[PATCH] arm64: efi: Force the use of SetVirtualAddressMap() on Altra Max machines

[PATCH] arm64: efi: Force the use of SetVirtualAddressMap() on Altra Max machines

[Darren Hart]

Commit 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap()
on Altra machines") identifies the Altra family via the family field in
the type#1 SMBIOS record. Altra Max machines are similarly affected but
not detected with the strict strcmp test.

Rather than risk greedy matching with strncmp, add a second test for
Altra Max. Do not refactor to handle multiple tests as these should be
the only two needed.

Fixes: 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap() on Altra machines")
Cc: <stable@vger.kernel.org> # 6.1.x
Cc: <linux-efi@vger.kernel.org>
Cc: Alexandru Elisei <alexandru.elisei@gmail.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Darren Hart <darren@os.amperecomputing.com>
---
 drivers/firmware/efi/libstub/arm64.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/firmware/efi/libstub/arm64.c b/drivers/firmware/efi/libstub/arm64.c
index ff2d18c42ee7..97f4423059c7 100644
--- a/drivers/firmware/efi/libstub/arm64.c
+++ b/drivers/firmware/efi/libstub/arm64.c
@@ -19,10 +19,10 @@ static bool system_needs_vamap(void)
 	const u8 *type1_family = efi_get_smbios_string(1, family);
 
 	/*
-	 * Ampere Altra machines crash in SetTime() if SetVirtualAddressMap()
-	 * has not been called prior.
+	 * Ampere Altra and Altra Max machines crash in SetTime() if
+	 * SetVirtualAddressMap() has not been called prior.
 	 */
-	if (!type1_family || strcmp(type1_family, "Altra"))
+	if (!type1_family || (strcmp(type1_family, "Altra") && strcmp(type1_family, "Altra Max")))
 		return false;
 
 	efi_warn("Working around broken SetVirtualAddressMap()\n");
-- 
2.34.3

Re: [PATCH] arm64: efi: Force the use of SetVirtualAddressMap() on Altra Max machines

[Ard Biesheuvel]

Hello Darren,

On Thu, 9 Feb 2023 at 00:14, Darren Hart <darren@os.amperecomputing.com> wrote:
>
> Commit 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap()
> on Altra machines") identifies the Altra family via the family field in
> the type#1 SMBIOS record. Altra Max machines are similarly affected but
> not detected with the strict strcmp test.
>
> Rather than risk greedy matching with strncmp, add a second test for
> Altra Max. Do not refactor to handle multiple tests as these should be
> the only two needed.
>

Famous last words ...

Unfortunately, I just had a report the other day that 'eMAG' and
'Server' (!) are also being used.

https://lore.kernel.org/all/20230131040355.3116-1-justin.he@arm.com/


> Fixes: 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap() on Altra machines")
> Cc: <stable@vger.kernel.org> # 6.1.x
> Cc: <linux-efi@vger.kernel.org>
> Cc: Alexandru Elisei <alexandru.elisei@gmail.com>
> Cc: Ard Biesheuvel <ardb@kernel.org>
> Signed-off-by: Darren Hart <darren@os.amperecomputing.com>
> ---
>  drivers/firmware/efi/libstub/arm64.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/firmware/efi/libstub/arm64.c b/drivers/firmware/efi/libstub/arm64.c
> index ff2d18c42ee7..97f4423059c7 100644
> --- a/drivers/firmware/efi/libstub/arm64.c
> +++ b/drivers/firmware/efi/libstub/arm64.c
> @@ -19,10 +19,10 @@ static bool system_needs_vamap(void)
>         const u8 *type1_family = efi_get_smbios_string(1, family);
>
>         /*
> -        * Ampere Altra machines crash in SetTime() if SetVirtualAddressMap()
> -        * has not been called prior.
> +        * Ampere Altra and Altra Max machines crash in SetTime() if
> +        * SetVirtualAddressMap() has not been called prior.
>          */
> -       if (!type1_family || strcmp(type1_family, "Altra"))
> +       if (!type1_family || (strcmp(type1_family, "Altra") && strcmp(type1_family, "Altra Max")))
>                 return false;
>
>         efi_warn("Working around broken SetVirtualAddressMap()\n");
> --
> 2.34.3
>

Re: [PATCH] arm64: efi: Force the use of SetVirtualAddressMap() on Altra Max machines

[Darren Hart]

On Thu, Feb 09, 2023 at 12:16:46AM +0100, Ard Biesheuvel wrote:
> Hello Darren,
> 
> On Thu, 9 Feb 2023 at 00:14, Darren Hart <darren@os.amperecomputing.com> wrote:
> >
> > Commit 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap()
> > on Altra machines") identifies the Altra family via the family field in
> > the type#1 SMBIOS record. Altra Max machines are similarly affected but
> > not detected with the strict strcmp test.
> >
> > Rather than risk greedy matching with strncmp, add a second test for
> > Altra Max. Do not refactor to handle multiple tests as these should be
> > the only two needed.
> >
> 
> Famous last words ...

Indeed, I nearly included that myself...

> 
> Unfortunately, I just had a report the other day that 'eMAG' and
> 'Server' (!) are also being used.
> 
> https://lore.kernel.org/all/20230131040355.3116-1-justin.he@arm.com/
> 

OK, so in order to workaround this in the kernel, we need a better way to match.
Unfortunately, this is specific to the oem platform, and the oem controls those
strings.

Thanks for the pointer, will go mull this over and see if I can come up with
something better.

In the meantime, would you consider matching on:
Altra
Altra Max
eMAG

to capture the bulk of the systems until we have a better solution?

Thanks,

Darren

> 
> > Fixes: 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap() on Altra machines")
> > Cc: <stable@vger.kernel.org> # 6.1.x
> > Cc: <linux-efi@vger.kernel.org>
> > Cc: Alexandru Elisei <alexandru.elisei@gmail.com>
> > Cc: Ard Biesheuvel <ardb@kernel.org>
> > Signed-off-by: Darren Hart <darren@os.amperecomputing.com>
> > ---
> >  drivers/firmware/efi/libstub/arm64.c | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/firmware/efi/libstub/arm64.c b/drivers/firmware/efi/libstub/arm64.c
> > index ff2d18c42ee7..97f4423059c7 100644
> > --- a/drivers/firmware/efi/libstub/arm64.c
> > +++ b/drivers/firmware/efi/libstub/arm64.c
> > @@ -19,10 +19,10 @@ static bool system_needs_vamap(void)
> >         const u8 *type1_family = efi_get_smbios_string(1, family);
> >
> >         /*
> > -        * Ampere Altra machines crash in SetTime() if SetVirtualAddressMap()
> > -        * has not been called prior.
> > +        * Ampere Altra and Altra Max machines crash in SetTime() if
> > +        * SetVirtualAddressMap() has not been called prior.
> >          */
> > -       if (!type1_family || strcmp(type1_family, "Altra"))
> > +       if (!type1_family || (strcmp(type1_family, "Altra") && strcmp(type1_family, "Altra Max")))
> >                 return false;
> >
> >         efi_warn("Working around broken SetVirtualAddressMap()\n");
> > --
> > 2.34.3
> >

-- 
Darren Hart
Ampere Computing / OS and Kernel

Re: [PATCH] arm64: efi: Force the use of SetVirtualAddressMap() on Altra Max machines

[Ard Biesheuvel]

On Thu, 9 Feb 2023 at 00:43, Darren Hart <darren@os.amperecomputing.com> wrote:
>
> On Thu, Feb 09, 2023 at 12:16:46AM +0100, Ard Biesheuvel wrote:
> > Hello Darren,
> >
> > On Thu, 9 Feb 2023 at 00:14, Darren Hart <darren@os.amperecomputing.com> wrote:
> > >
> > > Commit 550b33cfd445 ("arm64: efi: Force the use of SetVirtualAddressMap()
> > > on Altra machines") identifies the Altra family via the family field in
> > > the type#1 SMBIOS record. Altra Max machines are similarly affected but
> > > not detected with the strict strcmp test.
> > >
> > > Rather than risk greedy matching with strncmp, add a second test for
> > > Altra Max. Do not refactor to handle multiple tests as these should be
> > > the only two needed.
> > >
> >
> > Famous last words ...
>
> Indeed, I nearly included that myself...
>
> >
> > Unfortunately, I just had a report the other day that 'eMAG' and
> > 'Server' (!) are also being used.
> >
> > https://lore.kernel.org/all/20230131040355.3116-1-justin.he@arm.com/
> >
>
> OK, so in order to workaround this in the kernel, we need a better way to match.
> Unfortunately, this is specific to the oem platform, and the oem controls those
> strings.
>
> Thanks for the pointer, will go mull this over and see if I can come up with
> something better.
>
> In the meantime, would you consider matching on:
> Altra
> Altra Max
> eMAG
>
> to capture the bulk of the systems until we have a better solution?
>

Yeah that's fine.