linux-kernel - RE: [PATCH] x86/mce: Increase the size of the MCE pool from 2 to 8 pages

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] [day] [month] [year] [list]

Message-ID: <SJ1PR11MB60832922E4D036138FF390FAFCD7A@SJ1PR11MB6083.namprd11.prod.outlook.com>
Date:   Mon, 16 Oct 2023 16:14:25 +0000
From:   "Luck, Tony" <tony.luck@...el.com>
To:     Yazen Ghannam <yazen.ghannam@....com>,
        "Hansen, Dave" <dave.hansen@...el.com>,
        "Sironi, Filippo" <sironi@...zon.de>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
CC:     "bp@...en8.de" <bp@...en8.de>,
        "tglx@...utronix.de" <tglx@...utronix.de>,
        "mingo@...hat.com" <mingo@...hat.com>,
        "dave.hansen@...ux.intel.com" <dave.hansen@...ux.intel.com>,
        "x86@...nel.org" <x86@...nel.org>, "hpa@...or.com" <hpa@...or.com>,
        "linux-edac@...r.kernel.org" <linux-edac@...r.kernel.org>
Subject: RE: [PATCH] x86/mce: Increase the size of the MCE pool from 2 to 8
 pages

> > An automatic resizing one doesn't have to be fancy and only has to
> > expand once:
> >
>
> Why once?
>
> > static bool expanded = false;
> >
> > ...
> >
> >     if (full && !expanded) {
> >             expand();
> >             expanded = true;
> >     }
> >
> > It might be a _wee_ bit worse than that because you might have to queue
> > some work outside of #MC context but seriously we're talking 10-ish
> > lines of code.  It'd probably be even smaller than doing it when poked
> > by userspace and wouldn't involve new ABI.
>
> Okay, I'm with you here.

I'm not. The reason that pool exists is so that an error record (struct mce)
can be copied into it during machine check context and atomically added
to the mce_event_llist. See mce_gen_pool_add().

I don't see how your "expand()" function is going to safely allocate additional
pages to the pool while executing in the machine check handler.

Just make it one of:

1) CONFIG option
2) boot command line option.
3) Dynamic based on num_online_cpus()

-Tony