Hi Salvatore, Thanks for chasing this up, I've just built the 4.19.144 vanilla kernel with the referenced commit applied and can confirm that this issue no longer occurs on writing to a mounted gfs2 filesystem.
Cheers, Dan. ________________________________ From: Salvatore Bonaccorso <salvatore.bonacco...@gmail.com> on behalf of Salvatore Bonaccorso <car...@debian.org> Sent: Friday, 11 September 2020 11:15 PM To: Craig, Daniel (CASS, Marsfield) <daniel.cr...@csiro.au>; 968...@bugs.debian.org <968...@bugs.debian.org> Cc: Nicolas Courtel <cour...@cena.fr> Subject: Re: Bug#968567: linux-image-4.19.0-10-amd64: kernel failure when writing on a GFS2 partition Hi Daniel, hi Nicolas, On Thu, Sep 10, 2020 at 04:47:20AM +0000, Craig, Daniel (CASS, Marsfield) wrote: > Hi, > > I can confirm the existence of this CPU soft-lock bug with gfs2. > > I won't worry about reproducing the kernel bug message, but I have > done a bit of digging and if I revert the following commit, added in > the 4.19.130 release then this fixes issue for me. > > https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-4.19.y&id=c91cffd0fd010c06d67f3a9a528b858ce28c60fb > > Note that the problem is still present in the latest upstream > release in the 4.19 series (4.19.144) > > I've reported this bug in the kernel bugzilla, referenced here: > > https://bugzilla.kernel.org/show_bug.cgi?id=209217 Upstream did had a look at this see (you both were CC'ed on that thread though) the thread starting at https://lore.kernel.org/stable/20200910194319.GA131386@eldamar.local/ and suggested that upstream commit cbcc89b630447ec7836aa2b9242d9bb1725f5a61 is definitely needed. Would it be possible for you to test a build with that commit added on top? (Note the patch would need a slight refresh when applying on top of 4.19.144 for context changes). https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s4.2.2 explains how to test a single patch on top of the packaging, or you can try 4.19.144 with the cherry-picked commit. That would be much appreciated. Regards, Salvatore
