From: Wei Liu <wei.liu2@citrix.com>
To: Xen-devel <xen-devel@lists.xenproject.org>
Cc: wei.liu2@citrix.com
Subject: [PATCH v2 30/62] x86/guest: map per-cpu vcpu_info area.
Date: Fri, 12 Jan 2018 11:28:38 +0000 [thread overview]
Message-ID: <20180112112910.16762-31-wei.liu2@citrix.com> (raw)
In-Reply-To: <20180112112910.16762-1-wei.liu2@citrix.com>
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
next prev parent reply other threads:[~2018-01-12 12:06 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
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 ` [PATCH v2 03/62] tools/libxc: remove extraneous newline in xc_dom_load_acpi Wei Liu
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 ` [PATCH v2 05/62] tools/libxc: Multi modules support Wei Liu
2018-01-12 11:28 ` [PATCH v2 06/62] xen/common: Widen the guest logging buffer slightly 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
2018-01-12 11:28 ` [PATCH v2 08/62] x86/link: Introduce and use SECTION_ALIGN Wei Liu
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 ` [PATCH v2 10/62] xen/domctl: Return arch_config via getdomaininfo Wei Liu
2018-01-12 11:28 ` [PATCH v2 11/62] tools/ocaml: Expose arch_config in domaininfo Wei Liu
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 ` [PATCH v2 13/62] x86/fixmap: Modify fix_to_virt() to return a void pointer Wei Liu
2018-01-12 11:28 ` [PATCH v2 14/62] x86: Common cpuid faulting support Wei Liu
2018-01-12 11:28 ` [PATCH v2 15/62] x86/Kconfig: Options for Xen and PVH support Wei Liu
2018-01-12 11:28 ` [PATCH v2 16/62] x86/link: Relocate program headers Wei Liu
2018-01-12 11:28 ` [PATCH v2 17/62] x86: introduce ELFNOTE macro Wei Liu
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 ` [PATCH v2 19/62] x86/entry: Early PVH boot code Wei Liu
2018-01-12 11:28 ` [PATCH v2 20/62] x86/boot: Map more than the first 16MB Wei Liu
2018-01-12 11:28 ` [PATCH v2 21/62] x86/entry: Probe for Xen early during boot Wei Liu
2018-01-12 11:28 ` [PATCH v2 22/62] x86/guest: Hypercall support Wei Liu
2018-01-12 11:28 ` [PATCH v2 23/62] x86/shutdown: Support for using SCHEDOP_{shutdown, reboot} Wei Liu
2018-01-12 11:28 ` [PATCH v2 24/62] x86/pvh: Retrieve memory map from Xen Wei Liu
2018-01-12 11:28 ` [PATCH v2 25/62] xen/console: Introduce console=xen Wei Liu
2018-01-12 11:28 ` [PATCH v2 26/62] xen: introduce rangeset_claim_range Wei Liu
2018-01-12 11:28 ` [PATCH v2 27/62] xen/pvshim: keep track of used PFN ranges Wei Liu
2018-01-12 11:28 ` [PATCH v2 28/62] x86/guest: map shared_info page Wei Liu
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 [this message]
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
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 ` [PATCH v2 33/62] x86: read wallclock from Xen when running in pvh mode 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
2018-01-12 11:28 ` [PATCH v2 35/62] x86/guest: setup event channel upcall vector Wei Liu
2018-01-12 11:28 ` [PATCH v2 36/62] x86/guest: add PV console code Wei Liu
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 ` [PATCH v2 38/62] x86/shim: Kconfig and command line options Wei Liu
2018-01-12 11:28 ` [PATCH v2 39/62] tools/firmware: Build and install xen-shim Wei Liu
2018-01-12 11:28 ` [PATCH v2 40/62] xen/x86: make VGA support selectable Wei Liu
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 ` [PATCH v2 42/62] sched/null: skip vCPUs on the waitqueue that are blocked Wei Liu
2018-01-12 11:28 ` [PATCH v2 43/62] xen/pvshim: skip Dom0-only domain builder parts Wei Liu
2018-01-12 11:28 ` [PATCH v2 44/62] xen: mark xenstore/console pages as RAM Wei Liu
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 ` [PATCH v2 46/62] xen/pvshim: set correct domid value Wei Liu
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 ` [PATCH v2 48/62] xen/pvshim: add grant table operations Wei Liu
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 ` [PATCH v2 50/62] xen/pvshim: add migration support Wei Liu
2018-01-12 11:28 ` [PATCH v2 51/62] xen/pvshim: add shim_mem cmdline parameter Wei Liu
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 ` [PATCH v2 53/62] xen/pvshim: support vCPU hotplug Wei Liu
2018-01-12 11:29 ` [PATCH v2 54/62] xen/pvshim: memory hotplug Wei Liu
2018-01-12 11:29 ` [PATCH v2 55/62] xen/shim: modify shim_mem parameter behaviour Wei Liu
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 ` [PATCH v2 57/62] xen/shim: crash instead of reboot in shim mode Wei Liu
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 ` [PATCH v2 59/62] libxl: pvshim: Provide first-class config settings to enable shim mode 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
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
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180112112910.16762-31-wei.liu2@citrix.com \
--to=wei.liu2@citrix.com \
--cc=xen-devel@lists.xenproject.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).