Patchwork [05/11] Allocate physical memory in low virtual address space

login
register
mail settings
Submitter Alexander Graf
Date Nov. 26, 2009, 1:23 p.m.
Message ID <1259241800-2810-6-git-send-email-agraf@suse.de>
Download mbox | patch
Permalink /patch/39520/
State New
Headers show

Comments

Alexander Graf - Nov. 26, 2009, 1:23 p.m.
KVM on S390x requires the virtual address space of the guest's RAM to be
within the first 256GB.

The general direction I'd like to see KVM on S390 move is that this requirement
is losened, but for now that's what we're stuck with.

So let's just hack up qemu_ram_alloc until KVM behaves nicely :-).

Signed-off-by: Alexander Graf <agraf@suse.de>
---
 exec.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
Aurelien Jarno - Nov. 30, 2009, 6:18 p.m.
On Thu, Nov 26, 2009 at 02:23:14PM +0100, Alexander Graf wrote:
> KVM on S390x requires the virtual address space of the guest's RAM to be
> within the first 256GB.
> 

While I have no problem with this patch, I am curious to know the reason
of this. No need to change anything in the patch.

> The general direction I'd like to see KVM on S390 move is that this requirement
> is losened, but for now that's what we're stuck with.
> 
> So let's just hack up qemu_ram_alloc until KVM behaves nicely :-).
>
> Signed-off-by: Alexander Graf <agraf@suse.de>
> ---
>  exec.c |    6 ++++++
>  1 files changed, 6 insertions(+), 0 deletions(-)
> 
> diff --git a/exec.c b/exec.c
> index 076d26b..59150d0 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -2411,7 +2411,13 @@ ram_addr_t qemu_ram_alloc(ram_addr_t size)
>      size = TARGET_PAGE_ALIGN(size);
>      new_block = qemu_malloc(sizeof(*new_block));
>  
> +#if defined(TARGET_S390) && defined(CONFIG_KVM)
> +    /* XXX S390 KVM requires the topmost vma of the RAM to be < 256GB */
> +    new_block->host = mmap(0x1000000, size, PROT_EXEC|PROT_READ|PROT_WRITE,
> +                           MAP_SHARED | MAP_ANONYMOUS, -1, 0);
> +#else
>      new_block->host = qemu_vmalloc(size);
> +#endif
>  #ifdef MADV_MERGEABLE
>      madvise(new_block->host, size, MADV_MERGEABLE);
>  #endif
> -- 
> 1.6.0.2
> 
> 
> 
>
Alexander Graf - Nov. 30, 2009, 10:21 p.m.
On 30.11.2009, at 19:18, Aurelien Jarno wrote:

> On Thu, Nov 26, 2009 at 02:23:14PM +0100, Alexander Graf wrote:
>> KVM on S390x requires the virtual address space of the guest's RAM to be
>> within the first 256GB.
>> 
> 
> While I have no problem with this patch, I am curious to know the reason
> of this. No need to change anything in the patch.

For the technical reason I hope Carsten can provide some insight.
As for the implementation reason - S390 KVM (the kernel part) doesn't do memslots properly :-).

Alex
Carsten Otte - Dec. 1, 2009, 9:12 a.m.
Alexander Graf wrote:
> For the technical reason I hope Carsten can provide some insight.
> As for the implementation reason - S390 KVM (the kernel part) doesn't do memslots properly :-).
I'm sorry, I cannot comment on that :-(.

Patch

diff --git a/exec.c b/exec.c
index 076d26b..59150d0 100644
--- a/exec.c
+++ b/exec.c
@@ -2411,7 +2411,13 @@  ram_addr_t qemu_ram_alloc(ram_addr_t size)
     size = TARGET_PAGE_ALIGN(size);
     new_block = qemu_malloc(sizeof(*new_block));
 
+#if defined(TARGET_S390) && defined(CONFIG_KVM)
+    /* XXX S390 KVM requires the topmost vma of the RAM to be < 256GB */
+    new_block->host = mmap(0x1000000, size, PROT_EXEC|PROT_READ|PROT_WRITE,
+                           MAP_SHARED | MAP_ANONYMOUS, -1, 0);
+#else
     new_block->host = qemu_vmalloc(size);
+#endif
 #ifdef MADV_MERGEABLE
     madvise(new_block->host, size, MADV_MERGEABLE);
 #endif