linux-kernel - Re: [PATCH 09/11] EDAC/ghes: Implement DIMM mapping table for SMBIOS handles

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:   Mon, 16 Mar 2020 10:40:56 +0100
From:   Borislav Petkov <bp@...en8.de>
To:     Robert Richter <rrichter@...vell.com>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        Tony Luck <tony.luck@...el.com>,
        James Morse <james.morse@....com>,
        Aristeu Rozanski <aris@...hat.com>, linux-edac@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 09/11] EDAC/ghes: Implement DIMM mapping table for SMBIOS
 handles

On Fri, Mar 06, 2020 at 04:13:16PM +0100, Robert Richter wrote:
> GHES uses SMBIOS handles to identify the DIMM that was causing a hw
> error. Currently this is stored in smbios_handle of struct dimm_info.
> This implementation has several drawbacks. The information is private
> to the ghes driver, but struct dimm_info is for general use. The DIMMs
> are tied to a *mci struct, which makes a lockup inefficient. It is

"lookup"

> hard to dynamically allocate DIMMs and also to meet locking
> constraints when adding or removing them. This becomes even more
> challenging when having multiple MC instances that group a set of
> DIMMs, so the change is needed also to later support multiple MC
> instances.

Err, I don't understand: normally a bunch of DIMMs belong to a memory
controller and that gives you the hierarchy automatically. Why is that a
problem?

And normally you allocate the DIMM representations on MC init and
deallocate them when removing the MC.

So what is the problem here?

...

...
> @@ -72,6 +79,52 @@ struct memdev_dmi_entry {
>  	u16 conf_mem_clk_speed;
>  } __attribute__((__packed__));
>  
> +/* ghes_reg_mutex must be held. */

lockdep_assert_held()

...

-- 
Regards/Gruss,
    Boris.

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