2010-07-22 05:58:26

by Yasuaki Ishimatsu

Subject: [PATCH] x86: ioremap: fix wrong physical address handling in PAT code

From: Yasuaki Ishimatsu <[email protected]>

The following two commits fixed a problem that x86 ioremap() doesn't handle
physical address higher than 32-bit properly in X86_32 PAE mode.

ffa71f33a820d1ab3f2fc5723819ac60fb76080b (x86, ioremap: Fix incorrect physical address handling in
PAE mode)
35be1b716a475717611b2dc04185e9d80b9cb693 (x86, ioremap: Fix normal ram range check)

But these fixes are not enough, since pat_pagerange_is_ram() in PAT code
also has a same problem. This patch fixes it.

Signed-off-by: Yasuaki Ishimatsu <[email protected]>

---

It is for -tip tree.

Index: linux-2.6-x86/arch/x86/mm/pat.c
===================================================================
--- linux-2.6-x86.orig/arch/x86/mm/pat.c 2010-07-22 03:04:38.000000000 +0900
+++ linux-2.6-x86/arch/x86/mm/pat.c 2010-07-22 03:47:36.000000000 +0900
@@ -158,7 +158,7 @@ static unsigned long pat_x_mtrr_type(u64
return req_type;
}

-static int pat_pagerange_is_ram(unsigned long start, unsigned long end)
+static int pat_pagerange_is_ram(resource_size_t start, resource_size_t end)
{
int ram_page = 0, not_rampage = 0;
unsigned long page_nr;

2010-07-22 15:14:39

by Kenji Kaneshige

Subject: Re: [PATCH] x86: ioremap: fix wrong physical address handling in PAT code

Good catch!

Reviewed-by: Kenji Kaneshige <[email protected]>


(2010/07/22 14:57), Yasuaki Ishimatsu wrote:
> From: Yasuaki Ishimatsu<[email protected]>
>
> The following two commits fixed a problem that x86 ioremap() doesn't handle
> physical address higher than 32-bit properly in X86_32 PAE mode.
>
> ffa71f33a820d1ab3f2fc5723819ac60fb76080b (x86, ioremap: Fix incorrect physical address handling in
> PAE mode)
> 35be1b716a475717611b2dc04185e9d80b9cb693 (x86, ioremap: Fix normal ram range check)
>
> But these fixes are not enough, since pat_pagerange_is_ram() in PAT code
> also has a same problem. This patch fixes it.
>
> Signed-off-by: Yasuaki Ishimatsu<[email protected]>
>
> ---
>
> It is for -tip tree.
>
> Index: linux-2.6-x86/arch/x86/mm/pat.c
> ===================================================================
> --- linux-2.6-x86.orig/arch/x86/mm/pat.c 2010-07-22 03:04:38.000000000 +0900
> +++ linux-2.6-x86/arch/x86/mm/pat.c 2010-07-22 03:47:36.000000000 +0900
> @@ -158,7 +158,7 @@ static unsigned long pat_x_mtrr_type(u64
> return req_type;
> }
>
> -static int pat_pagerange_is_ram(unsigned long start, unsigned long end)
> +static int pat_pagerange_is_ram(resource_size_t start, resource_size_t end)
> {
> int ram_page = 0, not_rampage = 0;
> unsigned long page_nr;
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>

2010-07-29 05:59:53

by Yasuaki Ishimatsu

Subject: Re: [PATCH] x86: ioremap: fix wrong physical address handling in PAT code

Hi,

Does anybody see the following patch? It's obviously a bug. Please apply it.

Thanks,
Yasuaki Ishimatsu

> From: Yasuaki Ishimatsu <[email protected]>
>
> The following two commits fixed a problem that x86 ioremap() doesn't handle
> physical address higher than 32-bit properly in X86_32 PAE mode.
>
> ffa71f33a820d1ab3f2fc5723819ac60fb76080b (x86, ioremap: Fix incorrect physical address handling in
> PAE mode)
> 35be1b716a475717611b2dc04185e9d80b9cb693 (x86, ioremap: Fix normal ram range check)
>
> But these fixes are not enough, since pat_pagerange_is_ram() in PAT code
> also has a same problem. This patch fixes it.
>
> Signed-off-by: Yasuaki Ishimatsu <[email protected]>
>
> ---
>
> It is for -tip tree.
>
> Index: linux-2.6-x86/arch/x86/mm/pat.c
> ===================================================================
> --- linux-2.6-x86.orig/arch/x86/mm/pat.c 2010-07-22 03:04:38.000000000 +0900
> +++ linux-2.6-x86/arch/x86/mm/pat.c 2010-07-22 03:47:36.000000000 +0900
> @@ -158,7 +158,7 @@ static unsigned long pat_x_mtrr_type(u64
> return req_type;
> }
>
> -static int pat_pagerange_is_ram(unsigned long start, unsigned long end)
> +static int pat_pagerange_is_ram(resource_size_t start, resource_size_t end)
> {
> int ram_page = 0, not_rampage = 0;
> unsigned long page_nr;
>

2010-07-29 13:28:07

by Yasuaki Ishimatsu

Subject: [tip:x86/mm] x86: Ioremap: fix wrong physical address handling in PAT code

Commit-ID: 3709c857350976408953831f0cf89d19951394a1
Gitweb: http://git.kernel.org/tip/3709c857350976408953831f0cf89d19951394a1
Author: Yasuaki Ishimatsu <[email protected]>
AuthorDate: Thu, 22 Jul 2010 14:57:35 +0900
Committer: Thomas Gleixner <[email protected]>
CommitDate: Thu, 29 Jul 2010 15:26:11 +0200

x86: Ioremap: fix wrong physical address handling in PAT code

The following two commits fixed a problem that x86 ioremap() doesn't handle
physical address higher than 32-bit properly in X86_32 PAE mode.

ffa71f33a820d1ab3f2fc5723819ac60fb76080b (x86, ioremap: Fix incorrect
physical address handling in PAE mode)

35be1b716a475717611b2dc04185e9d80b9cb693 (x86, ioremap: Fix normal
ram range check)

But these fixes are not enough, since pat_pagerange_is_ram() in PAT code
also has a same problem. This patch fixes it.

Signed-off-by: Yasuaki Ishimatsu <[email protected]>
Reviewed-by: Kenji Kaneshige <[email protected]>
LKML-Reference: <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>

---
arch/x86/mm/pat.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index acc15b2..03b48c8 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -158,7 +158,7 @@ static unsigned long pat_x_mtrr_type(u64 start, u64 end, unsigned long req_type)
return req_type;
}

-static int pat_pagerange_is_ram(unsigned long start, unsigned long end)
+static int pat_pagerange_is_ram(resource_size_t start, resource_size_t end)
{
int ram_page = 0, not_rampage = 0;
unsigned long page_nr;