[PATCH v2 0/6] Xen PV on HVM fixes and improvements

[PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Stefano Stabellini]

Hi all,
this patch series is a collection of fixes and improvements for Linux
running as Xen PV on HVM guest.
Changes to the previous version:

- patch 5 and 6 have been squashed together;

- xen_hvm_spinlock_init has been merged into xen_hvm_smp_prepare_cpus.


The list of patches with diffstat follows:

Stefano Stabellini (6):
      xen: no need to delay xen_setup_shutdown_event for hvm guests anymore
      xen: do not use xen_info on HVM, set pv_info name to "Xen HVM"
      xen-blkfront: handle Xen major numbers other than XENVBD
      xen: make the ballon driver work for hvm domains
      xen: PV on HVM: support PV spinlocks and IPIs
      xen: fix compile issue if XEN is enabled but XEN_PVHVM is disabled

 arch/x86/xen/enlighten.c         |    6 ++-
 arch/x86/xen/smp.c               |   38 ++++++++++++++++++
 arch/x86/xen/suspend.c           |    2 +
 arch/x86/xen/xen-ops.h           |    2 +
 drivers/block/xen-blkfront.c     |   79 +++++++++++++++++++++++++++++++++++--
 drivers/xen/balloon.c            |   14 ++++--
 drivers/xen/manage.c             |   17 ++------
 drivers/xen/platform-pci.c       |    3 -
 include/xen/interface/io/blkif.h |   21 ++++++++++
 9 files changed, 154 insertions(+), 28 deletions(-)

A branch with these patches on 2.6.38-rc6 is available here:

git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git 2.6.38-rc6-pvhvm

Cheers,

Stefano

[PATCH v2 1/6] xen: no need to delay xen_setup_shutdown_event for hvm guests anymore

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Now that xenstore_ready is used correctly for PV on HVM guests too, we
don't need to delay the initialization of xen_setup_shutdown_event
anymore.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
---
 drivers/xen/manage.c       |   17 ++++-------------
 drivers/xen/platform-pci.c |    3 ---
 2 files changed, 4 insertions(+), 16 deletions(-)

diff --git a/drivers/xen/manage.c b/drivers/xen/manage.c
index 2417727..b2a8d78 100644
--- a/drivers/xen/manage.c
+++ b/drivers/xen/manage.c
@@ -291,27 +291,18 @@ static int shutdown_event(struct notifier_block *notifier,
 	return NOTIFY_DONE;
 }
 
-static int __init __setup_shutdown_event(void)
-{
-	/* Delay initialization in the PV on HVM case */
-	if (xen_hvm_domain())
-		return 0;
-
-	if (!xen_pv_domain())
-		return -ENODEV;
-
-	return xen_setup_shutdown_event();
-}
-
 int xen_setup_shutdown_event(void)
 {
 	static struct notifier_block xenstore_notifier = {
 		.notifier_call = shutdown_event
 	};
+
+	if (!xen_domain())
+		return -ENODEV;
 	register_xenstore_notifier(&xenstore_notifier);
 
 	return 0;
 }
 EXPORT_SYMBOL_GPL(xen_setup_shutdown_event);
 
-subsys_initcall(__setup_shutdown_event);
+subsys_initcall(xen_setup_shutdown_event);
diff --git a/drivers/xen/platform-pci.c b/drivers/xen/platform-pci.c
index afbe041..319dd0a 100644
--- a/drivers/xen/platform-pci.c
+++ b/drivers/xen/platform-pci.c
@@ -156,9 +156,6 @@ static int __devinit platform_pci_init(struct pci_dev *pdev,
 	if (ret)
 		goto out;
 	xenbus_probe(NULL);
-	ret = xen_setup_shutdown_event();
-	if (ret)
-		goto out;
 	return 0;
 
 out:
-- 
1.5.6.5

[PATCH v2 2/6] xen: do not use xen_info on HVM, set pv_info name to "Xen HVM"

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
---
 arch/x86/xen/enlighten.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)

diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 50542ef..2f67e2e 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1284,8 +1284,7 @@ static int init_hvm_pv_info(int *major, int *minor)
 
 	xen_setup_features();
 
-	pv_info = xen_info;
-	pv_info.kernel_rpl = 0;
+	pv_info.name = "Xen HVM";
 
 	xen_domain_type = XEN_HVM_DOMAIN;
 
-- 
1.5.6.5

[PATCH v2 3/6] xen-blkfront: handle Xen major numbers other than XENVBD

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

This patch makes sure blkfront handles correctly virtual device numbers
corresponding to Xen emulated IDE and SCSI disks: in those cases
blkfront translates the major number to XENVBD and the minor number to a
low xvd minor.

Note: this behaviour is different from what old xenlinux PV guests used
to do: they used to steal an IDE or SCSI major number and use it
instead.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
---
 drivers/block/xen-blkfront.c     |   79 +++++++++++++++++++++++++++++++++++--
 include/xen/interface/io/blkif.h |   21 ++++++++++
 2 files changed, 95 insertions(+), 5 deletions(-)

diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index d7aa39e..64d9c6d 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -120,6 +120,10 @@ static DEFINE_SPINLOCK(minor_lock);
 #define EXTENDED (1<<EXT_SHIFT)
 #define VDEV_IS_EXTENDED(dev) ((dev)&(EXTENDED))
 #define BLKIF_MINOR_EXT(dev) ((dev)&(~EXTENDED))
+#define EMULATED_HD_DISK_MINOR_OFFSET (0)
+#define EMULATED_HD_DISK_NAME_OFFSET (EMULATED_HD_DISK_MINOR_OFFSET / 256)
+#define EMULATED_SD_DISK_MINOR_OFFSET (EMULATED_HD_DISK_MINOR_OFFSET + (4 * 16))
+#define EMULATED_SD_DISK_NAME_OFFSET (EMULATED_HD_DISK_NAME_OFFSET + 4)
 
 #define DEV_NAME	"xvd"	/* name in /dev */
 
@@ -434,6 +438,65 @@ static void xlvbd_flush(struct blkfront_info *info)
 	       info->feature_flush ? "enabled" : "disabled");
 }
 
+static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset)
+{
+	int major;
+	major = BLKIF_MAJOR(vdevice);
+	*minor = BLKIF_MINOR(vdevice);
+	switch (major) {
+		case XEN_IDE0_MAJOR:
+			*offset = (*minor / 64) + EMULATED_HD_DISK_NAME_OFFSET;
+			*minor = ((*minor / 64) * PARTS_PER_DISK) +
+				EMULATED_HD_DISK_MINOR_OFFSET;
+			break;
+		case XEN_IDE1_MAJOR:
+			*offset = (*minor / 64) + 2 + EMULATED_HD_DISK_NAME_OFFSET;
+			*minor = (((*minor / 64) + 2) * PARTS_PER_DISK) +
+				EMULATED_HD_DISK_MINOR_OFFSET;
+			break;
+		case XEN_SCSI_DISK0_MAJOR:
+			*offset = (*minor / PARTS_PER_DISK) + EMULATED_SD_DISK_NAME_OFFSET;
+			*minor = *minor + EMULATED_SD_DISK_MINOR_OFFSET;
+			break;
+		case XEN_SCSI_DISK1_MAJOR:
+		case XEN_SCSI_DISK2_MAJOR:
+		case XEN_SCSI_DISK3_MAJOR:
+		case XEN_SCSI_DISK4_MAJOR:
+		case XEN_SCSI_DISK5_MAJOR:
+		case XEN_SCSI_DISK6_MAJOR:
+		case XEN_SCSI_DISK7_MAJOR:
+			*offset = (*minor / PARTS_PER_DISK) + 
+				((major - XEN_SCSI_DISK1_MAJOR + 1) * 16) +
+				EMULATED_SD_DISK_NAME_OFFSET;
+			*minor = *minor +
+				((major - XEN_SCSI_DISK1_MAJOR + 1) * 16 * PARTS_PER_DISK) +
+				EMULATED_SD_DISK_MINOR_OFFSET;
+			break;
+		case XEN_SCSI_DISK8_MAJOR:
+		case XEN_SCSI_DISK9_MAJOR:
+		case XEN_SCSI_DISK10_MAJOR:
+		case XEN_SCSI_DISK11_MAJOR:
+		case XEN_SCSI_DISK12_MAJOR:
+		case XEN_SCSI_DISK13_MAJOR:
+		case XEN_SCSI_DISK14_MAJOR:
+		case XEN_SCSI_DISK15_MAJOR:
+			*offset = (*minor / PARTS_PER_DISK) + 
+				((major - XEN_SCSI_DISK8_MAJOR + 8) * 16) +
+				EMULATED_SD_DISK_NAME_OFFSET;
+			*minor = *minor +
+				((major - XEN_SCSI_DISK8_MAJOR + 8) * 16 * PARTS_PER_DISK) +
+				EMULATED_SD_DISK_MINOR_OFFSET;
+			break;
+		case XENVBD_MAJOR:
+			*offset = *minor / PARTS_PER_DISK;
+			break;
+		default:
+			printk(KERN_WARNING "blkfront: your disk configuration is "
+					"incorrect, please use an xvd device instead\n");
+			return -ENODEV;
+	}
+	return 0;
+}
 
 static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 			       struct blkfront_info *info,
@@ -441,7 +504,7 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 {
 	struct gendisk *gd;
 	int nr_minors = 1;
-	int err = -ENODEV;
+	int err;
 	unsigned int offset;
 	int minor;
 	int nr_parts;
@@ -456,12 +519,20 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 	}
 
 	if (!VDEV_IS_EXTENDED(info->vdevice)) {
-		minor = BLKIF_MINOR(info->vdevice);
-		nr_parts = PARTS_PER_DISK;
+		err = xen_translate_vdev(info->vdevice, &minor, &offset);
+		if (err)
+			return err;		
+ 		nr_parts = PARTS_PER_DISK;
 	} else {
 		minor = BLKIF_MINOR_EXT(info->vdevice);
 		nr_parts = PARTS_PER_EXT_DISK;
+		offset = minor / nr_parts;
+		if (xen_hvm_domain() && offset <= EMULATED_HD_DISK_NAME_OFFSET + 4)
+			printk(KERN_WARNING "blkfront: vdevice 0x%x might conflict with "
+					"emulated IDE disks,\n\t choose an xvd device name"
+					"from xvde on\n", info->vdevice);
 	}
+	err = -ENODEV;
 
 	if ((minor % nr_parts) == 0)
 		nr_minors = nr_parts;
@@ -475,8 +546,6 @@ static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 	if (gd == NULL)
 		goto release;
 
-	offset = minor / nr_parts;
-
 	if (nr_minors > 1) {
 		if (offset < 26)
 			sprintf(gd->disk_name, "%s%c", DEV_NAME, 'a' + offset);
diff --git a/include/xen/interface/io/blkif.h b/include/xen/interface/io/blkif.h
index c2d1fa4..68dd2b4 100644
--- a/include/xen/interface/io/blkif.h
+++ b/include/xen/interface/io/blkif.h
@@ -91,4 +91,25 @@ DEFINE_RING_TYPES(blkif, struct blkif_request, struct blkif_response);
 #define VDISK_REMOVABLE    0x2
 #define VDISK_READONLY     0x4
 
+/* Xen-defined major numbers for virtual disks, they look strangely
+ * familiar */
+#define XEN_IDE0_MAJOR	3
+#define XEN_IDE1_MAJOR	22
+#define XEN_SCSI_DISK0_MAJOR	8
+#define XEN_SCSI_DISK1_MAJOR	65
+#define XEN_SCSI_DISK2_MAJOR	66
+#define XEN_SCSI_DISK3_MAJOR	67
+#define XEN_SCSI_DISK4_MAJOR	68
+#define XEN_SCSI_DISK5_MAJOR	69
+#define XEN_SCSI_DISK6_MAJOR	70
+#define XEN_SCSI_DISK7_MAJOR	71
+#define XEN_SCSI_DISK8_MAJOR	128
+#define XEN_SCSI_DISK9_MAJOR	129
+#define XEN_SCSI_DISK10_MAJOR	130
+#define XEN_SCSI_DISK11_MAJOR	131
+#define XEN_SCSI_DISK12_MAJOR	132
+#define XEN_SCSI_DISK13_MAJOR	133
+#define XEN_SCSI_DISK14_MAJOR	134
+#define XEN_SCSI_DISK15_MAJOR	135
+
 #endif /* __XEN_PUBLIC_IO_BLKIF_H__ */
-- 
1.5.6.5

[PATCH v2 4/6] xen: make the ballon driver work for hvm domains

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 drivers/xen/balloon.c |   14 +++++++++-----
 1 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/drivers/xen/balloon.c b/drivers/xen/balloon.c
index 43f9f02..9294f25 100644
--- a/drivers/xen/balloon.c
+++ b/drivers/xen/balloon.c
@@ -232,7 +232,7 @@ static int increase_reservation(unsigned long nr_pages)
 		set_phys_to_machine(pfn, frame_list[i]);
 
 		/* Link back into the page tables if not highmem. */
-		if (pfn < max_low_pfn) {
+		if (!xen_hvm_domain() && pfn < max_low_pfn) {
 			int ret;
 			ret = HYPERVISOR_update_va_mapping(
 				(unsigned long)__va(pfn << PAGE_SHIFT),
@@ -280,7 +280,7 @@ static int decrease_reservation(unsigned long nr_pages)
 
 		scrub_page(page);
 
-		if (!PageHighMem(page)) {
+		if (!xen_hvm_domain() && !PageHighMem(page)) {
 			ret = HYPERVISOR_update_va_mapping(
 				(unsigned long)__va(pfn << PAGE_SHIFT),
 				__pte_ma(0), 0);
@@ -392,15 +392,19 @@ static struct notifier_block xenstore_notifier;
 
 static int __init balloon_init(void)
 {
-	unsigned long pfn, extra_pfn_end;
+ 	unsigned long pfn, nr_pages, extra_pfn_end;
 	struct page *page;
 
-	if (!xen_pv_domain())
+	if (!xen_domain())
 		return -ENODEV;
 
 	pr_info("xen_balloon: Initialising balloon driver.\n");
 
-	balloon_stats.current_pages = min(xen_start_info->nr_pages, max_pfn);
+ 	if (xen_pv_domain())
+ 		nr_pages = xen_start_info->nr_pages;
+ 	else
+ 		nr_pages = max_pfn;
+ 	balloon_stats.current_pages = min(nr_pages, max_pfn);
 	balloon_stats.target_pages  = balloon_stats.current_pages;
 	balloon_stats.balloon_low   = 0;
 	balloon_stats.balloon_high  = 0;
-- 
1.5.6.5

[PATCH v2 5/6] xen: PV on HVM: support PV spinlocks and IPIs

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Initialize PV spinlocks on boot CPU right after native_smp_prepare_cpus
(that switch to APIC mode and initialize APIC routing); on secondary
CPUs on CPU_UP_PREPARE.

Enable the usage of event channels to send and receive IPIs when
running as a PV on HVM guest.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 arch/x86/xen/enlighten.c |    3 +++
 arch/x86/xen/smp.c       |   38 ++++++++++++++++++++++++++++++++++++++
 arch/x86/xen/xen-ops.h   |    2 ++
 3 files changed, 43 insertions(+), 0 deletions(-)

diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 2f67e2e..fe02574 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1330,6 +1330,8 @@ static int __cpuinit xen_hvm_cpu_notify(struct notifier_block *self,
 	switch (action) {
 	case CPU_UP_PREPARE:
 		per_cpu(xen_vcpu, cpu) = &HYPERVISOR_shared_info->vcpu_info[cpu];
+		if (xen_have_vector_callback)
+			xen_init_lock_cpu(cpu);
 		break;
 	default:
 		break;
@@ -1354,6 +1356,7 @@ static void __init xen_hvm_guest_init(void)
 
 	if (xen_feature(XENFEAT_hvm_callback_vector))
 		xen_have_vector_callback = 1;
+	xen_hvm_smp_init();
 	register_cpu_notifier(&xen_hvm_cpu_notifier);
 	xen_unplug_emulated_devices();
 	have_vcpu_info_placement = 0;
diff --git a/arch/x86/xen/smp.c b/arch/x86/xen/smp.c
index 72a4c79..3061244 100644
--- a/arch/x86/xen/smp.c
+++ b/arch/x86/xen/smp.c
@@ -509,3 +509,41 @@ void __init xen_smp_init(void)
 	xen_fill_possible_map();
 	xen_init_spinlocks();
 }
+
+static void __init xen_hvm_smp_prepare_cpus(unsigned int max_cpus)
+{
+	native_smp_prepare_cpus(max_cpus);
+	WARN_ON(xen_smp_intr_init(0));
+
+	if (!xen_have_vector_callback)
+		return;
+	xen_init_lock_cpu(0);
+	xen_init_spinlocks();
+}
+
+static int __cpuinit xen_hvm_cpu_up(unsigned int cpu)
+{
+	int rc;
+	rc = native_cpu_up(cpu);
+	WARN_ON (xen_smp_intr_init(cpu));
+	return rc;
+}
+
+static void xen_hvm_cpu_die(unsigned int cpu)
+{
+	unbind_from_irqhandler(per_cpu(xen_resched_irq, cpu), NULL);
+	unbind_from_irqhandler(per_cpu(xen_callfunc_irq, cpu), NULL);
+	unbind_from_irqhandler(per_cpu(xen_debug_irq, cpu), NULL);
+	unbind_from_irqhandler(per_cpu(xen_callfuncsingle_irq, cpu), NULL);
+	native_cpu_die(cpu);
+}
+
+void __init xen_hvm_smp_init(void)
+{
+	smp_ops.smp_prepare_cpus = xen_hvm_smp_prepare_cpus;
+	smp_ops.smp_send_reschedule = xen_smp_send_reschedule;
+	smp_ops.cpu_up = xen_hvm_cpu_up;
+	smp_ops.cpu_die = xen_hvm_cpu_die;
+	smp_ops.send_call_func_ipi = xen_smp_send_call_function_ipi;
+	smp_ops.send_call_func_single_ipi = xen_smp_send_call_function_single_ipi;
+}
diff --git a/arch/x86/xen/xen-ops.h b/arch/x86/xen/xen-ops.h
index 9d41bf9..3112f55 100644
--- a/arch/x86/xen/xen-ops.h
+++ b/arch/x86/xen/xen-ops.h
@@ -64,10 +64,12 @@ void xen_setup_vcpu_info_placement(void);
 
 #ifdef CONFIG_SMP
 void xen_smp_init(void);
+void __init xen_hvm_smp_init(void);
 
 extern cpumask_var_t xen_cpu_initialized_map;
 #else
 static inline void xen_smp_init(void) {}
+static inline void xen_hvm_smp_init(void) {}
 #endif
 
 #ifdef CONFIG_PARAVIRT_SPINLOCKS
-- 
1.5.6.5

[PATCH v2 6/6] xen: fix compile issue if XEN is enabled but XEN_PVHVM is disabled

[stefano.stabellini]

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 arch/x86/xen/suspend.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/arch/x86/xen/suspend.c b/arch/x86/xen/suspend.c
index 9bbd63a..4a3d3dd 100644
--- a/arch/x86/xen/suspend.c
+++ b/arch/x86/xen/suspend.c
@@ -28,6 +28,7 @@ void xen_pre_suspend(void)
 
 void xen_hvm_post_suspend(int suspend_cancelled)
 {
+#ifdef CONFIG_XEN_PVHVM
 	int cpu;
 	xen_hvm_init_shared_info();
 	xen_callback_vector();
@@ -37,6 +38,7 @@ void xen_hvm_post_suspend(int suspend_cancelled)
 			xen_setup_runstate_info(cpu);
 		}
 	}
+#endif
 }
 
 void xen_post_suspend(int suspend_cancelled)
-- 
1.5.6.5

Re: [PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Konrad Rzeszutek Wilk]

On Fri, Feb 25, 2011 at 05:11:33PM +0000, Stefano Stabellini wrote:
> Hi all,
> this patch series is a collection of fixes and improvements for Linux
> running as Xen PV on HVM guest.
> Changes to the previous version:

Looks good to me. I am surprised that it was that easy to get the balloon
code to work under HVM..

> 
> - patch 5 and 6 have been squashed together;
> 
> - xen_hvm_spinlock_init has been merged into xen_hvm_smp_prepare_cpus.
> 
> 
> The list of patches with diffstat follows:
> 
> Stefano Stabellini (6):
>       xen: no need to delay xen_setup_shutdown_event for hvm guests anymore
>       xen: do not use xen_info on HVM, set pv_info name to "Xen HVM"
>       xen-blkfront: handle Xen major numbers other than XENVBD
>       xen: make the ballon driver work for hvm domains
>       xen: PV on HVM: support PV spinlocks and IPIs
>       xen: fix compile issue if XEN is enabled but XEN_PVHVM is disabled
> 
>  arch/x86/xen/enlighten.c         |    6 ++-
>  arch/x86/xen/smp.c               |   38 ++++++++++++++++++
>  arch/x86/xen/suspend.c           |    2 +
>  arch/x86/xen/xen-ops.h           |    2 +
>  drivers/block/xen-blkfront.c     |   79 +++++++++++++++++++++++++++++++++++--
>  drivers/xen/balloon.c            |   14 ++++--
>  drivers/xen/manage.c             |   17 ++------
>  drivers/xen/platform-pci.c       |    3 -
>  include/xen/interface/io/blkif.h |   21 ++++++++++
>  9 files changed, 154 insertions(+), 28 deletions(-)
> 
> A branch with these patches on 2.6.38-rc6 is available here:
> 
> git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git 2.6.38-rc6-pvhvm
> 
> Cheers,
> 
> Stefano

Re: [Xen-devel] [PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Pasi Kärkkäinen]

On Thu, Mar 03, 2011 at 10:29:38AM +0800, topperxin wrote:
>    Hi stefano,
>       It's very great of your work.I think it't useful
>    to me, I have checked it out, and I will study it.
>      What I care mostly is the pv-on-hvm drivers, now
>    my main work is port pvonhvm drivers to various guest
>    OS, no matter what OS version, such as ubuntu 10.10,
>    debian 505 etc. But I found this work is difficult &
>    boring to me, I can't find a general method which can
>    easily port the pvonhvm drivers to all kinds guest os.
>    Do you have any good ideas?
>       I found there are not platform-pci source in your
>    project, as we know, in the old pvonhvm drivers version,
>    platform-pci module is important, it was used to establish
>    the communication mechanism, such as xenbus, evtchn,etc,
>    could you please tell me how can you get rid of it? and how
>    can you substitute platform-pci, with which?
>        Thanks a lot.

Hey,

Not sure if you've seen this wiki page..it has some links to
various versions of pv-on-hvm drivers for linux:

http://wiki.xen.org/xenwiki/XenLinuxPVonHVMdrivers

-- Pasi

> 
>  At 2011-02-26 01:11:33**"Stefano Stabellini" <Stefano.Stabellini@eu.citrix.com> wrote:
> 
>  >Hi all,
>  >this patch series is a collection of fixes and improvements for Linux
>  >running as Xen PV on HVM guest.
>  >Changes to the previous version:
>  >
>  >- patch 5 and 6 have been squashed together;
>  >
>  >- xen_hvm_spinlock_init has been merged into xen_hvm_smp_prepare_cpus.
>  >
>  >
>  >The list of patches with diffstat follows:
>  >
>  >Stefano Stabellini (6):
>  >      xen: no need to delay xen_setup_shutdown_event for hvm guests anymore
>  >      xen: do not use xen_info on HVM, set pv_info name to "Xen HVM"
>  >      xen-blkfront: handle Xen major numbers other than XENVBD
>  >      xen: make the ballon driver work for hvm domains
>  >      xen: PV on HVM: support PV spinlocks and IPIs
>  >      xen: fix compile issue if XEN is enabled but XEN_PVHVM is disabled
>  >
>  > arch/x86/xen/enlighten.c         |    6 ++-
>  > arch/x86/xen/smp.c               |   38 ++++++++++++++++++
>  > arch/x86/xen/suspend.c           |    2 +
>  > arch/x86/xen/xen-ops.h           |    2 +
>  > drivers/block/xen-blkfront.c     |   79 +++++++++++++++++++++++++++++++++++--
>  > drivers/xen/balloon.c            |   14 ++++--
>  > drivers/xen/manage.c             |   17 ++------
>  > drivers/xen/platform-pci.c       |    3 -
>  > include/xen/interface/io/blkif.h |   21 ++++++++++
>  > 9 files changed, 154 insertions(+), 28 deletions(-)
>  >
>  >A branch with these patches on 2.6.38-rc6 is available here:
>  >
>  >git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git 2.6.38-rc6-pvhvm
>  >
>  >Cheers,
>  >
>  >Stefano
>  >
>  >_______________________________________________
>  >Xen-devel mailing list
>  >Xen-devel@lists.xensource.com
>  >http://lists.xensource.com/xen-devel

> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re:[Xen-devel] [PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Stefano Stabellini]

[-- Attachment #1: Type: text/plain, Size: 1074 bytes --]

On Thu, 3 Mar 2011, topperxin wrote:
> Hi stefano,
> 
>    It's very great of your work.I think it't useful 
> to me, I have checked it out, and I will study it.
>   What I care mostly is the pv-on-hvm drivers, now 
> my main work is port pvonhvm drivers to various guest
> OS, no matter what OS version, such as ubuntu 10.10, 
> debian 505 etc. But I found this work is difficult & 
> boring to me, I can't find a general method which can 
> easily port the pvonhvm drivers to all kinds guest os.
> Do you have any good ideas?

PV on HVM is upstream now, so it is just a matter of time before all the
distros will have a kernel that supports PV on HVM out of the box.


>    I found there are not platform-pci source in your 
> project, as we know, in the old pvonhvm drivers version,
> platform-pci module is important, it was used to establish
> the communication mechanism, such as xenbus, evtchn,etc, 
> could you please tell me how can you get rid of it? and how 
> can you substitute platform-pci, with which?

Yes there is:
drivers/xen/platform-pci.c

Re:Re:[Xen-devel] [PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Stefano Stabellini]

[-- Attachment #1: Type: text/plain, Size: 1769 bytes --]

On Tue, 8 Mar 2011, topperxin wrote:
> At 2011-03-03 23:05:09，"Stefano Stabellini" <stefano.stabellini@eu.citrix.com> wrote:
> 
> >On Thu, 3 Mar 2011, topperxin wrote:
> >> Hi stefano,
> >> 
> >>    It's very great of your work.I think it't useful 
> >> to me, I have checked it out, and I will study it.
> >>   What I care mostly is the pv-on-hvm drivers, now 
> >> my main work is port pvonhvm drivers to various guest
> >> OS, no matter what OS version, such as ubuntu 10.10, 
> >> debian 505 etc. But I found this work is difficult & 
> >> boring to me, I can't find a general method which can 
> >> easily port the pvonhvm drivers to all kinds guest os.
> >> Do you have any good ideas?
> >
> >PV on HVM is upstream now, so it is just a matter of time before all the
> >distros will have a kernel that supports PV on HVM out of the box.
> >
> >
> >>    I found there are not platform-pci source in your 
> >> project, as we know, in the old pvonhvm drivers version,
> >> platform-pci module is important, it was used to establish
> >> the communication mechanism, such as xenbus, evtchn,etc, 
> >> could you please tell me how can you get rid of it? and how 
> >> can you substitute platform-pci, with which?
> >
> >Yes there is:
> >drivers/xen/platform-pci.c
> 
> 1.sorry, I can't find platform-pci.c in your release source.
> 2.6.38-rc5-pvhvm
> 2.another question is when I make; make install under 2.6.38-rc5-pvhvm
> I found the version number change from 2.6.38 to 2.6.34 under 
> /lib/modules/2.6.34-rc1
> why?

you must be working on the wrong tree :)

Re:Re:Re:[Xen-devel] [PATCH v2 0/6] Xen PV on HVM fixes and improvements

[Stefano Stabellini]

[-- Attachment #1: Type: text/plain, Size: 815 bytes --]

On Tue, 8 Mar 2011, topperxin wrote:
> >you must be working on the wrong tree :)
> 
> git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git 2.6.38-rc6-pvhvm
> isn't it? I get it from this address, you given in the mail.
> cloud you please give me a right address?
> thanks a lot.
 
You probably cloned my master branch: 2.6.38-rc6-pvhvm is the name of a
branch, if you executed:

git clone git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git 2.6.38-rc6-pvhvm

you just cloned the "master" branch in a directory named "2.6.38-rc6-pvhvm".

try:

git remote update
git branch -r

you should get a list of remote branches available, among which
"2.6.38-rc6-pvhvm".

At this point you can checkout the branch 2.6.38-rc6-pvhvm with this
command:

git checkout -b your_branch_name "origin/2.6.38-rc6-pvhvm"