linux-kernel - Re: microcode should perform update after suspend to disk

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]

Date:	Wed, 04 Jan 2012 11:37:43 +0530
From:	"Srivatsa S. Bhat" <srivatsa.bhat@...ux.vnet.ibm.com>
To:	Sven Köhler <sven.koehler@...il.com>
CC:	linux-kernel@...r.kernel.org, Borislav Petkov <bp@...64.org>,
	Linux PM mailing list <linux-pm@...r.kernel.org>,
	"Rafael J. Wysocki" <rjw@...k.pl>
Subject: Re: microcode should perform update after suspend to disk

On 01/01/2012 10:07 PM, Sven Köhler wrote:

> Hi,
> 
> at startup I load the microcode module and it correctly loads the
> microcode from /lib/firmware/intel-ucode/. In dmesg, I can see that the
> microcode has been updated from 0xc to 0xd.
> After suspend to ram, dmesg shows new messages from the microcode
> module, telling me that the microcode is still at revision 0xd. Hence,
> no update is performed.
> However, after suspend to disk, dmesg shows no new messages from the
> microcode module. After rmmod microcode; modprobe microcode dmesg tells
> me that the microcode had to be updated from 0xc to 0xd.
> 
> Somethings wrong here. My interpretation of the messages in dmesg is,
> that microcode is trying to update the microcode after suspend to ram
> but not after suspend to disk. It should try in both cases. Note that I
> don't unload the microcode module.
> 
> 

Which kernel version are you using?

As far as this part of the code is concerned, the path taken for both
suspend-to-ram and suspend-to-disk are more or less the same.
On the resume path, both call enable_nonboot_cpus() which brings the nonboot
cpus online, and due to the cpuhotplug notifications for these, we have a
flow something like:

[The following functions are all in arch/x86/kernel/microcode_core.c]

  mc_cpu_callback() //with action=CPU_ONLINE_FROZEN
  -> microcode_update_cpu()
     -> microcode_resume_cpu() //since uci->valid is true
        -> apply_microcode_on_target()

I tried doing suspend-to-ram and suspend-to-disk on latest kernel (3.2-rc7+)
and I saw exactly the same messages getting emitted from the microcode
module in both cases. Not sure what is going wrong in your case...
Are you really sure that your hibernation script (pm-hibernate) does not
unload the microcode module? You could look up the hooks installed in
/usr/lib/pm-utils/sleep.d/ and also /var/log/pm-suspend.log

If you want to try out by-passing all those hooks and attempting a
suspend-to-disk directly, do:

echo platform > /sys/power/disk  //or shutdown, if platform doesn't exist
echo none > /sys/power/pm_test
echo disk > /sys/power/state

Regards,
Srivatsa S. Bhat

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/