Re: [gentoo-dev] [PATCH 0/4] CPU_FLAGS_X86: Introduce 'rdrand' flag

[Francisco Blas Izquierdo Riera (klondike)]

El 13/7/20 a las 22:07, Francisco Blas Izquierdo Riera (klondike) escribió:
> El 13/7/20 a las 21:27, Michał Górny escribió:
>> On Mon, 2020-07-13 at 19:33 +0200, Francisco Blas Izquierdo Riera
>> (klondike) wrote:
>>> El 13/7/20 a las 19:23, Michał Górny escribió:
>>>> On Mon, 2020-07-13 at 19:07 +0200, Francisco Blas Izquierdo Riera
>>>> (klondike) wrote:
>>>>> Hi!
>>>>>
>>>>> We have currently two packages that have USE cpu-flags-x86-rdrand as there is no USE_EXPAND version available. This pattern is likely to confuse users as they may not be aware of the difference between dashes and underscores.
>>>>>
>>>>> Affected packages:
>>>>> dev-haskell/cryptonite
>>>>> dev-libs/json-c
>>>>>
>>>> I'm sorry but I haven't been following the big rdrand-AMD drama closely.
>>>>  Does the kernel mitigate broken RDRAND after resume these days?
>>>>
>>> AGESA patches do exist to address this issue: https://www.reddit.com/r/Amd/comments/cmza34/agesa_1003_abb_fixes_rdrandrdseed/ I suspect AMD microcode does too.
>>>
>> Lemme rephrase: is this something we should be warning our users before
>> they enable the flag?  I think we should aim to avoid the situation that
>> cpuid2cpuflags enables this behind user's backs and their programs
>> suddenly break as a result.
> You then mean the other rdrand bug.
>
> Yes, current stable linux kernel (5.4.48) in amd64 does include the patch, I'm uncertain about prior kernel branches though but I don't expect many users on that confluence.
>
> The patch: https://lore.kernel.org/linux-doc/776cb5c2d33e7fd0d2893904724c0e52b394f24a.1565817448.git.thomas.lendacky@×××.com/
> The bug: https://bugzilla.kernel.org/show_bug.cgi?id=85911
>
>
I have checked also stable versions for other kernel branches and all seem to have the patch icluded.