Re: [PATCH] EDAC: Fix global-out-of-bounds write when setting edac_mc_poll_msec
From: Luck, Tony
Date: Thu Jun 27 2019 - 13:36:57 EST
On Thu, Jun 27, 2019 at 06:11:18PM +0100, James Morse wrote:
> (CC: +Tony Luck.
> Original Patch: lore.kernel.org/r/20190626054011.30044-1-devel@xxxxxxxxxxxx )
Heh: My mail agent "helpfully" made that clickable, but as a "mailto:"
URL rather than an https: one!
> On 26/06/2019 06:40, Eiichi Tsukata wrote:
> > Commit 9da21b1509d8 ("EDAC: Poll timeout cannot be zero, p2") assumes
> > edac_mc_poll_msec to be unsigned long, but the type of the variable still
> > remained as int. Setting edac_mc_poll_msec can trigger out-of-bounds
> > write.
> Thanks for catching this!
Ditto & likewise.
> > Fix it by changing the type of edac_mc_poll_msec to unsigned int.
> This means reverting more of 9da21b1509d8, but it also fixes signed/unsigned issues:
> | root@debian-guest:/sys/module/edac_core/parameters# echo 4294967295 > edac_mc_poll_msec
> | root@debian-guest:/sys/module/edac_core/parameters# cat edac_mc_poll_msec
> | -1
> | root@debian-guest:/sys/module/edac_core/parameters# echo -1 > edac_mc_poll_msec
> | -bash: echo: write error: Invalid argument
> > The reason why this patch adopts unsigned int rather than unsigned long
> > is msecs_to_jiffies() assumes arg to be unsigned int.
> Ah, so the range is limited anyway.
> It looks like it was switched to long to be consistent with edac_mc_workq_setup(), which
> has since been removed in preference to msecs_to_jiffies().
> Reviewed-by: James Morse <james.morse@xxxxxxx>
Applied. (Boris left me in charge of EDAC bits for the
next few weeks). It will show up in the for_5.3 branch
after test builds complete.