lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YMfRxX/M4rJ5gM/R@zn.tnic>
Date:   Tue, 15 Jun 2021 00:01:41 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Smita Koralahalli <Smita.KoralahalliChannabasappa@....com>
Cc:     x86@...nel.org, linux-kernel@...r.kernel.org,
        linux-edac@...r.kernel.org, Tony Luck <tony.luck@...el.com>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        James Morse <james.morse@....com>, yazen.ghannam@....com,
        Robert Richter <rric@...nel.org>
Subject: Re: [PATCH] EDAC/mce_amd: Reduce unnecessary spew in dmesg if SMCA
 feature bit is not exposed

On Mon, Jun 14, 2021 at 04:21:29PM -0500, Smita Koralahalli wrote:
> The SMCA feature bit is not exposed on the guest.
> 
> This causes a lot of noise in dmesg as the warning is printed for each
> logical CPU.
> 
> $ dmesg |grep -i family
> 
> [    0.031000] smpboot: CPU0: AMD EPYC-Milan Processor (family: 0x19, model: 0x1, stepping: 0x1)
> [    4.653422] Huh? What family is it: 0x19?!
> [    4.720732] Huh? What family is it: 0x19?!
> [    6.171028] Huh? What family is it: 0x19?!
> [    6.766552] Huh? What family is it: 0x19?!
> [    6.811119] Huh? What family is it: 0x19?!
> [    6.839855] Huh? What family is it: 0x19?!
> 
> Give these messages debug severity and output once as it is mostly useful
> for module developers and just noise for users.

I'm getting a patch like that on a pretty regular basis and each time I
tell people simply not to load the module in guests. Hypervisors do not
support SMCA...

But apparently those wrong error messages bug people on a regular basis
and I'm tired of typing this each time so I'll take a different version
of this patch: check X86_FEATURE_HYPERVISOR on entry in mce_amd_init()
and return -ENODEV if set.

So that this is done once and for all.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ