diff mbox series

x86/pat: fix pat_x_mtrr_type() for MTRR disabled case

Message ID 20230110065427.20767-1-jgross@suse.com
State New
Headers
Series x86/pat: fix pat_x_mtrr_type() for MTRR disabled case |

Commit Message

Juergen Gross Jan. 10, 2023, 6:54 a.m. UTC 
  Since commit 72cbc8f04fe2 ("x86/PAT: Have pat_enabled() properly
reflect state when running on Xen") PAT can be enabled without MTRR.

This has resulted in problems e.g. for a SEV-SNP guest running under
Hyper-V, when trying to establish a new mapping via memremap() with
WB caching mode, as pat_x_mtrr_type() will call mtrr_type_lookup(),
which in turn is returning MTRR_TYPE_INVALID due to MTRR being
disabled in this configuration. The result is a mapping with UC-
caching, leading to severe performance degradation.

Fix that by handling MTRR_TYPE_INVALID the same way as MTRR_TYPE_WRBACK
in pat_x_mtrr_type().

Fixes: 72cbc8f04fe2 ("x86/PAT: Have pat_enabled() properly reflect state when running on Xen")
Reported-by: Michael Kelley (LINUX) <mikelley@microsoft.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
---
 arch/x86/mm/pat/memtype.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Michael Kelley (LINUX) Jan. 10, 2023, 3:22 p.m. UTC | #1 
From: Juergen Gross <jgross@suse.com> Sent: Monday, January 9, 2023 10:54 PM
> 
> Since commit 72cbc8f04fe2 ("x86/PAT: Have pat_enabled() properly
> reflect state when running on Xen") PAT can be enabled without MTRR.
> 
> This has resulted in problems e.g. for a SEV-SNP guest running under
> Hyper-V, when trying to establish a new mapping via memremap() with
> WB caching mode, as pat_x_mtrr_type() will call mtrr_type_lookup(),
> which in turn is returning MTRR_TYPE_INVALID due to MTRR being
> disabled in this configuration. The result is a mapping with UC-
> caching, leading to severe performance degradation.
> 
> Fix that by handling MTRR_TYPE_INVALID the same way as MTRR_TYPE_WRBACK
> in pat_x_mtrr_type().
> 
> Fixes: 72cbc8f04fe2 ("x86/PAT: Have pat_enabled() properly reflect state when running on Xen")
> Reported-by: Michael Kelley (LINUX) <mikelley@microsoft.com>
> Signed-off-by: Juergen Gross <jgross@suse.com>
> ---
>  arch/x86/mm/pat/memtype.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
> index 46de9cf5c91d..fb4b1b5e0dea 100644
> --- a/arch/x86/mm/pat/memtype.c
> +++ b/arch/x86/mm/pat/memtype.c
> @@ -387,7 +387,8 @@ static unsigned long pat_x_mtrr_type(u64 start, u64 end,
>  		u8 mtrr_type, uniform;
> 
>  		mtrr_type = mtrr_type_lookup(start, end, &uniform);
> -		if (mtrr_type != MTRR_TYPE_WRBACK)
> +		if (mtrr_type != MTRR_TYPE_WRBACK &&
> +		    mtrr_type != MTRR_TYPE_INVALID)
>  			return _PAGE_CACHE_MODE_UC_MINUS;
> 
>  		return _PAGE_CACHE_MODE_WB;
> --
> 2.35.3

This looks good to me.  I've tested my specific use case of a SEV-SNP guest
running under Hyper-V, where the MTRRs are not visible in the guest but the
PAT is functional.  Previously, this use case showed many "uncached-minus"
entries in /sys/kernel/debug/x86/pat_memtype_list even though the mapping
requests were made as WB.  With this patch, those entries show "write-back",
as expected.

Thanks!

Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Tested-by: Michael Kelley <mikelley@microsoft.com>
diff mbox series

Patch

diff --git a/arch/x86/mm/pat/memtype.c b/arch/x86/mm/pat/memtype.c
index 46de9cf5c91d..fb4b1b5e0dea 100644
--- a/arch/x86/mm/pat/memtype.c
+++ b/arch/x86/mm/pat/memtype.c
@@ -387,7 +387,8 @@  static unsigned long pat_x_mtrr_type(u64 start, u64 end,
 		u8 mtrr_type, uniform;
 
 		mtrr_type = mtrr_type_lookup(start, end, &uniform);
-		if (mtrr_type != MTRR_TYPE_WRBACK)
+		if (mtrr_type != MTRR_TYPE_WRBACK &&
+		    mtrr_type != MTRR_TYPE_INVALID)
 			return _PAGE_CACHE_MODE_UC_MINUS;
 
 		return _PAGE_CACHE_MODE_WB;