windows patch seems to have broken allocation on demand patches

windows patch seems to have broken allocation on demand patches

[Jes Sorensen]

Hi Xiantao,

The recent changes adding support for windows to KVM/ia64, seems to have
broken the allocation on demand patches, needed to boot with large
memory support.

Is there any chance you have an update version of these patches?

Cheers,
Jes

RE: windows patch seems to have broken allocation on demand patches

[Zhang, Xiantao]

Which patches ?
Xiantao 

-----Original Message-----
From: Jes Sorensen [mailto:jes@sgi.com] 
Sent: Saturday, March 28, 2009 12:00 AM
To: Zhang, Xiantao
Cc: kvm-ia64@vger.kernel.org
Subject: windows patch seems to have broken allocation on demand patches

Hi Xiantao,

The recent changes adding support for windows to KVM/ia64, seems to have
broken the allocation on demand patches, needed to boot with large
memory support.

Is there any chance you have an update version of these patches?

Cheers,
Jes

Re: windows patch seems to have broken allocation on demand patches

[Jes Sorensen]

[-- Attachment #1: Type: text/plain, Size: 781 bytes --]

Zhang, Xiantao wrote:
> Which patches ?
> Xiantao 

This is the one I keep applying - but modified to be against the latest
kernel, which is broken.

You came up with this due to the long startup time for KVM when booting
large guests, like 64GB+ etc.

Cheers,
Jes



> -----Original Message-----
> From: Jes Sorensen [mailto:jes@sgi.com] 
> Sent: Saturday, March 28, 2009 12:00 AM
> To: Zhang, Xiantao
> Cc: kvm-ia64@vger.kernel.org
> Subject: windows patch seems to have broken allocation on demand patches
> 
> Hi Xiantao,
> 
> The recent changes adding support for windows to KVM/ia64, seems to have
> broken the allocation on demand patches, needed to boot with large
> memory support.
> 
> Is there any chance you have an update version of these patches?
> 
> Cheers,
> Jes


[-- Attachment #2: 4000-kvm-ia64-allocate_on_demand.patch --]
[-- Type: text/x-patch, Size: 4992 bytes --]

---
 arch/ia64/include/asm/kvm_host.h |    8 ++++++
 arch/ia64/kvm/kvm-ia64.c         |   49 ++++++++++++++++++++++++++++++++-------
 arch/ia64/kvm/misc.h             |    7 +++++
 arch/ia64/kvm/vtlb.c             |   29 ++++++++++++++++++++++-
 4 files changed, 84 insertions(+), 9 deletions(-)

Index: linux-2.6.git/arch/ia64/include/asm/kvm_host.h
===================================================================
--- linux-2.6.git.orig/arch/ia64/include/asm/kvm_host.h
+++ linux-2.6.git/arch/ia64/include/asm/kvm_host.h
@@ -39,6 +39,7 @@
 #define EXIT_REASON_EXTERNAL_INTERRUPT	6
 #define EXIT_REASON_IPI			7
 #define EXIT_REASON_PTC_G		8
+#define EXIT_REASON_ALLOC_MEM		9
 #define EXIT_REASON_DEBUG		20
 
 /*Define vmm address space and vm data space.*/
@@ -313,6 +314,12 @@
 	struct kvm_vcpu *vcpu;
 };
 
+/* Alloc real memory exit */
+struct kvm_alloc_mem {
+	unsigned long gpfn;
+	unsigned long pmt_val;
+};
+
 /*Exit control data */
 struct exit_ctl_data{
 	uint32_t exit_reason;
@@ -324,6 +331,7 @@
 		struct kvm_switch_rr6	rr_data;
 		struct kvm_ipi_data	ipi_data;
 		struct kvm_ptc_g	ptc_g_data;
+		struct kvm_alloc_mem    alloc_mem;
 	} u;
 };
 
Index: linux-2.6.git/arch/ia64/kvm/kvm-ia64.c
===================================================================
--- linux-2.6.git.orig/arch/ia64/kvm/kvm-ia64.c
+++ linux-2.6.git/arch/ia64/kvm/kvm-ia64.c
@@ -212,6 +212,8 @@
 {
 	kvm_run->exit_reason = KVM_EXIT_UNKNOWN;
 	kvm_run->hw.hardware_exit_reason = 1;
+	printk(KERN_ERR"KVM: VM error occurs!");
+
 	return 0;
 }
 
@@ -481,6 +483,40 @@
 	return 1;
 }
 
+static int handle_mem_alloc(struct kvm_vcpu *vcpu,
+		struct kvm_run *kvm_run)
+{
+	unsigned long pmt_val, gpfn, pfn, gpfn_off;
+	struct kvm_memory_slot *memslot;
+	struct exit_ctl_data *p = kvm_get_exit_data(vcpu);
+
+	gpfn = p->u.alloc_mem.gpfn;
+	
+	spin_lock(&vcpu->kvm->mmu_lock);
+	pmt_val = kvm_get_pmt_entry(vcpu->kvm, gpfn);
+	if (!pmt_val) {
+
+		pfn = gfn_to_pfn(vcpu->kvm, gpfn);
+		if (!pfn_valid(pfn))
+			goto out;
+
+		kvm_set_pmt_entry(vcpu->kvm, gpfn, pfn << PAGE_SHIFT,
+						_PAGE_AR_RWX | _PAGE_MA_WB);
+
+		memslot = gfn_to_memslot(vcpu->kvm, gpfn);
+		if (!memslot)
+			goto out;
+		gpfn_off =  gpfn - memslot->base_gfn;
+		memslot->rmap[gpfn_off] = (unsigned long)pfn_to_page(pfn);
+		pmt_val = kvm_get_pmt_entry(vcpu->kvm, gpfn);
+	}
+out:
+	spin_unlock(&vcpu->kvm->mmu_lock);
+	p->u.alloc_mem.pmt_val = pmt_val;
+
+	return 1;
+}
+
 static int (*kvm_vti_exit_handlers[])(struct kvm_vcpu *vcpu,
 		struct kvm_run *kvm_run) = {
 	[EXIT_REASON_VM_PANIC]              = handle_vm_error,
@@ -492,6 +528,7 @@
 	[EXIT_REASON_EXTERNAL_INTERRUPT]    = handle_external_interrupt,
 	[EXIT_REASON_IPI]		    = handle_ipi,
 	[EXIT_REASON_PTC_G]		    = handle_global_purge,
+	[EXIT_REASON_ALLOC_MEM]		    = handle_mem_alloc,
 	[EXIT_REASON_DEBUG]		    = handle_vcpu_debug,
 
 };
@@ -1554,18 +1591,14 @@
 
 	for (i = 0; i < npages; i++) {
 		pfn = gfn_to_pfn(kvm, base_gfn + i);
-		if (!kvm_is_mmio_pfn(pfn)) {
-			kvm_set_pmt_entry(kvm, base_gfn + i,
-					pfn << PAGE_SHIFT,
-				_PAGE_AR_RWX | _PAGE_MA_WB);
-			memslot->rmap[i] = (unsigned long)pfn_to_page(pfn);
-		} else {
+		if (kvm_is_mmio_pfn(pfn)) {
 			kvm_set_pmt_entry(kvm, base_gfn + i,
 					GPFN_PHYS_MMIO | (pfn << PAGE_SHIFT),
 					_PAGE_MA_UC);
-			memslot->rmap[i] = 0;
-			}
+		}
+		memslot->rmap[i] = 0;
 	}
+	spin_unlock(&kvm->mmu_lock);
 
 	return 0;
 }
Index: linux-2.6.git/arch/ia64/kvm/misc.h
===================================================================
--- linux-2.6.git.orig/arch/ia64/kvm/misc.h
+++ linux-2.6.git/arch/ia64/kvm/misc.h
@@ -41,6 +41,13 @@
 	pmt_base[gfn] = pte;
 }
 
+static inline uint64_t kvm_get_pmt_entry(struct kvm *kvm, gfn_t gfn)
+{
+	uint64_t *pmt_base = kvm_host_get_pmt(kvm);
+
+	return pmt_base[gfn];
+}
+
 /*Function for translating host address to guest address*/
 
 static inline void *to_guest(struct kvm *kvm, void *addr)
Index: linux-2.6.git/arch/ia64/kvm/vtlb.c
===================================================================
--- linux-2.6.git.orig/arch/ia64/kvm/vtlb.c
+++ linux-2.6.git/arch/ia64/kvm/vtlb.c
@@ -558,14 +558,41 @@
 	}
 }
 
+static unsigned long alloc_real_maddr(unsigned long gpfn)
+{
+	struct exit_ctl_data *p = &current_vcpu->arch.exit_data;
+	unsigned long psr;
+
+	local_irq_save(psr);
+
+	p->exit_reason = EXIT_REASON_ALLOC_MEM;
+	p->u.alloc_mem.gpfn = gpfn;
+	p->u.alloc_mem.pmt_val = 0;
+	vmm_transition(current_vcpu);
+
+	local_irq_restore(psr);
+
+	return p->u.alloc_mem.pmt_val;
+}
+
 u64 kvm_get_mpt_entry(u64 gpfn)
 {
 	u64 *base = (u64 *) KVM_P2M_BASE;
+	u64 pmt_val;
 
 	if (gpfn >= (KVM_P2M_SIZE >> 3))
 		panic_vm(current_vcpu, "Invalid gpfn =%lx\n", gpfn);
 
-	return *(base + gpfn);
+	pmt_val = *(base + gpfn);
+
+	if (!pmt_val) {
+		pmt_val = alloc_real_maddr(gpfn);
+		if (!pmt_val) {
+			//printk(KERN_ERR"kvm: NO Enough memory!\n");
+			panic_vm(current_vcpu, "not enough memory\n");
+		}
+	}
+	return pmt_val;
 }
 
 u64 kvm_lookup_mpa(u64 gpfn)

RE: windows patch seems to have broken allocation on demand patches

[Zhang, Xiantao]

Could you attach the kernel message info dumped by dmesg after breaking windows ? Actually, I am not aware of this. 
Xiantao 

-----Original Message-----
From: Jes Sorensen [mailto:jes@sgi.com] 
Sent: Saturday, March 28, 2009 12:20 AM
To: Zhang, Xiantao
Cc: kvm-ia64@vger.kernel.org
Subject: Re: windows patch seems to have broken allocation on demand patches

Zhang, Xiantao wrote:
> Which patches ?
> Xiantao 

This is the one I keep applying - but modified to be against the latest
kernel, which is broken.

You came up with this due to the long startup time for KVM when booting
large guests, like 64GB+ etc.

Cheers,
Jes



> -----Original Message-----
> From: Jes Sorensen [mailto:jes@sgi.com] 
> Sent: Saturday, March 28, 2009 12:00 AM
> To: Zhang, Xiantao
> Cc: kvm-ia64@vger.kernel.org
> Subject: windows patch seems to have broken allocation on demand patches
> 
> Hi Xiantao,
> 
> The recent changes adding support for windows to KVM/ia64, seems to have
> broken the allocation on demand patches, needed to boot with large
> memory support.
> 
> Is there any chance you have an update version of these patches?
> 
> Cheers,
> Jes

Re: windows patch seems to have broken allocation on demand patches

[Jes Sorensen]

Zhang, Xiantao wrote:
> Could you attach the kernel message info dumped by dmesg after breaking windows ? Actually, I am not aware of this. 
> Xiantao 

Hi Xiantao,

No sorry, I think I wasn't clear. After I did a pull and the windows
patches were applied, it seems KVM no longer works here. I had a
conflict in the attached patch with the windows changes and I believe
that caused it - I will try and investigate further on Monday to see if
I can track it down.

Regards,
Jes

RE: windows patch seems to have broken allocation on demand patches

[Zhang, Xiantao]

Jes Sorensen wrote:
> Zhang, Xiantao wrote:
>> Could you attach the kernel message info dumped by dmesg after
>> breaking windows ? Actually, I am not aware of this. Xiantao
> 
> Hi Xiantao,
> 
> No sorry, I think I wasn't clear. After I did a pull and the windows
> patches were applied, it seems KVM no longer works here. I had a
> conflict in the attached patch with the windows changes and I believe
> that caused it - I will try and investigate further on Monday to see
> if I can track it down.

Okay, I think they have no much big changes in windows patch, should be easy to make it work again. Thanks!
Xiantao