[PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

This series is a followup to the single patch 'modversions: treat symbol
CRCs as 32 bit quantities on 64 bit archs', of which two versions have
been sent out so far [0][1]

As pointed out by Michael, GNU ld behaves a bit differently between arm64
and PowerPC64, and where the former gets rid of all runtime relocations
related to CRCs, the latter is not as easily convinced.

Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
routines for 32-bit PowerPC, for which the original fix was effectively
reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
causes perf issues")

Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
symbol entry to the PPC64 runtime relocation routines, so it is prepared to
deal with CRCs being emitted as 32-bit quantities.

Patch #3 is the original patch from the v1 and v2 submissions.

Changes since v2:
- added #1 and #2
- updated #3 to deal with CRC entries being emitted from assembler
- added Rusty's ack (#3)

[0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
[1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2

Ard Biesheuvel (3):
  powerpc/reloc32: fix corrupted modversion CRCs
  powerpc/reloc64: add support for 32-bit CRC pseudo-symbols
  modversions: treat symbol CRCs as 32 bit quantities on 64 bit archs

 arch/powerpc/include/asm/module.h |  4 --
 arch/powerpc/kernel/module_64.c   |  8 ----
 arch/powerpc/kernel/reloc_32.S    | 36 +++++++++++++--
 arch/powerpc/kernel/reloc_64.S    | 22 +++++++--
 arch/powerpc/relocs_check.sh      |  3 +-
 include/asm-generic/export.h      |  7 +--
 include/linux/export.h            |  8 ++++
 include/linux/module.h            | 14 +++---
 kernel/module.c                   | 47 +++++++-------------
 9 files changed, 85 insertions(+), 64 deletions(-)

-- 
2.7.4

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Michael Ellerman]

Ard Biesheuvel <ard.biesheuvel@linaro.org> writes:

> On 15 November 2016 at 09:13, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>> On 10 November 2016 at 05:22, Michael Ellerman <mpe@ellerman.id.au> wrote:
>>> Ard Biesheuvel <ard.biesheuvel@linaro.org> writes:
>>>
>>>> On 27 October 2016 at 17:27, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>>>>> This series is a followup to the single patch 'modversions: treat symbol
>>>>> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
>>>>> been sent out so far [0][1]
>>>>>
>>>>> As pointed out by Michael, GNU ld behaves a bit differently between arm64
>>>>> and PowerPC64, and where the former gets rid of all runtime relocations
>>>>> related to CRCs, the latter is not as easily convinced.
>>>>>
>>>>> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
>>>>> routines for 32-bit PowerPC, for which the original fix was effectively
>>>>> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
>>>>> causes perf issues")
>>>>>
>>>>> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
>>>>> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
>>>>> deal with CRCs being emitted as 32-bit quantities.
>>>>>
>>>>> Patch #3 is the original patch from the v1 and v2 submissions.
>>>>>
>>>>> Changes since v2:
>>>>> - added #1 and #2
>>>>> - updated #3 to deal with CRC entries being emitted from assembler
>>>>> - added Rusty's ack (#3)
>>>>>
>>>>> Branch can be found here:
>>>>> https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc
>>>>>
>>>>> [0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
>>>>> [1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2
>>>>
>>>> Ping?
>>>
>>> Sorry, you didn't cc linuxppc-dev, so it's not in my patchwork list
>>> which tends to mean I miss it.
>>
>> Ah, my mistake. Apologies.
>>
>>> Will try and test and get back to you.
>
> Ping?

Sorry :/

I tried testing it last week or so but it was interacting badly with the
other modversion CRC problems we were having (fixed recently in my fixes
branch).

I'll rebase on top of those fixes and try again.

cheers

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

On 15 November 2016 at 09:13, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> On 10 November 2016 at 05:22, Michael Ellerman <mpe@ellerman.id.au> wrote:
>> Ard Biesheuvel <ard.biesheuvel@linaro.org> writes:
>>
>>> On 27 October 2016 at 17:27, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>>>> This series is a followup to the single patch 'modversions: treat symbol
>>>> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
>>>> been sent out so far [0][1]
>>>>
>>>> As pointed out by Michael, GNU ld behaves a bit differently between arm64
>>>> and PowerPC64, and where the former gets rid of all runtime relocations
>>>> related to CRCs, the latter is not as easily convinced.
>>>>
>>>> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
>>>> routines for 32-bit PowerPC, for which the original fix was effectively
>>>> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
>>>> causes perf issues")
>>>>
>>>> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
>>>> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
>>>> deal with CRCs being emitted as 32-bit quantities.
>>>>
>>>> Patch #3 is the original patch from the v1 and v2 submissions.
>>>>
>>>> Changes since v2:
>>>> - added #1 and #2
>>>> - updated #3 to deal with CRC entries being emitted from assembler
>>>> - added Rusty's ack (#3)
>>>>
>>>> Branch can be found here:
>>>> https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc
>>>>
>>>> [0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
>>>> [1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2
>>>
>>> Ping?
>>
>> Sorry, you didn't cc linuxppc-dev, so it's not in my patchwork list
>> which tends to mean I miss it.
>>
>
> Ah, my mistake. Apologies.
>
>> Will try and test and get back to you.
>>
>

Ping?

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

On 16 November 2016 at 19:23, Uwe Kleine-König
<u.kleine-koenig@pengutronix.de> wrote:
> On Thu, Oct 27, 2016 at 05:27:08PM +0100, Ard Biesheuvel wrote:
>> This series is a followup to the single patch 'modversions: treat symbol
>> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
>> been sent out so far [0][1]
>>
>> As pointed out by Michael, GNU ld behaves a bit differently between arm64
>> and PowerPC64, and where the former gets rid of all runtime relocations
>> related to CRCs, the latter is not as easily convinced.
>>
>> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
>> routines for 32-bit PowerPC, for which the original fix was effectively
>> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
>> causes perf issues")
>>
>> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
>> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
>> deal with CRCs being emitted as 32-bit quantities.
>>
>> Patch #3 is the original patch from the v1 and v2 submissions.
>
> Is this related to me seeing
>
> [    2.111424] mvneta: module verification failed: signature and/or required key missing - tainting kernel
> [    2.126061] scsi_mod: no symbol version for _clear_bit
> [    2.131257] scsi_mod: Unknown symbol _clear_bit (err -22)
> [    2.138093] mvneta: no symbol version for _clear_bit
> [    2.143117] mvneta: Unknown symbol _clear_bit (err -22)
> [    2.144135] mvmdio: no symbol version for __gnu_mcount_nc
> [    2.144138] mvmdio: Unknown symbol __gnu_mcount_nc (err -22)
> ...
>
> ? If so, this would be great to mention it in the commit log to make
> people searching for this issue actually find this patch set.
>

No, I don't think it is. My guess would be that it is caused by

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=4dd1837d75

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Uwe Kleine-König]

On Thu, Oct 27, 2016 at 05:27:08PM +0100, Ard Biesheuvel wrote:
> This series is a followup to the single patch 'modversions: treat symbol
> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
> been sent out so far [0][1]
> 
> As pointed out by Michael, GNU ld behaves a bit differently between arm64
> and PowerPC64, and where the former gets rid of all runtime relocations
> related to CRCs, the latter is not as easily convinced.
> 
> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
> routines for 32-bit PowerPC, for which the original fix was effectively
> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
> causes perf issues")
> 
> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
> deal with CRCs being emitted as 32-bit quantities.
> 
> Patch #3 is the original patch from the v1 and v2 submissions.

Is this related to me seeing 

[    2.111424] mvneta: module verification failed: signature and/or required key missing - tainting kernel
[    2.126061] scsi_mod: no symbol version for _clear_bit
[    2.131257] scsi_mod: Unknown symbol _clear_bit (err -22)
[    2.138093] mvneta: no symbol version for _clear_bit
[    2.143117] mvneta: Unknown symbol _clear_bit (err -22)
[    2.144135] mvmdio: no symbol version for __gnu_mcount_nc
[    2.144138] mvmdio: Unknown symbol __gnu_mcount_nc (err -22)
...

? If so, this would be great to mention it in the commit log to make
people searching for this issue actually find this patch set.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

On 10 November 2016 at 05:22, Michael Ellerman <mpe@ellerman.id.au> wrote:
> Ard Biesheuvel <ard.biesheuvel@linaro.org> writes:
>
>> On 27 October 2016 at 17:27, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>>> This series is a followup to the single patch 'modversions: treat symbol
>>> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
>>> been sent out so far [0][1]
>>>
>>> As pointed out by Michael, GNU ld behaves a bit differently between arm64
>>> and PowerPC64, and where the former gets rid of all runtime relocations
>>> related to CRCs, the latter is not as easily convinced.
>>>
>>> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
>>> routines for 32-bit PowerPC, for which the original fix was effectively
>>> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
>>> causes perf issues")
>>>
>>> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
>>> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
>>> deal with CRCs being emitted as 32-bit quantities.
>>>
>>> Patch #3 is the original patch from the v1 and v2 submissions.
>>>
>>> Changes since v2:
>>> - added #1 and #2
>>> - updated #3 to deal with CRC entries being emitted from assembler
>>> - added Rusty's ack (#3)
>>>
>>> Branch can be found here:
>>> https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc
>>>
>>> [0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
>>> [1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2
>>
>> Ping?
>
> Sorry, you didn't cc linuxppc-dev, so it's not in my patchwork list
> which tends to mean I miss it.
>

Ah, my mistake. Apologies.

> Will try and test and get back to you.
>

Thanks!

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Michael Ellerman]

Ard Biesheuvel <ard.biesheuvel@linaro.org> writes:

> On 27 October 2016 at 17:27, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
>> This series is a followup to the single patch 'modversions: treat symbol
>> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
>> been sent out so far [0][1]
>>
>> As pointed out by Michael, GNU ld behaves a bit differently between arm64
>> and PowerPC64, and where the former gets rid of all runtime relocations
>> related to CRCs, the latter is not as easily convinced.
>>
>> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
>> routines for 32-bit PowerPC, for which the original fix was effectively
>> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
>> causes perf issues")
>>
>> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
>> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
>> deal with CRCs being emitted as 32-bit quantities.
>>
>> Patch #3 is the original patch from the v1 and v2 submissions.
>>
>> Changes since v2:
>> - added #1 and #2
>> - updated #3 to deal with CRC entries being emitted from assembler
>> - added Rusty's ack (#3)
>>
>> Branch can be found here:
>> https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc
>>
>> [0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
>> [1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2
>
> Ping?

Sorry, you didn't cc linuxppc-dev, so it's not in my patchwork list
which tends to mean I miss it.

Will try and test and get back to you.

cheers

Re: [PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

On 27 October 2016 at 17:27, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> This series is a followup to the single patch 'modversions: treat symbol
> CRCs as 32 bit quantities on 64 bit archs', of which two versions have
> been sent out so far [0][1]
>
> As pointed out by Michael, GNU ld behaves a bit differently between arm64
> and PowerPC64, and where the former gets rid of all runtime relocations
> related to CRCs, the latter is not as easily convinced.
>
> Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
> routines for 32-bit PowerPC, for which the original fix was effectively
> reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
> causes perf issues")
>
> Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
> symbol entry to the PPC64 runtime relocation routines, so it is prepared to
> deal with CRCs being emitted as 32-bit quantities.
>
> Patch #3 is the original patch from the v1 and v2 submissions.
>
> Changes since v2:
> - added #1 and #2
> - updated #3 to deal with CRC entries being emitted from assembler
> - added Rusty's ack (#3)
>
> Branch can be found here:
> https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc
>
> [0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
> [1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2
>

Ping?

> Ard Biesheuvel (3):
>   powerpc/reloc32: fix corrupted modversion CRCs
>   powerpc/reloc64: add support for 32-bit CRC pseudo-symbols
>   modversions: treat symbol CRCs as 32 bit quantities on 64 bit archs
>
>  arch/powerpc/include/asm/module.h |  4 --
>  arch/powerpc/kernel/module_64.c   |  8 ----
>  arch/powerpc/kernel/reloc_32.S    | 36 +++++++++++++--
>  arch/powerpc/kernel/reloc_64.S    | 22 +++++++--
>  arch/powerpc/relocs_check.sh      |  3 +-
>  include/asm-generic/export.h      |  7 +--
>  include/linux/export.h            |  8 ++++
>  include/linux/module.h            | 14 +++---
>  kernel/module.c                   | 47 +++++++-------------
>  9 files changed, 85 insertions(+), 64 deletions(-)
>
> --
> 2.7.4
>

[PATCH v3 0/3] modversions: Fix CRC mangling under CONFIG_RELOCATABLE=y

[Ard Biesheuvel]

This series is a followup to the single patch 'modversions: treat symbol
CRCs as 32 bit quantities on 64 bit archs', of which two versions have
been sent out so far [0][1]

As pointed out by Michael, GNU ld behaves a bit differently between arm64
and PowerPC64, and where the former gets rid of all runtime relocations
related to CRCs, the latter is not as easily convinced.

Patch #1 fixes the issue where CRCs are corrupted by the runtime relocation
routines for 32-bit PowerPC, for which the original fix was effectively
reverted by commit 0e0ed6406e61 ("powerpc/modules: Module CRC relocation fix
causes perf issues")

Patch #2 adds handling of R_PPC64_ADDR32 relocations against the NULL .dynsym
symbol entry to the PPC64 runtime relocation routines, so it is prepared to
deal with CRCs being emitted as 32-bit quantities.

Patch #3 is the original patch from the v1 and v2 submissions.

Changes since v2:
- added #1 and #2
- updated #3 to deal with CRC entries being emitted from assembler
- added Rusty's ack (#3)

Branch can be found here:
https://git.kernel.org/cgit/linux/kernel/git/ardb/linux.git/log/?h=kcrctab-reloc

[0] http://marc.info/?l=linux-kernel&m=147652300207369&w=2
[1] http://marc.info/?l=linux-kernel&m=147695629614409&w=2

Ard Biesheuvel (3):
  powerpc/reloc32: fix corrupted modversion CRCs
  powerpc/reloc64: add support for 32-bit CRC pseudo-symbols
  modversions: treat symbol CRCs as 32 bit quantities on 64 bit archs

 arch/powerpc/include/asm/module.h |  4 --
 arch/powerpc/kernel/module_64.c   |  8 ----
 arch/powerpc/kernel/reloc_32.S    | 36 +++++++++++++--
 arch/powerpc/kernel/reloc_64.S    | 22 +++++++--
 arch/powerpc/relocs_check.sh      |  3 +-
 include/asm-generic/export.h      |  7 +--
 include/linux/export.h            |  8 ++++
 include/linux/module.h            | 14 +++---
 kernel/module.c                   | 47 +++++++-------------
 9 files changed, 85 insertions(+), 64 deletions(-)

-- 
2.7.4