[PATCH] arm[64]/memremap: fix arch_memremap_can_ram_remap()

[PATCH] arm[64]/memremap: fix arch_memremap_can_ram_remap()

[Ross Stutterheim]

commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
presence of linear map") added the definition of
arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
can be used from the linear mapping. memblock_is_map_memory() was called
with the pfn of the address given to arch_memremap_can_ram_remap();
however, memblock_is_map_memory() expects to be given an address, not a
pfn.

This results in calls to memremap() returning a newly mapped area when
it should return an address in the existing linear mapping.

Fix this by removing the address to pfn translation and pass the
address directly.

Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
---
 arch/arm/mm/ioremap.c   | 4 +---
 arch/arm64/mm/ioremap.c | 4 +---
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
index 748698e91a4b..27e64f782cb3 100644
--- a/arch/arm/mm/ioremap.c
+++ b/arch/arm/mm/ioremap.c
@@ -515,7 +515,5 @@ void __init early_ioremap_init(void)
 bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
 				 unsigned long flags)
 {
-	unsigned long pfn = PHYS_PFN(offset);
-
-	return memblock_is_map_memory(pfn);
+	return memblock_is_map_memory(offset);
 }
diff --git a/arch/arm64/mm/ioremap.c b/arch/arm64/mm/ioremap.c
index 10e246f11271..48c38c986b95 100644
--- a/arch/arm64/mm/ioremap.c
+++ b/arch/arm64/mm/ioremap.c
@@ -51,7 +51,5 @@ void __init early_ioremap_init(void)
 bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
 				 unsigned long flags)
 {
-	unsigned long pfn = PHYS_PFN(offset);
-
-	return pfn_is_map_memory(pfn);
+	return pfn_is_map_memory(offset);
 }
-- 
2.49.0

Re: [PATCH] arm[64]/memremap: fix arch_memremap_can_ram_remap()

[Catalin Marinas]

Please cc the maintainers and the original contributor of the commit you
are fixing, otherwise the patch may not be noticed.

On Mon, Apr 14, 2025 at 08:32:19AM -0500, Ross Stutterheim wrote:
> commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
> presence of linear map") added the definition of
> arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
> can be used from the linear mapping. memblock_is_map_memory() was called
> with the pfn of the address given to arch_memremap_can_ram_remap();
> however, memblock_is_map_memory() expects to be given an address, not a
> pfn.
> 
> This results in calls to memremap() returning a newly mapped area when
> it should return an address in the existing linear mapping.
> 
> Fix this by removing the address to pfn translation and pass the
> address directly.
> 
> Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
> ---
>  arch/arm/mm/ioremap.c   | 4 +---
>  arch/arm64/mm/ioremap.c | 4 +---
>  2 files changed, 2 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
> index 748698e91a4b..27e64f782cb3 100644
> --- a/arch/arm/mm/ioremap.c
> +++ b/arch/arm/mm/ioremap.c
> @@ -515,7 +515,5 @@ void __init early_ioremap_init(void)
>  bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
>  				 unsigned long flags)
>  {
> -	unsigned long pfn = PHYS_PFN(offset);
> -
> -	return memblock_is_map_memory(pfn);
> +	return memblock_is_map_memory(offset);
>  }

This indeed needs fixing.

> diff --git a/arch/arm64/mm/ioremap.c b/arch/arm64/mm/ioremap.c
> index 10e246f11271..48c38c986b95 100644
> --- a/arch/arm64/mm/ioremap.c
> +++ b/arch/arm64/mm/ioremap.c
> @@ -51,7 +51,5 @@ void __init early_ioremap_init(void)
>  bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
>  				 unsigned long flags)
>  {
> -	unsigned long pfn = PHYS_PFN(offset);
> -
> -	return pfn_is_map_memory(pfn);
> +	return pfn_is_map_memory(offset);

This is already correct.

-- 
Catalin

Re: [PATCH] arm[64]/memremap: fix arch_memremap_can_ram_remap()

[Ross Stutterheim]

On 4/14/25 08:49, Catalin Marinas wrote:
> Please cc the maintainers and the original contributor of the commit you
> are fixing, otherwise the patch may not be noticed.
Thanks for adding them on your reply. I will add to V2.

>> diff --git a/arch/arm64/mm/ioremap.c b/arch/arm64/mm/ioremap.c
>> index 10e246f11271..48c38c986b95 100644
>> --- a/arch/arm64/mm/ioremap.c
>> +++ b/arch/arm64/mm/ioremap.c
>> @@ -51,7 +51,5 @@ void __init early_ioremap_init(void)
>>   bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
>>   				 unsigned long flags)
>>   {
>> -	unsigned long pfn = PHYS_PFN(offset);
>> -
>> -	return pfn_is_map_memory(pfn);
>> +	return pfn_is_map_memory(offset);
> 
> This is already correct.
Will remove on V2.

--
Ross

[PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Ross Stutterheim]

commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
presence of linear map") added the definition of
arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
can be used from the linear mapping. memblock_is_map_memory() was called
with the pfn of the address given to arch_memremap_can_ram_remap();
however, memblock_is_map_memory() expects to be given an address for arm,
not a pfn.

This results in calls to memremap() returning a newly mapped area when
it should return an address in the existing linear mapping.

Fix this by removing the address to pfn translation and pass the
address directly.

Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
---
 arch/arm/mm/ioremap.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
index 748698e91a4b..27e64f782cb3 100644
--- a/arch/arm/mm/ioremap.c
+++ b/arch/arm/mm/ioremap.c
@@ -515,7 +515,5 @@ void __init early_ioremap_init(void)
 bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
 				 unsigned long flags)
 {
-	unsigned long pfn = PHYS_PFN(offset);
-
-	return memblock_is_map_memory(pfn);
+	return memblock_is_map_memory(offset);
 }
-- 
2.49.0

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Catalin Marinas]

On Mon, Apr 14, 2025 at 09:21:40AM -0500, Ross Stutterheim wrote:
> commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
> presence of linear map") added the definition of
> arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
> can be used from the linear mapping. memblock_is_map_memory() was called
> with the pfn of the address given to arch_memremap_can_ram_remap();
> however, memblock_is_map_memory() expects to be given an address for arm,
> not a pfn.
> 
> This results in calls to memremap() returning a newly mapped area when
> it should return an address in the existing linear mapping.
> 
> Fix this by removing the address to pfn translation and pass the
> address directly.
> 
> Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>

I think you could also add:

Cc: <stable@vger.kernel.org>

> ---
>  arch/arm/mm/ioremap.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
> index 748698e91a4b..27e64f782cb3 100644
> --- a/arch/arm/mm/ioremap.c
> +++ b/arch/arm/mm/ioremap.c
> @@ -515,7 +515,5 @@ void __init early_ioremap_init(void)
>  bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
>  				 unsigned long flags)
>  {
> -	unsigned long pfn = PHYS_PFN(offset);
> -
> -	return memblock_is_map_memory(pfn);
> +	return memblock_is_map_memory(offset);
>  }

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>

Not sure how Russell picks patches up these days (I used to send them to
the patch system -
https://www.arm.linux.org.uk/developer/patches/info.php).

It might be simpler with git send-email (that's the alias I had):

  git send-email --add-header="KernelVersion: $(git describe --abbrev=0)" --no-thread --suppress-cc=all --to="patches@armlinux.org.uk"

-- 
Catalin

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Ross Stutterheim]

On 4/16/25 05:09, Catalin Marinas wrote:
>> Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
>> Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
> 
> I think you could also add:
> 
> Cc: <stable@vger.kernel.org>
Done. I also added some other Cc: entries on V3 along with Reviewed-by: 
lines . I'm new here, so I'm not sure I've used those 100% properly.

> Not sure how Russell picks patches up these days (I used to send them to
> the patch system -
> https://www.arm.linux.org.uk/developer/patches/info.php).
> 
> It might be simpler with git send-email (that's the alias I had):
> 
>    git send-email --add-header="KernelVersion: $(git describe --abbrev=0)" --no-thread --suppress-cc=all --to="patches@armlinux.org.uk"
> 
Thanks. I created an account there and submitted V3 through the web 
interface (to avoid my SMTP server appending more stuff on the end).

--
Ross

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Russell King (Oracle)]

On Wed, Apr 16, 2025 at 08:57:09AM -0500, Ross Stutterheim wrote:
> On 4/16/25 05:09, Catalin Marinas wrote:
> > > Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> > > Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
> > 
> > I think you could also add:
> > 
> > Cc: <stable@vger.kernel.org>
> Done. I also added some other Cc: entries on V3 along with Reviewed-by:
> lines . I'm new here, so I'm not sure I've used those 100% properly.
> 
> > Not sure how Russell picks patches up these days (I used to send them to
> > the patch system -
> > https://www.arm.linux.org.uk/developer/patches/info.php).
> > 
> > It might be simpler with git send-email (that's the alias I had):
> > 
> >    git send-email --add-header="KernelVersion: $(git describe --abbrev=0)" --no-thread --suppress-cc=all --to="patches@armlinux.org.uk"
> > 
> Thanks. I created an account there and submitted V3 through the web
> interface (to avoid my SMTP server appending more stuff on the end).

Thanks, seems all good, applied and pushed out. I'll send it to Linus
tomorrow as it's certainly a serious regression.

I'm surprised Mike hasn't responded...

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Catalin Marinas]

On Wed, Apr 16, 2025 at 04:48:06PM +0100, Russell King wrote:
> On Wed, Apr 16, 2025 at 08:57:09AM -0500, Ross Stutterheim wrote:
> > On 4/16/25 05:09, Catalin Marinas wrote:
> > > > Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> > > > Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
> > > 
> > > I think you could also add:
> > > 
> > > Cc: <stable@vger.kernel.org>
> > Done. I also added some other Cc: entries on V3 along with Reviewed-by:
> > lines . I'm new here, so I'm not sure I've used those 100% properly.
> > 
> > > Not sure how Russell picks patches up these days (I used to send them to
> > > the patch system -
> > > https://www.arm.linux.org.uk/developer/patches/info.php).
> > > 
> > > It might be simpler with git send-email (that's the alias I had):
> > > 
> > >    git send-email --add-header="KernelVersion: $(git describe --abbrev=0)" --no-thread --suppress-cc=all --to="patches@armlinux.org.uk"
> > > 
> > Thanks. I created an account there and submitted V3 through the web
> > interface (to avoid my SMTP server appending more stuff on the end).
> 
> Thanks, seems all good, applied and pushed out. I'll send it to Linus
> tomorrow as it's certainly a serious regression.
> 
> I'm surprised Mike hasn't responded...

Ah, I think he joined Microsoft. Adding his kernel.org address.

-- 
Catalin

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Mike Rapoport]

On Wed, Apr 16, 2025 at 04:48:06PM +0100, Russell King (Oracle) wrote:
> On Wed, Apr 16, 2025 at 08:57:09AM -0500, Ross Stutterheim wrote:
> > On 4/16/25 05:09, Catalin Marinas wrote:
> > > > Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> > > > Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
> > > 
> > > I think you could also add:
> > > 
> > > Cc: <stable@vger.kernel.org>
> > Done. I also added some other Cc: entries on V3 along with Reviewed-by:
> > lines . I'm new here, so I'm not sure I've used those 100% properly.
> > 
> > > Not sure how Russell picks patches up these days (I used to send them to
> > > the patch system -
> > > https://www.arm.linux.org.uk/developer/patches/info.php).
> > > 
> > > It might be simpler with git send-email (that's the alias I had):
> > > 
> > >    git send-email --add-header="KernelVersion: $(git describe --abbrev=0)" --no-thread --suppress-cc=all --to="patches@armlinux.org.uk"
> > > 
> > Thanks. I created an account there and submitted V3 through the web
> > interface (to avoid my SMTP server appending more stuff on the end).
> 
> Thanks, seems all good, applied and pushed out. I'll send it to Linus
> tomorrow as it's certainly a serious regression.
> 
> I'm surprised Mike hasn't responded...

Passover vacation :)
 
> -- 
> RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!

-- 
Sincerely yours,
Mike.

Re: [PATCH v2] arm/memremap: fix arch_memremap_can_ram_remap()

[Linus Walleij]

On Mon, Apr 14, 2025 at 4:32 PM Ross Stutterheim
<ross.stutterheim@garmin.com> wrote:

> commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
> presence of linear map") added the definition of
> arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
> can be used from the linear mapping. memblock_is_map_memory() was called
> with the pfn of the address given to arch_memremap_can_ram_remap();
> however, memblock_is_map_memory() expects to be given an address for arm,
> not a pfn.
>
> This results in calls to memremap() returning a newly mapped area when
> it should return an address in the existing linear mapping.
>
> Fix this by removing the address to pfn translation and pass the
> address directly.
>
> Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
> Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>

OMG
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>

Also add Cc: stable@vger.kernel.org as pointed out by Catalin.

Yours,
Linus Walleij

[PATCH v3] arm/memremap: fix arch_memremap_can_ram_remap()

[Ross Stutterheim]

commit 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure
presence of linear map") added the definition of
arch_memremap_can_ram_remap() for arm[64] specific filtering of what pages
can be used from the linear mapping. memblock_is_map_memory() was called
with the pfn of the address given to arch_memremap_can_ram_remap();
however, memblock_is_map_memory() expects to be given an address for arm,
not a pfn.

This results in calls to memremap() returning a newly mapped area when
it should return an address in the existing linear mapping.

Fix this by removing the address to pfn translation and pass the
address directly.

Fixes: 260364d112bc ("arm[64]/memremap: don't abuse pfn_valid() to ensure presence of linear map")
Signed-off-by: Ross Stutterheim <ross.stutterheim@garmin.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: stable@vger.kernel.org
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/mm/ioremap.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
index 748698e91a4b..27e64f782cb3 100644
--- a/arch/arm/mm/ioremap.c
+++ b/arch/arm/mm/ioremap.c
@@ -515,7 +515,5 @@ void __init early_ioremap_init(void)
 bool arch_memremap_can_ram_remap(resource_size_t offset, size_t size,
 				 unsigned long flags)
 {
-	unsigned long pfn = PHYS_PFN(offset);
-
-	return memblock_is_map_memory(pfn);
+	return memblock_is_map_memory(offset);
 }
-- 
2.49.0