 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
})
SET_AND_MAP_PARAM(HVM_PARAM_STORE_PFN, si->store_mfn, store_va);
SET_AND_MAP_PARAM(HVM_PARAM_STORE_EVTCHN, si->store_evtchn, 0);
+ SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_EVTCHN, si->console.domU.evtchn, 0);
if ( !pv_console )
- {
SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_PFN, si->console.domU.mfn,
console_va);
- SET_AND_MAP_PARAM(HVM_PARAM_CONSOLE_EVTCHN, si->console.domU.evtchn, 0);
- }
#undef SET_AND_MAP_PARAM
+ else
+ {
+ /* Allocate a new page for DomU's PV console */
+ void *page = alloc_xenheap_pages(0, MEMF_bits(32));
+ uint64_t console_mfn;
+
+ ASSERT(page);
+ clear_page(page);
+ console_mfn = virt_to_mfn(page);
+ si->console.domU.mfn = console_mfn;
+ share_xen_page_with_guest(mfn_to_page(console_mfn), d,
+ XENSHARE_writable);
+ replace_va_mapping(d, l4start, console_va, console_mfn);
+ dom0_update_physmap(d, (console_va - va_start) >> PAGE_SHIFT,
+ console_mfn, vphysmap);
+ consoled_set_ring_addr(page);
+ }
pv_hypercall_table_replace(__HYPERVISOR_event_channel_op,
(hypercall_fn_t *)pv_shim_event_channel_op,
(hypercall_fn_t *)pv_shim_event_channel_op);
@@ -341,7 +358,13 @@ static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
if ( copy_from_guest(&send, arg, 1) != 0 )
return -EFAULT;
- rc = xen_hypercall_event_channel_op(EVTCHNOP_send, &send);
+ if ( pv_console && send.port == pv_console_evtchn() )
+ {
+ consoled_guest_rx();
+ rc = 0;
+ }
+ else
+ rc = xen_hypercall_event_channel_op(EVTCHNOP_send, &send);
break;
}
diff --git a/xen/drivers/char/Makefile b/xen/drivers/char/Makefile
index 9d48d0f2dc..0d48b16e8d 100644
--- a/xen/drivers/char/Makefile
+++ b/xen/drivers/char/Makefile
@@ -9,3 +9,4 @@ obj-$(CONFIG_HAS_EHCI) += ehci-dbgp.o
obj-$(CONFIG_ARM) += arm-uart.o
obj-y += serial.o
obj-$(CONFIG_XEN_GUEST) += xen_pv_console.o
+obj-$(CONFIG_PV_SHIM) += consoled.o
diff --git a/xen/drivers/char/console.c b/xen/drivers/char/console.c
index 8acd358395..18f5b7f7aa 100644
--- a/xen/drivers/char/console.c
+++ b/xen/drivers/char/console.c
@@ -32,6 +32,7 @@
#include <xen/warning.h>
#ifdef CONFIG_X86
+#include <xen/consoled.h>
#include <xen/pv_console.h>
#include <asm/guest.h>
#endif
@@ -415,6 +416,11 @@ static void __serial_rx(char c, struct cpu_user_regs *regs)
serial_rx_ring[SERIAL_RX_MASK(serial_rx_prod++)] = c;
/* Always notify the guest: prevents receive path from getting stuck. */
send_global_virq(VIRQ_CONSOLE);
+
+#ifdef CONFIG_X86
+ if ( pv_shim && pv_console )
+ consoled_guest_tx(c);
+#endif
}
static void serial_rx(char c, struct cpu_user_regs *regs)
diff --git a/xen/drivers/char/consoled.c b/xen/drivers/char/consoled.c
new file mode 100644
index 0000000000..552abf5766
--- /dev/null
+++ b/xen/drivers/char/consoled.c
@@ -0,0 +1,148 @@
+/******************************************************************************
+ * drivers/char/consoled.c
+ *
+ * A backend driver for Xen's PV console.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (c) 2017 Citrix Systems Ltd.
+ */
+
+#include <xen/lib.h>
+#include <xen/event.h>
+#include <xen/pv_console.h>
+#include <xen/consoled.h>
+
+#include <asm/guest.h>
+
+static struct xencons_interface *cons_ring;
+static DEFINE_SPINLOCK(rx_lock);
+
+void consoled_set_ring_addr(struct xencons_interface *ring)
+{
+ cons_ring = ring;
+}
+
+struct xencons_interface *consoled_get_ring_addr(void)
+{
+ return cons_ring;
+}
+
+#define BUF_SZ 255
+static char buf[BUF_SZ + 1];
+
+/* Receives characters from a domain's PV console */
+size_t consoled_guest_rx(void)
+{
+ size_t recv = 0, idx = 0;
+ XENCONS_RING_IDX cons, prod;
+
+ if ( !cons_ring )
+ return 0;
+
+ spin_lock(&rx_lock);
+
+ cons = cons_ring->out_cons;
+ prod = ACCESS_ONCE(cons_ring->out_prod);
+
+ /*
+ * Latch pointers before accessing the ring. Included compiler barrier also
+ * ensures that pointers are really read only once into local variables.
+ */
+ smp_rmb();
+
+ ASSERT((prod - cons) <= sizeof(cons_ring->out));
+
+ /* Is the ring empty? */
+ if ( cons == prod )
+ goto out;
+
+ while ( cons != prod )
+ {
+ char c = cons_ring->out[MASK_XENCONS_IDX(cons++, cons_ring->out)];
+
+ buf[idx++] = c;
+ recv++;
+
+ if ( idx >= BUF_SZ )
+ {
+ pv_console_puts(buf);
+ idx = 0;
+ }
+ }
+
+ if ( idx )
+ {
+ buf[idx] = '\0';
+ pv_console_puts(buf);
+ }
+
+ /* No need for a mem barrier because every character was already consumed */
+ barrier();
+ ACCESS_ONCE(cons_ring->out_cons) = cons;
+ pv_shim_inject_evtchn(pv_console_evtchn());
+
+ out:
+ spin_unlock(&rx_lock);
+
+ return recv;
+}
+
+/* Sends a character into a domain's PV console */
+size_t consoled_guest_tx(char c)
+{
+ size_t sent = 0;
+ XENCONS_RING_IDX cons, prod;
+
+ if ( !cons_ring )
+ return 0;
+
+ cons = ACCESS_ONCE(cons_ring->in_cons);
+ prod = cons_ring->in_prod;
+
+ /*
+ * Latch pointers before accessing the ring. Included compiler barrier also
+ * ensures that pointers are really read only once into local variables.
+ */
+ smp_rmb();
+
+ ASSERT((prod - cons) <= sizeof(cons_ring->in));
+
+ /* Is the ring out of space? */
+ if ( sizeof(cons_ring->in) - (prod - cons) == 0 )
+ goto notify;
+
+ cons_ring->in[MASK_XENCONS_IDX(prod++, cons_ring->in)] = c;
+ sent++;
+
+ /* Write to the ring before updating the pointer */
+ smp_wmb();
+ ACCESS_ONCE(cons_ring->in_prod) = prod;
+
+ notify:
+ /* Always notify the guest: prevents receive path from getting stuck. */
+ pv_shim_inject_evtchn(pv_console_evtchn());
+
+ return sent;
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
diff --git a/xen/include/xen/consoled.h b/xen/include/xen/consoled.h
new file mode 100644
index 0000000000..fd5d220a8a
--- /dev/null
+++ b/xen/include/xen/consoled.h
@@ -0,0 +1,27 @@
+#ifndef __XEN_CONSOLED_H__
+#define __XEN_CONSOLED_H__
+
+#include <public/io/console.h>
+
+#ifdef CONFIG_PV_SHIM
+
+void consoled_set_ring_addr(struct xencons_interface *ring);
+struct xencons_interface *consoled_get_ring_addr(void);
+size_t consoled_guest_rx(void);
+size_t consoled_guest_tx(char c);
+
+#else
+
+size_t consoled_guest_tx(char c) { return 0; }
+
+#endif /* !CONFIG_PV_SHIM */
+#endif /* __XEN_CONSOLED_H__ */
+/*
+ * Local variables:
+ * mode: C
+ * c-file-style: "BSD"
+ * c-basic-offset: 4
+ * tab-width: 4
+ * indent-tabs-mode: nil
+ * End:
+ */
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 50/62] xen/pvshim: add migration support
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (48 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 49/62] x86/pv-shim: shadow PV console's page for L2 DomU Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:28 ` [PATCH v2 51/62] xen/pvshim: add shim_mem cmdline parameter Wei Liu
` (11 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Use bitmap_zero instead of memset.
- Don't drop the __init attribute of unshare_xen_page_with_guest,
it's not needed for migration.
- Remove BUG_ON to check correct mapping, map_domain_page cannot
fail.
- Reduce indentation level of pv_shim_shutdown.
---
xen/arch/x86/guest/xen.c | 29 +++++++
xen/arch/x86/pv/shim.c | 155 +++++++++++++++++++++++++++++++++++++-
xen/common/domain.c | 11 ++-
xen/common/schedule.c | 3 +-
xen/drivers/char/xen_pv_console.c | 2 +-
xen/include/asm-x86/guest/xen.h | 5 ++
xen/include/asm-x86/pv/shim.h | 5 +-
xen/include/xen/sched.h | 2 +-
8 files changed, 197 insertions(+), 15 deletions(-)
diff --git a/xen/arch/x86/guest/xen.c b/xen/arch/x86/guest/xen.c
index 57b297ad47..2a5554ab26 100644
--- a/xen/arch/x86/guest/xen.c
+++ b/xen/arch/x86/guest/xen.c
@@ -348,6 +348,35 @@ uint32_t hypervisor_cpuid_base(void)
return xen_cpuid_base;
}
+static void ap_resume(void *unused)
+{
+ map_vcpuinfo();
+ init_evtchn();
+}
+
+void hypervisor_resume(void)
+{
+ /* Reset shared info page. */
+ map_shared_info();
+
+ /*
+ * Reset vcpu_info. Just clean the mapped bitmap and try to map the vcpu
+ * area again. On failure to map (when it was previously mapped) panic
+ * since it's impossible to safely shut down running guest vCPUs in order
+ * to meet the new XEN_LEGACY_MAX_VCPUS requirement.
+ */
+ bitmap_zero(vcpu_info_mapped, NR_CPUS);
+ if ( map_vcpuinfo() && nr_cpu_ids > XEN_LEGACY_MAX_VCPUS )
+ panic("unable to remap vCPU info and vCPUs > legacy limit");
+
+ /* Setup event channel upcall vector. */
+ init_evtchn();
+ smp_call_function(ap_resume, NULL, 1);
+
+ if ( pv_console )
+ pv_console_init();
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 986f9da58a..c53a4ca407 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -160,10 +160,159 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
guest = d;
}
-void pv_shim_shutdown(uint8_t reason)
+static void write_start_info(struct domain *d)
{
- /* XXX: handle suspend */
- xen_hypercall_shutdown(reason);
+ struct cpu_user_regs *regs = guest_cpu_user_regs();
+ start_info_t *si = map_domain_page(_mfn(is_pv_32bit_domain(d) ? regs->edx
+ : regs->rdx));
+ uint64_t param;
+
+ snprintf(si->magic, sizeof(si->magic), "xen-3.0-x86_%s",
+ is_pv_32bit_domain(d) ? "32p" : "64");
+ si->nr_pages = d->tot_pages;
+ si->shared_info = virt_to_maddr(d->shared_info);
+ si->flags = 0;
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_STORE_PFN, &si->store_mfn));
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_STORE_EVTCHN, ¶m));
+ si->store_evtchn = param;
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_EVTCHN, ¶m));
+ si->console.domU.evtchn = param;
+ if ( pv_console )
+ si->console.domU.mfn = virt_to_mfn(consoled_get_ring_addr());
+ else if ( xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_PFN,
+ &si->console.domU.mfn) )
+ BUG();
+
+ if ( is_pv_32bit_domain(d) )
+ xlat_start_info(si, XLAT_start_info_console_domU);
+
+ unmap_domain_page(si);
+}
+
+int pv_shim_shutdown(uint8_t reason)
+{
+ struct domain *d = current->domain;
+ struct vcpu *v;
+ unsigned int i;
+ uint64_t old_store_pfn, old_console_pfn = 0, store_pfn, console_pfn;
+ uint64_t store_evtchn, console_evtchn;
+ long rc;
+
+ if ( reason != SHUTDOWN_suspend )
+ /* Forward to L0. */
+ return xen_hypercall_shutdown(reason);
+
+ BUG_ON(current->vcpu_id != 0);
+
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_STORE_PFN, &old_store_pfn));
+ if ( !pv_console )
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_PFN,
+ &old_console_pfn));
+
+ /* Pause the other vcpus before starting the migration. */
+ for_each_vcpu(d, v)
+ if ( v != current )
+ vcpu_pause_by_systemcontroller(v);
+
+ rc = xen_hypercall_shutdown(SHUTDOWN_suspend);
+ if ( rc )
+ {
+ for_each_vcpu(d, v)
+ if ( v != current )
+ vcpu_unpause_by_systemcontroller(v);
+
+ return rc;
+ }
+
+ /* Resume the shim itself first. */
+ hypervisor_resume();
+
+ /*
+ * ATM there's nothing Xen can do if the console/store pfn changes,
+ * because Xen won't have a page_info struct for it.
+ */
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_STORE_PFN, &store_pfn));
+ BUG_ON(old_store_pfn != store_pfn);
+ if ( !pv_console )
+ {
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_PFN,
+ &console_pfn));
+ BUG_ON(old_console_pfn != console_pfn);
+ }
+
+ /* Update domain id. */
+ d->domain_id = get_initial_domain_id();
+
+ /* Clean the iomem range. */
+ BUG_ON(iomem_deny_access(d, 0, ~0UL));
+
+ /* Clean grant frames. */
+ xfree(grant_frames);
+ grant_frames = NULL;
+ nr_grant_list = 0;
+
+ /* Clean event channels. */
+ for ( i = 0; i < EVTCHN_2L_NR_CHANNELS; i++ )
+ {
+ if ( !port_is_valid(d, i) )
+ continue;
+
+ if ( evtchn_handled(d, i) )
+ evtchn_close(d, i, false);
+ else
+ evtchn_free(d, evtchn_from_port(d, i));
+ }
+
+ /* Reserve store/console event channel. */
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_STORE_EVTCHN, &store_evtchn));
+ BUG_ON(evtchn_allocate_port(d, store_evtchn));
+ evtchn_reserve(d, store_evtchn);
+ BUG_ON(xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_EVTCHN,
+ &console_evtchn));
+ BUG_ON(evtchn_allocate_port(d, console_evtchn));
+ evtchn_reserve(d, console_evtchn);
+
+ /* Clean watchdogs. */
+ watchdog_domain_destroy(d);
+ watchdog_domain_init(d);
+
+ /* Clean the PIRQ EOI page. */
+ if ( d->arch.pirq_eoi_map != NULL )
+ {
+ unmap_domain_page_global(d->arch.pirq_eoi_map);
+ put_page_and_type(mfn_to_page(d->arch.pirq_eoi_map_mfn));
+ d->arch.pirq_eoi_map = NULL;
+ d->arch.pirq_eoi_map_mfn = 0;
+ d->arch.auto_unmask = 0;
+ }
+
+ /*
+ * NB: there's no need to fixup the p2m, since the mfns assigned
+ * to the PV guest have not changed at all. Just re-write the
+ * start_info fields with the appropriate value.
+ */
+ write_start_info(d);
+
+ for_each_vcpu(d, v)
+ {
+ /* Unmap guest vcpu_info pages. */
+ unmap_vcpu_info(v);
+
+ /* Reset the periodic timer to the default value. */
+ v->periodic_period = MILLISECS(10);
+ /* Stop the singleshot timer. */
+ stop_timer(&v->singleshot_timer);
+
+ if ( test_bit(_VPF_down, &v->pause_flags) )
+ BUG_ON(vcpu_reset(v));
+
+ if ( v != current )
+ vcpu_unpause_by_systemcontroller(v);
+ else
+ vcpu_force_reschedule(v);
+ }
+
+ return 0;
}
static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
diff --git a/xen/common/domain.c b/xen/common/domain.c
index ef566f7f45..6ed21b2536 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -697,16 +697,13 @@ void __domain_crash_synchronous(void)
}
-void domain_shutdown(struct domain *d, u8 reason)
+int domain_shutdown(struct domain *d, u8 reason)
{
struct vcpu *v;
#ifdef CONFIG_X86
if ( pv_shim )
- {
- pv_shim_shutdown(reason);
- return;
- }
+ return pv_shim_shutdown(reason);
#endif
spin_lock(&d->shutdown_lock);
@@ -721,7 +718,7 @@ void domain_shutdown(struct domain *d, u8 reason)
if ( d->is_shutting_down )
{
spin_unlock(&d->shutdown_lock);
- return;
+ return 0;
}
d->is_shutting_down = 1;
@@ -743,6 +740,8 @@ void domain_shutdown(struct domain *d, u8 reason)
__domain_finalise_shutdown(d);
spin_unlock(&d->shutdown_lock);
+
+ return 0;
}
void domain_resume(struct domain *d)
diff --git a/xen/common/schedule.c b/xen/common/schedule.c
index 88279213e8..b7884263f2 100644
--- a/xen/common/schedule.c
+++ b/xen/common/schedule.c
@@ -1149,11 +1149,10 @@ ret_t do_sched_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
if ( copy_from_guest(&sched_shutdown, arg, 1) )
break;
- ret = 0;
TRACE_3D(TRC_SCHED_SHUTDOWN,
current->domain->domain_id, current->vcpu_id,
sched_shutdown.reason);
- domain_shutdown(current->domain, (u8)sched_shutdown.reason);
+ ret = domain_shutdown(current->domain, (u8)sched_shutdown.reason);
break;
}
diff --git a/xen/drivers/char/xen_pv_console.c b/xen/drivers/char/xen_pv_console.c
index 948343303e..cc1c1d743f 100644
--- a/xen/drivers/char/xen_pv_console.c
+++ b/xen/drivers/char/xen_pv_console.c
@@ -37,7 +37,7 @@ static DEFINE_SPINLOCK(tx_lock);
bool pv_console;
-void __init pv_console_init(void)
+void pv_console_init(void)
{
long r;
uint64_t raw_pfn = 0, raw_evtchn = 0;
diff --git a/xen/include/asm-x86/guest/xen.h b/xen/include/asm-x86/guest/xen.h
index ac48dcbe44..11243fe60d 100644
--- a/xen/include/asm-x86/guest/xen.h
+++ b/xen/include/asm-x86/guest/xen.h
@@ -39,6 +39,7 @@ int hypervisor_free_unused_page(mfn_t mfn);
void hypervisor_fixup_e820(struct e820map *e820);
const unsigned long *hypervisor_reserved_pages(unsigned int *size);
uint32_t hypervisor_cpuid_base(void);
+void hypervisor_resume(void);
DECLARE_PER_CPU(unsigned int, vcpu_id);
DECLARE_PER_CPU(struct vcpu_info *, vcpu_info);
@@ -72,6 +73,10 @@ static inline uint32_t hypervisor_cpuid_base(void)
ASSERT_UNREACHABLE();
return 0;
};
+static inline void hypervisor_resume(void)
+{
+ ASSERT_UNREACHABLE();
+};
#endif /* CONFIG_XEN_GUEST */
#endif /* __X86_GUEST_XEN_H__ */
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index ab656fd854..4d5f0b43fc 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -35,7 +35,7 @@ void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
unsigned long va_start, unsigned long store_va,
unsigned long console_va, unsigned long vphysmap,
start_info_t *si);
-void pv_shim_shutdown(uint8_t reason);
+int pv_shim_shutdown(uint8_t reason);
void pv_shim_inject_evtchn(unsigned int port);
domid_t get_initial_domain_id(void);
@@ -50,9 +50,10 @@ static inline void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
{
ASSERT_UNREACHABLE();
}
-static inline void pv_shim_shutdown(uint8_t reason)
+static inline int pv_shim_shutdown(uint8_t reason)
{
ASSERT_UNREACHABLE();
+ return 0;
}
static inline void pv_shim_inject_evtchn(unsigned int port)
{
diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h
index 64abc1df6c..2541ecb04f 100644
--- a/xen/include/xen/sched.h
+++ b/xen/include/xen/sched.h
@@ -603,7 +603,7 @@ static inline struct domain *rcu_lock_current_domain(void)
struct domain *get_domain_by_id(domid_t dom);
void domain_destroy(struct domain *d);
int domain_kill(struct domain *d);
-void domain_shutdown(struct domain *d, u8 reason);
+int domain_shutdown(struct domain *d, u8 reason);
void domain_resume(struct domain *d);
void domain_pause_for_debugger(void);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 51/62] xen/pvshim: add shim_mem cmdline parameter
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (49 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 50/62] xen/pvshim: add migration support Wei Liu
@ 2018-01-12 11:28 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 52/62] xen/pvshim: set max_pages to the value of tot_pages Wei Liu
` (10 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:28 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Sergey Dyasli <sergey.dyasli@citrix.com>
Signed-off-by: Sergey Dyasli <sergey.dyasli@citrix.com>
---
docs/misc/xen-command-line.markdown | 16 +++++++++++++
xen/arch/x86/dom0_build.c | 18 ++++++++++++++-
xen/arch/x86/pv/shim.c | 46 +++++++++++++++++++++++++++++++++++++
xen/include/asm-x86/pv/shim.h | 6 +++++
4 files changed, 85 insertions(+), 1 deletion(-)
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index 3a1a9c1fba..9f51710a46 100644
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -686,6 +686,8 @@ any dom0 autoballooning feature present in your toolstack. See the
_xl.conf(5)_ man page or [Xen Best
Practices](http://wiki.xen.org/wiki/Xen_Best_Practices#Xen_dom0_dedicated_memory_and_preventing_dom0_memory_ballooning).
+This option doesn't have effect if pv-shim mode is enabled.
+
### dom0\_nodes
> `= List of [ <integer> | relaxed | strict ]`
@@ -1456,6 +1458,20 @@ guest compatibly inside an HVM container.
In this mode, the kernel and initrd passed as modules to the hypervisor are
constructed into a plain unprivileged PV domain.
+### shim\_mem (x86)
+> `= List of ( min:<size> | max:<size> | <size> )`
+
+Set the amount of memory that xen-shim reserves for itself. Only has effect
+if pv-shim mode is enabled.
+
+* `min:<size>` specifies the minimum amount of memory. Ignored if greater
+ than max. Default: 10M.
+* `max:<size>` specifies the maximum amount of memory. Default: 128M.
+* `<size>` specifies the exact amount of memory. Overrides both min and max.
+
+By default, 1/16th of total HVM container's memory is reserved for xen-shim
+with minimum amount being 10MB and maximum amount 128MB.
+
### rcu-idle-timer-period-ms
> `= <integer>`
diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 8600123516..2aac0803fc 100644
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -51,6 +51,13 @@ static long __init parse_amt(const char *s, const char **ps)
static int __init parse_dom0_mem(const char *s)
{
+ /* xen-shim uses shim_mem parameter instead of dom0_mem */
+ if ( pv_shim )
+ {
+ printk("Ignoring dom0_mem param in pv-shim mode\n");
+ return 0;
+ }
+
do {
if ( !strncmp(s, "min:", 4) )
dom0_min_nrpages = parse_amt(s+4, &s);
@@ -284,7 +291,16 @@ unsigned long __init dom0_compute_nr_pages(
* maximum of 128MB.
*/
if ( nr_pages == 0 )
- nr_pages = -min(avail / 16, 128UL << (20 - PAGE_SHIFT));
+ {
+ uint64_t rsvd = min(avail / 16, 128UL << (20 - PAGE_SHIFT));
+ if ( pv_shim )
+ {
+ rsvd = pv_shim_mem(avail);
+ printk("Reserved %lu pages for xen-shim\n", rsvd);
+
+ }
+ nr_pages = -rsvd;
+ }
/* Negative specification means "all memory - specified amount". */
if ( (long)nr_pages < 0 ) nr_pages += avail;
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index c53a4ca407..6dc1ee45d7 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -53,6 +53,52 @@ static long pv_shim_grant_table_op(unsigned int cmd,
XEN_GUEST_HANDLE_PARAM(void) uop,
unsigned int count);
+/*
+ * By default, 1/16th of total HVM container's memory is reserved for xen-shim
+ * with minimum amount being 10MB and maximum amount 128MB. Some users may wish
+ * to tune this constants for better memory utilization. This can be achieved
+ * using the following xen-shim's command line option:
+ *
+ * shim_mem=[min:<min_amt>,][max:<max_amt>,][<amt>]
+ *
+ * <min_amt>: The minimum amount of memory that should be allocated for xen-shim
+ * (ignored if greater than max)
+ * <max_amt>: The maximum amount of memory that should be allocated for xen-shim
+ * <amt>: The precise amount of memory to allocate for xen-shim
+ * (overrides both min and max)
+ */
+static uint64_t __initdata shim_nrpages;
+static uint64_t __initdata shim_min_nrpages = 10UL << (20 - PAGE_SHIFT);
+static uint64_t __initdata shim_max_nrpages = 128UL << (20 - PAGE_SHIFT);
+
+static int __init parse_shim_mem(const char *s)
+{
+ do {
+ if ( !strncmp(s, "min:", 4) )
+ shim_min_nrpages = parse_size_and_unit(s+4, &s) >> PAGE_SHIFT;
+ else if ( !strncmp(s, "max:", 4) )
+ shim_max_nrpages = parse_size_and_unit(s+4, &s) >> PAGE_SHIFT;
+ else
+ shim_nrpages = parse_size_and_unit(s, &s) >> PAGE_SHIFT;
+ } while ( *s++ == ',' );
+
+ return s[-1] ? -EINVAL : 0;
+}
+custom_param("shim_mem", parse_shim_mem);
+
+uint64_t pv_shim_mem(uint64_t avail)
+{
+ uint64_t rsvd = min(avail / 16, shim_max_nrpages);
+
+ if ( shim_nrpages )
+ return shim_nrpages;
+
+ if ( shim_min_nrpages <= shim_max_nrpages )
+ rsvd = max(rsvd, shim_min_nrpages);
+
+ return rsvd;
+}
+
#define L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_USER| \
_PAGE_GUEST_KERNEL)
#define COMPAT_L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED)
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index 4d5f0b43fc..0ef452158e 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -38,6 +38,7 @@ void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
int pv_shim_shutdown(uint8_t reason);
void pv_shim_inject_evtchn(unsigned int port);
domid_t get_initial_domain_id(void);
+uint64_t pv_shim_mem(uint64_t avail);
#else
@@ -63,6 +64,11 @@ static inline domid_t get_initial_domain_id(void)
{
return 0;
}
+static inline uint64_t pv_shim_mem(uint64_t avail)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+}
#endif
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 52/62] xen/pvshim: set max_pages to the value of tot_pages
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (50 preceding siblings ...)
2018-01-12 11:28 ` [PATCH v2 51/62] xen/pvshim: add shim_mem cmdline parameter Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 53/62] xen/pvshim: support vCPU hotplug Wei Liu
` (9 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
So that the guest is not able to deplete the memory pool of the shim
itself by trying to balloon up.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
---
xen/arch/x86/pv/shim.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 6dc1ee45d7..e3e101a5b1 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -204,6 +204,12 @@ void __init pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
(hypercall_fn_t *)pv_shim_grant_table_op,
(hypercall_fn_t *)pv_shim_grant_table_op);
guest = d;
+
+ /*
+ * Set the max pages to the current number of pages to prevent the
+ * guest from depleting the shim memory pool.
+ */
+ d->max_pages = d->tot_pages;
}
static void write_start_info(struct domain *d)
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 53/62] xen/pvshim: support vCPU hotplug
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (51 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 52/62] xen/pvshim: set max_pages to the value of tot_pages Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 54/62] xen/pvshim: memory hotplug Wei Liu
` (8 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Fix hotplug so that v->initialized is checked before attempting to
bring up the physical CPU.
- Fix ARM compilation.
---
xen/arch/x86/pv/shim.c | 63 +++++++++++++++++++++++++++++++++++++++++++
xen/common/domain.c | 38 +++++++++++++++++---------
xen/include/asm-x86/pv/shim.h | 12 +++++++++
xen/include/xen/domain.h | 1 +
4 files changed, 102 insertions(+), 12 deletions(-)
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index e3e101a5b1..68ec7bed8e 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -751,6 +751,69 @@ static long pv_shim_grant_table_op(unsigned int cmd,
return rc;
}
+long pv_shim_cpu_up(void *data)
+{
+ struct vcpu *v = data;
+ struct domain *d = v->domain;
+ bool wake;
+
+ BUG_ON(smp_processor_id() != 0);
+
+ domain_lock(d);
+ if ( !v->is_initialised )
+ {
+ domain_unlock(d);
+ return -EINVAL;
+ }
+
+ if ( !cpu_online(v->vcpu_id) )
+ {
+ long rc = cpu_up_helper((void *)(unsigned long)v->vcpu_id);
+
+ if ( rc )
+ {
+ domain_unlock(d);
+ gprintk(XENLOG_ERR, "Failed to bring up CPU#%u: %ld\n",
+ v->vcpu_id, rc);
+ return rc;
+ }
+ }
+
+ wake = test_and_clear_bit(_VPF_down, &v->pause_flags);
+ domain_unlock(d);
+ if ( wake )
+ vcpu_wake(v);
+
+ return 0;
+}
+
+long pv_shim_cpu_down(void *data)
+{
+ struct vcpu *v = data;
+ long rc;
+
+ BUG_ON(smp_processor_id() != 0);
+
+ if ( !test_and_set_bit(_VPF_down, &v->pause_flags) )
+ vcpu_sleep_sync(v);
+
+ if ( cpu_online(v->vcpu_id) )
+ {
+ rc = cpu_down_helper((void *)(unsigned long)v->vcpu_id);
+ if ( rc )
+ gprintk(XENLOG_ERR, "Failed to bring down CPU#%u: %ld\n",
+ v->vcpu_id, rc);
+ /*
+ * NB: do not propagate errors from cpu_down_helper failing. The shim
+ * is going to run with extra CPUs, but that's not going to prevent
+ * normal operation. OTOH most guests are not prepared to handle an
+ * error on VCPUOP_down failing, and will likely panic.
+ */
+ }
+
+ return 0;
+}
+
domid_t get_initial_domain_id(void)
{
uint32_t eax, ebx, ecx, edx;
diff --git a/xen/common/domain.c b/xen/common/domain.c
index 6ed21b2536..558318e852 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -1301,22 +1301,36 @@ long do_vcpu_op(int cmd, unsigned int vcpuid, XEN_GUEST_HANDLE_PARAM(void) arg)
break;
- case VCPUOP_up: {
- bool_t wake = 0;
- domain_lock(d);
- if ( !v->is_initialised )
- rc = -EINVAL;
+ case VCPUOP_up:
+#ifdef CONFIG_X86
+ if ( pv_shim )
+ rc = continue_hypercall_on_cpu(0, pv_shim_cpu_up, v);
else
- wake = test_and_clear_bit(_VPF_down, &v->pause_flags);
- domain_unlock(d);
- if ( wake )
- vcpu_wake(v);
+#endif
+ {
+ bool wake = false;
+
+ domain_lock(d);
+ if ( !v->is_initialised )
+ rc = -EINVAL;
+ else
+ wake = test_and_clear_bit(_VPF_down, &v->pause_flags);
+ domain_unlock(d);
+ if ( wake )
+ vcpu_wake(v);
+ }
+
break;
- }
case VCPUOP_down:
- if ( !test_and_set_bit(_VPF_down, &v->pause_flags) )
- vcpu_sleep_nosync(v);
+#ifdef CONFIG_X86
+ if ( pv_shim )
+ rc = continue_hypercall_on_cpu(0, pv_shim_cpu_down, v);
+ else
+#endif
+ if ( !test_and_set_bit(_VPF_down, &v->pause_flags) )
+ vcpu_sleep_nosync(v);
+
break;
case VCPUOP_is_up:
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index 0ef452158e..eb59ddd38a 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -37,6 +37,8 @@ void pv_shim_setup_dom(struct domain *d, l4_pgentry_t *l4start,
start_info_t *si);
int pv_shim_shutdown(uint8_t reason);
void pv_shim_inject_evtchn(unsigned int port);
+long pv_shim_cpu_up(void *data);
+long pv_shim_cpu_down(void *data);
domid_t get_initial_domain_id(void);
uint64_t pv_shim_mem(uint64_t avail);
@@ -60,6 +62,16 @@ static inline void pv_shim_inject_evtchn(unsigned int port)
{
ASSERT_UNREACHABLE();
}
+static inline long pv_shim_cpu_up(void *data)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+}
+static inline long pv_shim_cpu_down(void *data)
+{
+ ASSERT_UNREACHABLE();
+ return 0;
+}
static inline domid_t get_initial_domain_id(void)
{
return 0;
diff --git a/xen/include/xen/domain.h b/xen/include/xen/domain.h
index 347f264047..eb62f1dab1 100644
--- a/xen/include/xen/domain.h
+++ b/xen/include/xen/domain.h
@@ -17,6 +17,7 @@ struct vcpu *alloc_vcpu(
struct domain *d, unsigned int vcpu_id, unsigned int cpu_id);
struct vcpu *alloc_dom0_vcpu0(struct domain *dom0);
int vcpu_reset(struct vcpu *);
+int vcpu_up(struct vcpu *v);
struct xen_domctl_getdomaininfo;
void getdomaininfo(struct domain *d, struct xen_domctl_getdomaininfo *info);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 54/62] xen/pvshim: memory hotplug
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (52 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 53/62] xen/pvshim: support vCPU hotplug Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 55/62] xen/shim: modify shim_mem parameter behaviour Wei Liu
` (7 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Add an order parameter to batch_memory_op.
- Add a TODO item regarding high order memory chunks to
pv_shim_online_memory.
- Use page_list_splice.
- Make sure the shim handlers are not called multiple times when
the hypercall is preempted.
---
xen/arch/x86/pv/shim.c | 112 ++++++++++++++++++++++++++++++++++++++++++
xen/common/memory.c | 21 ++++++++
xen/include/asm-x86/pv/shim.h | 10 ++++
3 files changed, 143 insertions(+)
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 68ec7bed8e..4120cc550e 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -48,6 +48,9 @@ static unsigned int nr_grant_list;
static unsigned long *grant_frames;
static DEFINE_SPINLOCK(grant_lock);
+static PAGE_LIST_HEAD(balloon);
+static DEFINE_SPINLOCK(balloon_lock);
+
static long pv_shim_event_channel_op(int cmd, XEN_GUEST_HANDLE_PARAM(void) arg);
static long pv_shim_grant_table_op(unsigned int cmd,
XEN_GUEST_HANDLE_PARAM(void) uop,
@@ -814,6 +817,115 @@ long pv_shim_cpu_down(void *data)
return 0;
}
+static unsigned long batch_memory_op(unsigned int cmd, unsigned int order,
+ const struct page_list_head *list)
+{
+ struct xen_memory_reservation xmr = {
+ .domid = DOMID_SELF,
+ .extent_order = order,
+ };
+ unsigned long pfns[64];
+ const struct page_info *pg;
+ unsigned long done = 0;
+
+ set_xen_guest_handle(xmr.extent_start, pfns);
+ page_list_for_each ( pg, list )
+ {
+ pfns[xmr.nr_extents++] = page_to_mfn(pg);
+ if ( xmr.nr_extents == ARRAY_SIZE(pfns) || !page_list_next(pg, list) )
+ {
+ long nr = xen_hypercall_memory_op(cmd, &xmr);
+
+ done += nr > 0 ? nr : 0;
+ if ( nr != xmr.nr_extents )
+ break;
+ xmr.nr_extents = 0;
+ }
+ }
+
+ return done;
+}
+
+void pv_shim_online_memory(unsigned int nr, unsigned int order)
+{
+ struct page_info *page, *tmp;
+ PAGE_LIST_HEAD(list);
+
+ spin_lock(&balloon_lock);
+ page_list_for_each_safe ( page, tmp, &balloon )
+ {
+ /* TODO: add support for splitting high order memory chunks. */
+ if ( page->v.free.order != order )
+ continue;
+
+ page_list_del(page, &balloon);
+ page_list_add_tail(page, &list);
+ if ( !--nr )
+ break;
+ }
+ spin_unlock(&balloon_lock);
+
+ if ( nr )
+ gprintk(XENLOG_WARNING,
+ "failed to allocate %u extents of order %u for onlining\n",
+ nr, order);
+
+ nr = batch_memory_op(XENMEM_populate_physmap, order, &list);
+ while ( nr-- )
+ {
+ BUG_ON((page = page_list_remove_head(&list)) == NULL);
+ free_domheap_pages(page, order);
+ }
+
+ if ( !page_list_empty(&list) )
+ {
+ gprintk(XENLOG_WARNING,
+ "failed to online some of the memory regions\n");
+ spin_lock(&balloon_lock);
+ page_list_splice(&list, &balloon);
+ spin_unlock(&balloon_lock);
+ }
+}
+
+void pv_shim_offline_memory(unsigned int nr, unsigned int order)
+{
+ struct page_info *page;
+ PAGE_LIST_HEAD(list);
+
+ while ( nr-- )
+ {
+ page = alloc_domheap_pages(NULL, order, 0);
+ if ( !page )
+ break;
+
+ page_list_add_tail(page, &list);
+ page->v.free.order = order;
+ }
+
+ if ( nr + 1 )
+ gprintk(XENLOG_WARNING,
+ "failed to reserve %u extents of order %u for offlining\n",
+ nr + 1, order);
+
+
+ nr = batch_memory_op(XENMEM_decrease_reservation, order, &list);
+ spin_lock(&balloon_lock);
+ while ( nr-- )
+ {
+ BUG_ON((page = page_list_remove_head(&list)) == NULL);
+ page_list_add_tail(page, &balloon);
+ }
+ spin_unlock(&balloon_lock);
+
+ if ( !page_list_empty(&list) )
+ {
+ gprintk(XENLOG_WARNING,
+ "failed to offline some of the memory regions\n");
+ while ( (page = page_list_remove_head(&list)) != NULL )
+ free_domheap_pages(page, order);
+ }
+}
+
domid_t get_initial_domain_id(void)
{
uint32_t eax, ebx, ecx, edx;
diff --git a/xen/common/memory.c b/xen/common/memory.c
index 5a1508a292..71e19aa629 100644
--- a/xen/common/memory.c
+++ b/xen/common/memory.c
@@ -29,6 +29,10 @@
#include <public/memory.h>
#include <xsm/xsm.h>
+#ifdef CONFIG_X86
+#include <asm/guest.h>
+#endif
+
struct memop_args {
/* INPUT */
struct domain *domain; /* Domain to be affected. */
@@ -993,6 +997,12 @@ long do_memory_op(unsigned long cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
return start_extent;
}
+#ifdef CONFIG_X86
+ if ( pv_shim && op != XENMEM_decrease_reservation && !args.preempted )
+ /* Avoid calling pv_shim_online_memory when preempted. */
+ pv_shim_online_memory(args.nr_extents, args.extent_order);
+#endif
+
switch ( op )
{
case XENMEM_increase_reservation:
@@ -1015,6 +1025,17 @@ long do_memory_op(unsigned long cmd, XEN_GUEST_HANDLE_PARAM(void) arg)
__HYPERVISOR_memory_op, "lh",
op | (rc << MEMOP_EXTENT_SHIFT), arg);
+#ifdef CONFIG_X86
+ if ( pv_shim && op == XENMEM_decrease_reservation )
+ /*
+ * Only call pv_shim_offline_memory when the hypercall has
+ * finished. Note that nr_done is used to cope in case the
+ * hypercall has failed and only part of the extents where
+ * processed.
+ */
+ pv_shim_offline_memory(args.nr_extents, args.nr_done);
+#endif
+
break;
case XENMEM_exchange:
diff --git a/xen/include/asm-x86/pv/shim.h b/xen/include/asm-x86/pv/shim.h
index eb59ddd38a..fb739772df 100644
--- a/xen/include/asm-x86/pv/shim.h
+++ b/xen/include/asm-x86/pv/shim.h
@@ -39,6 +39,8 @@ int pv_shim_shutdown(uint8_t reason);
void pv_shim_inject_evtchn(unsigned int port);
long pv_shim_cpu_up(void *data);
long pv_shim_cpu_down(void *data);
+void pv_shim_online_memory(unsigned int nr, unsigned int order);
+void pv_shim_offline_memory(unsigned int nr, unsigned int order);
domid_t get_initial_domain_id(void);
uint64_t pv_shim_mem(uint64_t avail);
@@ -72,6 +74,14 @@ static inline long pv_shim_cpu_down(void *data)
ASSERT_UNREACHABLE();
return 0;
}
+static inline void pv_shim_online_memory(unsigned int nr, unsigned int order)
+{
+ ASSERT_UNREACHABLE();
+}
+static inline void pv_shim_offline_memory(unsigned int nr, unsigned int order)
+{
+ ASSERT_UNREACHABLE();
+}
static inline domid_t get_initial_domain_id(void)
{
return 0;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 55/62] xen/shim: modify shim_mem parameter behaviour
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (53 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 54/62] xen/pvshim: memory hotplug Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 56/62] xen/pvshim: use default position for the m2p mappings Wei Liu
` (6 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
shim_mem will now account for both the memory used by the hypervisor
loaded in memory and the free memory slack given to the shim for
runtime usage.
From experimental testing it seems like the total amount of MiB used
by the shim (giving it ~1MB of free memory for runtime) is:
memory/113 + 20
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
docs/misc/xen-command-line.markdown | 13 +++++++------
xen/arch/x86/dom0_build.c | 14 +++-----------
xen/arch/x86/pv/shim.c | 30 +++++++++++++++++++-----------
3 files changed, 29 insertions(+), 28 deletions(-)
diff --git a/docs/misc/xen-command-line.markdown b/docs/misc/xen-command-line.markdown
index 9f51710a46..68ec52b5c2 100644
--- a/docs/misc/xen-command-line.markdown
+++ b/docs/misc/xen-command-line.markdown
@@ -1461,16 +1461,17 @@ constructed into a plain unprivileged PV domain.
### shim\_mem (x86)
> `= List of ( min:<size> | max:<size> | <size> )`
-Set the amount of memory that xen-shim reserves for itself. Only has effect
-if pv-shim mode is enabled.
+Set the amount of memory that xen-shim uses. Only has effect if pv-shim mode is
+enabled. Note that this value accounts for the memory used by the shim itself
+plus the free memory slack given to the shim for runtime allocations.
* `min:<size>` specifies the minimum amount of memory. Ignored if greater
- than max. Default: 10M.
-* `max:<size>` specifies the maximum amount of memory. Default: 128M.
+ than max.
+* `max:<size>` specifies the maximum amount of memory.
* `<size>` specifies the exact amount of memory. Overrides both min and max.
-By default, 1/16th of total HVM container's memory is reserved for xen-shim
-with minimum amount being 10MB and maximum amount 128MB.
+By default, the amount of free memory slack given to the shim for runtime usage
+is 1MB.
### rcu-idle-timer-period-ms
> `= <integer>`
diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 2aac0803fc..593bdbc18b 100644
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -290,17 +290,9 @@ unsigned long __init dom0_compute_nr_pages(
* for things like DMA buffers. This reservation is clamped to a
* maximum of 128MB.
*/
- if ( nr_pages == 0 )
- {
- uint64_t rsvd = min(avail / 16, 128UL << (20 - PAGE_SHIFT));
- if ( pv_shim )
- {
- rsvd = pv_shim_mem(avail);
- printk("Reserved %lu pages for xen-shim\n", rsvd);
-
- }
- nr_pages = -rsvd;
- }
+ if ( !nr_pages )
+ nr_pages = -(pv_shim ? pv_shim_mem(avail)
+ : min(avail / 16, 128UL << (20 - PAGE_SHIFT)));
/* Negative specification means "all memory - specified amount". */
if ( (long)nr_pages < 0 ) nr_pages += avail;
diff --git a/xen/arch/x86/pv/shim.c b/xen/arch/x86/pv/shim.c
index 4120cc550e..702249719e 100644
--- a/xen/arch/x86/pv/shim.c
+++ b/xen/arch/x86/pv/shim.c
@@ -57,9 +57,8 @@ static long pv_shim_grant_table_op(unsigned int cmd,
unsigned int count);
/*
- * By default, 1/16th of total HVM container's memory is reserved for xen-shim
- * with minimum amount being 10MB and maximum amount 128MB. Some users may wish
- * to tune this constants for better memory utilization. This can be achieved
+ * By default give the shim 1MB of free memory slack. Some users may wish to
+ * tune this constants for better memory utilization. This can be achieved
* using the following xen-shim's command line option:
*
* shim_mem=[min:<min_amt>,][max:<max_amt>,][<amt>]
@@ -71,8 +70,8 @@ static long pv_shim_grant_table_op(unsigned int cmd,
* (overrides both min and max)
*/
static uint64_t __initdata shim_nrpages;
-static uint64_t __initdata shim_min_nrpages = 10UL << (20 - PAGE_SHIFT);
-static uint64_t __initdata shim_max_nrpages = 128UL << (20 - PAGE_SHIFT);
+static uint64_t __initdata shim_min_nrpages;
+static uint64_t __initdata shim_max_nrpages;
static int __init parse_shim_mem(const char *s)
{
@@ -91,15 +90,24 @@ custom_param("shim_mem", parse_shim_mem);
uint64_t pv_shim_mem(uint64_t avail)
{
- uint64_t rsvd = min(avail / 16, shim_max_nrpages);
+ if ( !shim_nrpages )
+ {
+ shim_nrpages = max(shim_min_nrpages,
+ total_pages - avail + (1UL << (20 - PAGE_SHIFT)));
+ if ( shim_max_nrpages )
+ shim_max_nrpages = min(shim_nrpages, shim_max_nrpages);
+ }
+
+ if ( total_pages - avail > shim_nrpages )
+ panic("pages used by shim > shim_nrpages (%#lx > %#lx)",
+ total_pages - avail, shim_nrpages);
- if ( shim_nrpages )
- return shim_nrpages;
+ shim_nrpages -= total_pages - avail;
- if ( shim_min_nrpages <= shim_max_nrpages )
- rsvd = max(rsvd, shim_min_nrpages);
+ printk("shim used pages %#lx reserving %#lx free pages\n",
+ total_pages - avail, shim_nrpages);
- return rsvd;
+ return shim_nrpages;
}
#define L1_PROT (_PAGE_PRESENT|_PAGE_RW|_PAGE_ACCESSED|_PAGE_USER| \
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 56/62] xen/pvshim: use default position for the m2p mappings
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (54 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 55/62] xen/shim: modify shim_mem parameter behaviour Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 57/62] xen/shim: crash instead of reboot in shim mode Wei Liu
` (5 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
When running a 32bit kernel as Dom0 on a 64bit hypervisor the
hypervisor will try to shrink the hypervisor hole to the minimum
needed, and thus requires the Dom0 to use XENMEM_machphys_mapping in
order to fetch the position of the start of the hypervisor virtual
mappings.
Disable this feature when running as a PV shim, since some DomU
kernels don't implemented XENMEM_machphys_mapping and break if the m2p
doesn't begin at the default address.
NB: support for the XENMEM_machphys_mapping was added in Linux by
commit 7e7750.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
---
xen/arch/x86/pv/dom0_build.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/xen/arch/x86/pv/dom0_build.c b/xen/arch/x86/pv/dom0_build.c
index 52576795cc..e0b6b14e58 100644
--- a/xen/arch/x86/pv/dom0_build.c
+++ b/xen/arch/x86/pv/dom0_build.c
@@ -397,7 +397,8 @@ int __init dom0_construct_pv(struct domain *d,
if ( parms.pae == XEN_PAE_EXTCR3 )
set_bit(VMASST_TYPE_pae_extended_cr3, &d->vm_assist);
- if ( (parms.virt_hv_start_low != UNSET_ADDR) && elf_32bit(&elf) )
+ if ( !pv_shim && (parms.virt_hv_start_low != UNSET_ADDR) &&
+ elf_32bit(&elf) )
{
unsigned long mask = (1UL << L2_PAGETABLE_SHIFT) - 1;
value = (parms.virt_hv_start_low + mask) & ~mask;
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 57/62] xen/shim: crash instead of reboot in shim mode
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (55 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 56/62] xen/pvshim: use default position for the m2p mappings Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 58/62] xen/shim: allow DomU to have as many vcpus as available Wei Liu
` (4 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
All guest shutdown operations are forwarded to L0, so the only native
calls to machine_restart happen from crash related paths inside the
hypervisor, hence switch the reboot code to instead issue a crash
shutdown.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
[ wei: fix arm build ]
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
Changes since v1:
- Use the ternary operator.
---
xen/arch/x86/shutdown.c | 7 ++++++-
xen/drivers/char/console.c | 4 ++++
2 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/xen/arch/x86/shutdown.c b/xen/arch/x86/shutdown.c
index 689f6f137d..a350714319 100644
--- a/xen/arch/x86/shutdown.c
+++ b/xen/arch/x86/shutdown.c
@@ -642,7 +642,12 @@ void machine_restart(unsigned int delay_millisecs)
break;
case BOOT_XEN:
- xen_hypercall_shutdown(SHUTDOWN_reboot);
+ /*
+ * When running in PV shim mode guest shutdown calls are
+ * forwarded to L0, hence the only way to get here is if a
+ * shim crash happens.
+ */
+ xen_hypercall_shutdown(pv_shim ? SHUTDOWN_crash : SHUTDOWN_reboot);
break;
}
}
diff --git a/xen/drivers/char/console.c b/xen/drivers/char/console.c
index 18f5b7f7aa..121073c8ed 100644
--- a/xen/drivers/char/console.c
+++ b/xen/drivers/char/console.c
@@ -1253,7 +1253,11 @@ void panic(const char *fmt, ...)
if ( opt_noreboot )
printk("Manual reset required ('noreboot' specified)\n");
else
+#ifdef CONFIG_X86
+ printk("%s in five seconds...\n", pv_shim ? "Crash" : "Reboot");
+#else
printk("Reboot in five seconds...\n");
+#endif
spin_unlock_irqrestore(&lock, flags);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 58/62] xen/shim: allow DomU to have as many vcpus as available
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (56 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 57/62] xen/shim: crash instead of reboot in shim mode Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 59/62] libxl: pvshim: Provide first-class config settings to enable shim mode Wei Liu
` (3 subsequent siblings)
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Roger Pau Monne <roger.pau@citrix.com>
Since the shim VCPUOP_{up/down} hypercall is wired to the plug/unplug
of CPUs to the shim itself, start the shim DomU with only the BSP
online, and let the guest bring up other CPUs as it needs them.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
---
Changes since v1:
- Fix single line comment style.
- Print Dom%u d->domain_id.
- Change position of __start_xen comment.
---
xen/arch/x86/dom0_build.c | 30 +++++++++++++++++++++++++++---
xen/arch/x86/pv/dom0_build.c | 2 +-
xen/arch/x86/setup.c | 28 ++++++++++++++++++----------
3 files changed, 46 insertions(+), 14 deletions(-)
diff --git a/xen/arch/x86/dom0_build.c b/xen/arch/x86/dom0_build.c
index 593bdbc18b..17cb1272c1 100644
--- a/xen/arch/x86/dom0_build.c
+++ b/xen/arch/x86/dom0_build.c
@@ -138,9 +138,18 @@ struct vcpu *__init dom0_setup_vcpu(struct domain *d,
if ( v )
{
- if ( !d->is_pinned && !dom0_affinity_relaxed )
- cpumask_copy(v->cpu_hard_affinity, &dom0_cpus);
- cpumask_copy(v->cpu_soft_affinity, &dom0_cpus);
+ if ( pv_shim )
+ {
+
+ cpumask_setall(v->cpu_hard_affinity);
+ cpumask_setall(v->cpu_soft_affinity);
+ }
+ else
+ {
+ if ( !d->is_pinned && !dom0_affinity_relaxed )
+ cpumask_copy(v->cpu_hard_affinity, &dom0_cpus);
+ cpumask_copy(v->cpu_soft_affinity, &dom0_cpus);
+ }
}
return v;
@@ -153,6 +162,21 @@ unsigned int __init dom0_max_vcpus(void)
unsigned int i, max_vcpus, limit;
nodeid_t node;
+ if ( pv_shim )
+ {
+ nodes_setall(dom0_nodes);
+
+ /*
+ * When booting in shim mode APs are not started until the guest brings
+ * other vCPUs up.
+ */
+ cpumask_set_cpu(0, &dom0_cpus);
+
+ /* On PV shim mode allow the guest to have as many CPUs as available. */
+ return nr_cpu_ids;
+ }
+
+
for ( i = 0; i < dom0_nr_pxms; ++i )
if ( (node = pxm_to_node(dom0_pxms[i])) != NUMA_NO_NODE )
node_set(node, dom0_nodes);
diff --git a/xen/arch/x86/pv/dom0_build.c b/xen/arch/x86/pv/dom0_build.c
index e0b6b14e58..e7065bfde3 100644
--- a/xen/arch/x86/pv/dom0_build.c
+++ b/xen/arch/x86/pv/dom0_build.c
@@ -696,7 +696,7 @@ int __init dom0_construct_pv(struct domain *d,
for ( i = 0; i < XEN_LEGACY_MAX_VCPUS; i++ )
shared_info(d, vcpu_info[i].evtchn_upcall_mask) = 1;
- printk("Dom0 has maximum %u VCPUs\n", d->max_vcpus);
+ printk("Dom%u has maximum %u VCPUs\n", d->domain_id, d->max_vcpus);
cpu = v->processor;
for ( i = 1; i < d->max_vcpus; i++ )
diff --git a/xen/arch/x86/setup.c b/xen/arch/x86/setup.c
index 0f4282c42f..08796f4256 100644
--- a/xen/arch/x86/setup.c
+++ b/xen/arch/x86/setup.c
@@ -1584,18 +1584,26 @@ void __init noreturn __start_xen(unsigned long mbi_p)
do_presmp_initcalls();
- for_each_present_cpu ( i )
+ /*
+ * NB: when running as a PV shim VCPUOP_up/down is wired to the shim
+ * physical cpu_add/remove functions, so launch the guest with only
+ * the BSP online and let it bring up the other CPUs as required.
+ */
+ if ( !pv_shim )
{
- /* Set up cpu_to_node[]. */
- srat_detect_node(i);
- /* Set up node_to_cpumask based on cpu_to_node[]. */
- numa_add_cpu(i);
-
- if ( (num_online_cpus() < max_cpus) && !cpu_online(i) )
+ for_each_present_cpu ( i )
{
- int ret = cpu_up(i);
- if ( ret != 0 )
- printk("Failed to bring up CPU %u (error %d)\n", i, ret);
+ /* Set up cpu_to_node[]. */
+ srat_detect_node(i);
+ /* Set up node_to_cpumask based on cpu_to_node[]. */
+ numa_add_cpu(i);
+
+ if ( (num_online_cpus() < max_cpus) && !cpu_online(i) )
+ {
+ int ret = cpu_up(i);
+ if ( ret != 0 )
+ printk("Failed to bring up CPU %u (error %d)\n", i, ret);
+ }
}
}
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 59/62] libxl: pvshim: Provide first-class config settings to enable shim mode
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (57 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 58/62] xen/shim: allow DomU to have as many vcpus as available Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 14:45 ` [PATCH] fixup! " Ian Jackson
2018-01-12 11:29 ` [PATCH v2 60/62] libxl: pvshim: Introduce pvshim_extra Wei Liu
` (2 subsequent siblings)
61 siblings, 1 reply; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Ian Jackson <Ian.Jackson@eu.citrix.com>
This is API-compatible because old callers are supposed to call
libxl_*_init to initialise the struct; and the updated function clears
these members.
It is ABI-compatible because the new fields make this member of the
guest type union larger but only within the existing size of that
union.
Unfortunately it is not easy to backport because it depends on the PVH
domain type. Attempts to avoid use of the PVH domain type involved
working with two views of the configuration: the "underlying" domain
type and the "visible" type (and corresponding config info). Also
there are different sets of config settings for PV and PVH, which
callers would have to know to set.
And, unfortunately, it will not be possible, with this approach, to
enable the shim by default for all libxl callers. (Although it could
perhaps be done in xl.)
For now, our config defaults are:
* if enabled, path is "xen-shim" in the xen firmware directory
* if enabled, cmdline is the one we are currently debugging with
The debugging arguments will be rationalised in a moment.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Signed-off-by: George Dunlap <george.dunlap@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2: pvshim, not pvhshim
works with type "pvh", not type "pv"
---
tools/libxl/libxl.h | 8 +++++++
tools/libxl/libxl_create.c | 19 +++++++++++----
tools/libxl/libxl_dom.c | 57 +++++++++++++++++++++++++++++++++++---------
tools/libxl/libxl_internal.h | 4 ++++
tools/libxl/libxl_types.idl | 5 +++-
5 files changed, 77 insertions(+), 16 deletions(-)
diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
index 5e9aed739d..9632fd6d2f 100644
--- a/tools/libxl/libxl.h
+++ b/tools/libxl/libxl.h
@@ -1101,6 +1101,14 @@ void libxl_mac_copy(libxl_ctx *ctx, libxl_mac *dst, const libxl_mac *src);
*/
#define LIBXL_HAVE_SET_PARAMETERS 1
+/*
+ * LIBXL_HAVE_PV_SHIM
+ *
+ * If this is defined, libxl_domain_build_info's pvh type information
+ * contains members pvshim, pvshim_path, pvshim_cmdline.
+ */
+#define LIBXL_HAVE_PV_SHIM 1
+
typedef char **libxl_string_list;
void libxl_string_list_dispose(libxl_string_list *sl);
int libxl_string_list_length(const libxl_string_list *sl);
diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
index f15fb215c2..63cdc33991 100644
--- a/tools/libxl/libxl_create.c
+++ b/tools/libxl/libxl_create.c
@@ -389,6 +389,18 @@ int libxl__domain_build_info_setdefault(libxl__gc *gc,
}
break;
case LIBXL_DOMAIN_TYPE_PVH:
+ libxl_defbool_setdefault(&b_info->u.pvh.pvshim, false);
+ if (libxl_defbool_val(b_info->u.pvh.pvshim)) {
+ if (!b_info->u.pvh.pvshim_path)
+ b_info->u.pvh.pvshim_path =
+ libxl__sprintf(NOGC, "%s/%s",
+ libxl__xenfirmwaredir_path(),
+ PVSHIM_BASENAME);
+ if (!b_info->u.pvh.pvshim_cmdline)
+ b_info->u.pvh.pvshim_cmdline =
+ libxl__strdup(NOGC, PVSHIM_CMDLINE);
+ }
+
break;
default:
LOG(ERROR, "invalid domain type %s in create info",
@@ -476,10 +488,6 @@ int libxl__domain_build(libxl__gc *gc,
break;
case LIBXL_DOMAIN_TYPE_PV:
- ret = libxl__build_pv(gc, domid, info, state);
- if (ret)
- goto out;
-
vments = libxl__calloc(gc, 11, sizeof(char *));
i = 0;
vments[i++] = "image/ostype";
@@ -499,6 +507,9 @@ int libxl__domain_build(libxl__gc *gc,
break;
case LIBXL_DOMAIN_TYPE_PVH:
+ state->shim_path = info->u.pvh.pvshim_path;
+ state->shim_cmdline = info->u.pvh.pvshim_cmdline;
+
ret = libxl__build_hvm(gc, domid, d_config, state);
if (ret)
goto out;
diff --git a/tools/libxl/libxl_dom.c b/tools/libxl/libxl_dom.c
index fbbdb9ec2f..b03386409f 100644
--- a/tools/libxl/libxl_dom.c
+++ b/tools/libxl/libxl_dom.c
@@ -1025,22 +1025,51 @@ static int libxl__domain_firmware(libxl__gc *gc,
if (state->pv_kernel.path != NULL &&
info->type == LIBXL_DOMAIN_TYPE_PVH) {
- /* Try to load a kernel instead of the firmware. */
- if (state->pv_kernel.mapped) {
- rc = xc_dom_kernel_mem(dom, state->pv_kernel.data,
- state->pv_kernel.size);
+
+ if (state->shim_path) {
+ rc = xc_dom_kernel_file(dom, state->shim_path);
if (rc) {
- LOGE(ERROR, "xc_dom_kernel_mem failed");
+ LOGE(ERROR, "xc_dom_kernel_file failed");
goto out;
}
+
+ /* We've loaded the shim, so load the kernel as a secondary module */
+ if (state->pv_kernel.mapped) {
+ LOG(WARN, "xc_dom_module_mem, cmdline %s",
+ state->pv_cmdline);
+ rc = xc_dom_module_mem(dom, state->pv_kernel.data,
+ state->pv_kernel.size, state->pv_cmdline);
+ if (rc) {
+ LOGE(ERROR, "xc_dom_kernel_mem failed");
+ goto out;
+ }
+ } else {
+ LOG(WARN, "xc_dom_module_file, path %s cmdline %s",
+ state->pv_kernel.path, state->pv_cmdline);
+ rc = xc_dom_module_file(dom, state->pv_kernel.path, state->pv_cmdline);
+ if (rc) {
+ LOGE(ERROR, "xc_dom_kernel_file failed");
+ goto out;
+ }
+ }
} else {
- rc = xc_dom_kernel_file(dom, state->pv_kernel.path);
- if (rc) {
- LOGE(ERROR, "xc_dom_kernel_file failed");
- goto out;
+ /* No shim, so load the kernel directly */
+ if (state->pv_kernel.mapped) {
+ rc = xc_dom_kernel_mem(dom, state->pv_kernel.data,
+ state->pv_kernel.size);
+ if (rc) {
+ LOGE(ERROR, "xc_dom_kernel_mem failed");
+ goto out;
+ }
+ } else {
+ rc = xc_dom_kernel_file(dom, state->pv_kernel.path);
+ if (rc) {
+ LOGE(ERROR, "xc_dom_kernel_file failed");
+ goto out;
+ }
}
}
-
+
if (state->pv_ramdisk.path && strlen(state->pv_ramdisk.path)) {
if (state->pv_ramdisk.mapped) {
rc = xc_dom_module_mem(dom, state->pv_ramdisk.data,
@@ -1154,8 +1183,14 @@ int libxl__build_hvm(libxl__gc *gc, uint32_t domid,
xc_dom_loginit(ctx->xch);
+ /*
+ * If PVH and we have a shim override, use the shim cmdline.
+ * If PVH and no shim override, use the pv cmdline.
+ * If not PVH, use info->cmdline.
+ */
dom = xc_dom_allocate(ctx->xch, info->type == LIBXL_DOMAIN_TYPE_PVH ?
- state->pv_cmdline : info->cmdline, NULL);
+ (state->shim_path ? state->shim_cmdline : state->pv_cmdline) :
+ info->cmdline, NULL);
if (!dom) {
LOGE(ERROR, "xc_dom_allocate failed");
rc = ERROR_NOMEM;
diff --git a/tools/libxl/libxl_internal.h b/tools/libxl/libxl_internal.h
index bfa95d8619..2454efa621 100644
--- a/tools/libxl/libxl_internal.h
+++ b/tools/libxl/libxl_internal.h
@@ -118,6 +118,8 @@
#define TAP_DEVICE_SUFFIX "-emu"
#define DOMID_XS_PATH "domid"
#define INVALID_DOMID ~0
+#define PVSHIM_BASENAME "xen-shim"
+#define PVSHIM_CMDLINE "pv-shim console=xen,pv sched=null loglvl=all guest_loglvl=all apic_verbosity=debug e820-verbose"
/* Size macros. */
#define __AC(X,Y) (X##Y)
@@ -1136,6 +1138,8 @@ typedef struct {
libxl__file_reference pv_kernel;
libxl__file_reference pv_ramdisk;
+ const char * shim_path;
+ const char * shim_cmdline;
const char * pv_cmdline;
xen_vmemrange_t *vmemranges;
diff --git a/tools/libxl/libxl_types.idl b/tools/libxl/libxl_types.idl
index a239324341..6d060edc0d 100644
--- a/tools/libxl/libxl_types.idl
+++ b/tools/libxl/libxl_types.idl
@@ -592,7 +592,10 @@ libxl_domain_build_info = Struct("domain_build_info",[
# Use host's E820 for PCI passthrough.
("e820_host", libxl_defbool),
])),
- ("pvh", None),
+ ("pvh", Struct(None, [("pvshim", libxl_defbool),
+ ("pvshim_path", string),
+ ("pvshim_cmdline", string),
+ ])),
("invalid", None),
], keyvar_init_val = "LIBXL_DOMAIN_TYPE_INVALID")),
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH] fixup! libxl: pvshim: Provide first-class config settings to enable shim mode
2018-01-12 11:29 ` [PATCH v2 59/62] libxl: pvshim: Provide first-class config settings to enable shim mode Wei Liu
@ 2018-01-12 14:45 ` Ian Jackson
0 siblings, 0 replies; 66+ messages in thread
From: Ian Jackson @ 2018-01-12 14:45 UTC (permalink / raw)
To: xen-devel; +Cc: Wei Liu, Ian Jackson
Earlier versions of this patch have an erroneous hunk which will stop
PV guests from working at all!
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
CC: Wei Liu <wei.liu2@citrix.com>
---
tools/libxl/libxl_create.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
index 278ad6a..1fa1d3a 100644
--- a/tools/libxl/libxl_create.c
+++ b/tools/libxl/libxl_create.c
@@ -488,6 +488,10 @@ int libxl__domain_build(libxl__gc *gc,
break;
case LIBXL_DOMAIN_TYPE_PV:
+ ret = libxl__build_pv(gc, domid, info, state);
+ if (ret)
+ goto out;
+
vments = libxl__calloc(gc, 11, sizeof(char *));
i = 0;
vments[i++] = "image/ostype";
--
2.1.4
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread
* [PATCH v2 60/62] libxl: pvshim: Introduce pvshim_extra
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (58 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 59/62] libxl: pvshim: Provide first-class config settings to enable shim mode Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 61/62] xl: pvshim: Provide and document xl config Wei Liu
2018-01-12 11:29 ` [PATCH v2 62/62] xl: Default guest mode changed from PV to PVH with PV shim Wei Liu
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Ian Jackson <ian.jackson@eu.citrix.com>
And move the debugging options from the default config into a doc
comment in libxl_types.idl.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
---
v2: pvshim, not pvhshim
works with type "pvh", not type "pv"
---
tools/libxl/libxl.h | 2 +-
tools/libxl/libxl_create.c | 5 ++++-
tools/libxl/libxl_internal.h | 2 +-
tools/libxl/libxl_types.idl | 1 +
4 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
index 9632fd6d2f..3c0ae6de47 100644
--- a/tools/libxl/libxl.h
+++ b/tools/libxl/libxl.h
@@ -1105,7 +1105,7 @@ void libxl_mac_copy(libxl_ctx *ctx, libxl_mac *dst, const libxl_mac *src);
* LIBXL_HAVE_PV_SHIM
*
* If this is defined, libxl_domain_build_info's pvh type information
- * contains members pvshim, pvshim_path, pvshim_cmdline.
+ * contains members pvshim, pvshim_path, pvshim_cmdline, pvshim_extra.
*/
#define LIBXL_HAVE_PV_SHIM 1
diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
index 63cdc33991..278ad6a24e 100644
--- a/tools/libxl/libxl_create.c
+++ b/tools/libxl/libxl_create.c
@@ -508,7 +508,10 @@ int libxl__domain_build(libxl__gc *gc,
break;
case LIBXL_DOMAIN_TYPE_PVH:
state->shim_path = info->u.pvh.pvshim_path;
- state->shim_cmdline = info->u.pvh.pvshim_cmdline;
+ state->shim_cmdline = GCSPRINTF("%s%s%s",
+ info->u.pvh.pvshim_cmdline,
+ info->u.pvh.pvshim_extra ? " " : "",
+ info->u.pvh.pvshim_extra ? info->u.pvh.pvshim_extra : "");
ret = libxl__build_hvm(gc, domid, d_config, state);
if (ret)
diff --git a/tools/libxl/libxl_internal.h b/tools/libxl/libxl_internal.h
index 2454efa621..0f89364466 100644
--- a/tools/libxl/libxl_internal.h
+++ b/tools/libxl/libxl_internal.h
@@ -119,7 +119,7 @@
#define DOMID_XS_PATH "domid"
#define INVALID_DOMID ~0
#define PVSHIM_BASENAME "xen-shim"
-#define PVSHIM_CMDLINE "pv-shim console=xen,pv sched=null loglvl=all guest_loglvl=all apic_verbosity=debug e820-verbose"
+#define PVSHIM_CMDLINE "pv-shim console=xen,pv sched=null"
/* Size macros. */
#define __AC(X,Y) (X##Y)
diff --git a/tools/libxl/libxl_types.idl b/tools/libxl/libxl_types.idl
index 6d060edc0d..d74fac7d30 100644
--- a/tools/libxl/libxl_types.idl
+++ b/tools/libxl/libxl_types.idl
@@ -595,6 +595,7 @@ libxl_domain_build_info = Struct("domain_build_info",[
("pvh", Struct(None, [("pvshim", libxl_defbool),
("pvshim_path", string),
("pvshim_cmdline", string),
+ ("pvshim_extra", string), # eg "loglvl=all guest_loglvl=all apic_verbosity=debug e820-verbose"
])),
("invalid", None),
], keyvar_init_val = "LIBXL_DOMAIN_TYPE_INVALID")),
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 61/62] xl: pvshim: Provide and document xl config
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (59 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 60/62] libxl: pvshim: Introduce pvshim_extra Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
2018-01-12 11:29 ` [PATCH v2 62/62] xl: Default guest mode changed from PV to PVH with PV shim Wei Liu
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Ian Jackson <ian.jackson@eu.citrix.com>
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
---
v2: pvshim, not pvhshim
works with type "pvh", not type "pv"
pvshim_etc. options in config are not erroneously ignored
---
docs/man/xl.cfg.pod.5.in | 35 +++++++++++++++++++++++++++++++++++
tools/xl/xl_parse.c | 14 ++++++++++++++
2 files changed, 49 insertions(+)
diff --git a/docs/man/xl.cfg.pod.5.in b/docs/man/xl.cfg.pod.5.in
index b7b91d8627..bf6c266de1 100644
--- a/docs/man/xl.cfg.pod.5.in
+++ b/docs/man/xl.cfg.pod.5.in
@@ -508,6 +508,41 @@ Load the specified file as firmware for the guest.
Currently there's no firmware available for PVH guests, they should be
booted using the B<Direct Kernel Boot> method or the B<bootloader> option.
+=over 4
+
+=item B<pvshim=BOOLEAN>
+
+Whether to boot this guest as a PV guest within a PVH container.
+Ie, the guest will experience a PV environment,
+but
+processor hardware extensions are used to
+separate its address space
+to mitigate the Meltdown attack (CVE-2017-5754).
+
+Default is false.
+
+=item B<pvshim_path="PATH">
+
+The PV shim is a specially-built firmware-like executable
+constructed from the hypervisor source tree.
+This option specifies to use a non-default shim.
+Ignored if pvhsim is false.
+
+=item B<pvshim_cmdline="STRING">
+
+Command line for the shim.
+Default is "pv-shim console=xen,pv sched=null".
+Ignored if pvhsim is false.
+
+=item B<pvshim_extra="STRING">
+
+Extra command line arguments for the shim.
+If supplied, appended to the value for pvshim_cmdline.
+Default is empty.
+Ignored if pvhsim is false.
+
+=back
+
=head3 Other Options
=over 4
diff --git a/tools/xl/xl_parse.c b/tools/xl/xl_parse.c
index 9a692d5ae6..fdfe693de1 100644
--- a/tools/xl/xl_parse.c
+++ b/tools/xl/xl_parse.c
@@ -964,6 +964,20 @@ void parse_config_data(const char *config_source,
xlu_cfg_replace_string(config, "pool", &c_info->pool_name, 0);
libxl_domain_build_info_init_type(b_info, c_info->type);
+
+ if (b_info->type == LIBXL_DOMAIN_TYPE_PVH) {
+ xlu_cfg_get_defbool(config, "pvshim", &b_info->u.pvh.pvshim, 0);
+ if (!xlu_cfg_get_string(config, "pvshim_path", &buf, 0))
+ xlu_cfg_replace_string(config, "pvshim_path",
+ &b_info->u.pvh.pvshim_path, 0);
+ if (!xlu_cfg_get_string(config, "pvshim_cmdline", &buf, 0))
+ xlu_cfg_replace_string(config, "pvshim_cmdline",
+ &b_info->u.pvh.pvshim_cmdline, 0);
+ if (!xlu_cfg_get_string(config, "pvshim_extra", &buf, 0))
+ xlu_cfg_replace_string(config, "pvshim_extra",
+ &b_info->u.pvh.pvshim_extra, 0);
+ }
+
if (blkdev_start)
b_info->blkdev_start = strdup(blkdev_start);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread* [PATCH v2 62/62] xl: Default guest mode changed from PV to PVH with PV shim
2018-01-12 11:28 [PATCH v2 00/62] Comet: Run PV in PVH container Wei Liu
` (60 preceding siblings ...)
2018-01-12 11:29 ` [PATCH v2 61/62] xl: pvshim: Provide and document xl config Wei Liu
@ 2018-01-12 11:29 ` Wei Liu
61 siblings, 0 replies; 66+ messages in thread
From: Wei Liu @ 2018-01-12 11:29 UTC (permalink / raw)
To: Xen-devel; +Cc: wei.liu2
From: Ian Jackson <ian.jackson@eu.citrix.com>
If the config file specifies a type (or builder), it overrides this
default. But if it doesn't, you now get a PV-in-PVH guest.
Signed-off-by: Ian Jackson <Ian.Jackson@eu.citrix.com>
---
v2: New patch
---
tools/xl/xl_parse.c | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/tools/xl/xl_parse.c b/tools/xl/xl_parse.c
index fdfe693de1..2a53ac0015 100644
--- a/tools/xl/xl_parse.c
+++ b/tools/xl/xl_parse.c
@@ -939,8 +939,12 @@ void parse_config_data(const char *config_source,
c_info->type = builder_type;
}
- if (c_info->type == LIBXL_DOMAIN_TYPE_INVALID)
- c_info->type = LIBXL_DOMAIN_TYPE_PV;
+ static bool pvshim_default_enable = 0;
+
+ if (c_info->type == LIBXL_DOMAIN_TYPE_INVALID) {
+ c_info->type = LIBXL_DOMAIN_TYPE_PVH;
+ pvshim_default_enable = 1;
+ }
xlu_cfg_get_defbool(config, "hap", &c_info->hap, 0);
@@ -966,7 +970,10 @@ void parse_config_data(const char *config_source,
libxl_domain_build_info_init_type(b_info, c_info->type);
if (b_info->type == LIBXL_DOMAIN_TYPE_PVH) {
- xlu_cfg_get_defbool(config, "pvshim", &b_info->u.pvh.pvshim, 0);
+ if (xlu_cfg_get_defbool(config, "pvshim", &b_info->u.pvh.pvshim, 0)
+ && pvshim_default_enable)
+ libxl_defbool_set(&b_info->u.pvh.pvshim, 1);
+
if (!xlu_cfg_get_string(config, "pvshim_path", &buf, 0))
xlu_cfg_replace_string(config, "pvshim_path",
&b_info->u.pvh.pvshim_path, 0);
--
2.11.0
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
^ permalink raw reply related [flat|nested] 66+ messages in thread