Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 952 bytes --]

Hi!

> Resent without non-upstream patches.
> 
> This backport patchset fixed the spectre issue, it's original branch:
> https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
> A few dependency or fixingpatches are also picked up, if they are necessary
>  and no functional changes.
> 
> No bug found from kernelci.org and lkft testing. It also could be gotten from:
> 
> git://git.linaro.org/kernel/linux-linaro-stable.git v4.9-spectre-upstream-only
> 
> Comments are appreciated!

Not entirely related to this patched, but... I have few older ARM
boards here, and Nokia N9000 I really care about.

AFAICT Meltdown is arm64 only?

Spectre affects the older boards, too, right? Was there any work done
on that? cpuinfo says "ARMv7" for N900.

Thanks,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

[Pali Rohár]

On Thursday 08 March 2018 13:27:15 Pavel Machek wrote:
> Hi!
> 
> > Resent without non-upstream patches.
> > 
> > This backport patchset fixed the spectre issue, it's original branch:
> > https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
> > A few dependency or fixingpatches are also picked up, if they are necessary
> >  and no functional changes.
> > 
> > No bug found from kernelci.org and lkft testing. It also could be gotten from:
> > 
> > git://git.linaro.org/kernel/linux-linaro-stable.git v4.9-spectre-upstream-only
> > 
> > Comments are appreciated!
> 
> Not entirely related to this patched, but... I have few older ARM
> boards here, and Nokia N9000 I really care about.
> 
> AFAICT Meltdown is arm64 only?

IIRC ARMv7 is not affected by meltdown.

> Spectre affects the older boards, too, right? Was there any work done
> on that? cpuinfo says "ARMv7" for N900.

I remember that I saw some spectre patches for ARMv7 on LKML.

In general for ARMv7 it is problematic as mitigation needs to change IBE
bit which is not possible on OMAP HS devices. But for Nokia N900 there
is special code which do it via smc instruction (function
rx51_secure_update_aux_cr(), see also nokia_n900_legacy_init()).

-- 
Pali Rohár
pali.rohar@gmail.com

Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

[Ivaylo Dimitrov]

Hi,

On  8.03.2018 15:21, Pali Rohár wrote:
> On Thursday 08 March 2018 13:27:15 Pavel Machek wrote:
>> Hi!
>>
>>> Resent without non-upstream patches.
>>>
>>> This backport patchset fixed the spectre issue, it's original branch:
>>> https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
>>> A few dependency or fixingpatches are also picked up, if they are necessary
>>>   and no functional changes.
>>>
>>> No bug found from kernelci.org and lkft testing. It also could be gotten from:
>>>
>>> git://git.linaro.org/kernel/linux-linaro-stable.git v4.9-spectre-upstream-only
>>>
>>> Comments are appreciated!
>>
>> Not entirely related to this patched, but... I have few older ARM
>> boards here, and Nokia N9000 I really care about.
>>
>> AFAICT Meltdown is arm64 only?
> 
> IIRC ARMv7 is not affected by meltdown.
> 
>> Spectre affects the older boards, too, right? Was there any work done
>> on that? cpuinfo says "ARMv7" for N900.
> 
> I remember that I saw some spectre patches for ARMv7 on LKML.
> 
> In general for ARMv7 it is problematic as mitigation needs to change IBE
> bit which is not possible on OMAP HS devices. But for Nokia N900 there
> is special code which do it via smc instruction (function
> rx51_secure_update_aux_cr(), see also nokia_n900_legacy_init()).
> 

I *think* N900 should be safe from spectre attack as a side effect of 
thumb2 errata (430973) workaround (BTB is flushed on every context 
switch), but don't quote me on that one, lets more knowledgeable than me 
to confirm first.

Ivo

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

[Marc Zyngier]

On 08/03/18 13:21, Pali Rohár wrote:
> On Thursday 08 March 2018 13:27:15 Pavel Machek wrote:
>> Hi!
>>
>>> Resent without non-upstream patches.
>>>
>>> This backport patchset fixed the spectre issue, it's original branch:
>>> https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
>>> A few dependency or fixingpatches are also picked up, if they are necessary
>>>  and no functional changes.
>>>
>>> No bug found from kernelci.org and lkft testing. It also could be gotten from:
>>>
>>> git://git.linaro.org/kernel/linux-linaro-stable.git v4.9-spectre-upstream-only
>>>
>>> Comments are appreciated!
>>
>> Not entirely related to this patched, but... I have few older ARM
>> boards here, and Nokia N9000 I really care about.
>>
>> AFAICT Meltdown is arm64 only?
> 
> IIRC ARMv7 is not affected by meltdown.
> 
>> Spectre affects the older boards, too, right? Was there any work done
>> on that? cpuinfo says "ARMv7" for N900.
> 
> I remember that I saw some spectre patches for ARMv7 on LKML.
> 
> In general for ARMv7 it is problematic as mitigation needs to change IBE

Not quite. IBE is part of ACTLR, which is an implementation defined
register, so generalizing this statement to "ARMv7" is plain wrong. Case
in point: A9, A12/A17 do not need any bit to be set. A8 needs IBE to
allow BPIALL to work, A15 needs it for ICIALLUI to perform a BP
invalidation.

	M.
-- 
Jazz is not dead. It just smells funny...

Re: [PATCH 00/45] arm Spectre fix backport review for LTS 4.9

[Pali Rohár]

[-- Attachment #1: Type: text/plain, Size: 1706 bytes --]

On Thursday 08 March 2018 13:35:49 Marc Zyngier wrote:
> On 08/03/18 13:21, Pali Rohár wrote:
> > On Thursday 08 March 2018 13:27:15 Pavel Machek wrote:
> >> Hi!
> >>
> >>> Resent without non-upstream patches.
> >>>
> >>> This backport patchset fixed the spectre issue, it's original branch:
> >>> https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
> >>> A few dependency or fixingpatches are also picked up, if they are necessary
> >>>  and no functional changes.
> >>>
> >>> No bug found from kernelci.org and lkft testing. It also could be gotten from:
> >>>
> >>> git://git.linaro.org/kernel/linux-linaro-stable.git v4.9-spectre-upstream-only
> >>>
> >>> Comments are appreciated!
> >>
> >> Not entirely related to this patched, but... I have few older ARM
> >> boards here, and Nokia N9000 I really care about.
> >>
> >> AFAICT Meltdown is arm64 only?
> > 
> > IIRC ARMv7 is not affected by meltdown.
> > 
> >> Spectre affects the older boards, too, right? Was there any work done
> >> on that? cpuinfo says "ARMv7" for N900.
> > 
> > I remember that I saw some spectre patches for ARMv7 on LKML.

This is that patch series:
https://www.spinics.net/lists/arm-kernel/msg632065.html

> > In general for ARMv7 it is problematic as mitigation needs to change IBE
> 
> Not quite. IBE is part of ACTLR, which is an implementation defined
> register, so generalizing this statement to "ARMv7" is plain wrong. Case
> in point: A9, A12/A17 do not need any bit to be set. A8 needs IBE to
> allow BPIALL to work, A15 needs it for ICIALLUI to perform a BP
> invalidation.

Ok, Nokia N900 has Cortex A8.

-- 
Pali Rohár
pali.rohar@gmail.com

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]