Hi Coly, Yes, 7d6b902ea0e0 seems to be the fix for it. I cannot reproduce this deadlock when I build this patch into 5.15.50. Thanks for pointing me to this patch! Regards, Nikhil. On Wed, 29 Jun 2022 at 18:33, Coly Li <colyli@xxxxxxx> wrote: > > > > > 2022年6月29日 20:30，Nikhil Kshirsagar <nkshirsagar@xxxxxxxxx> 写道： > > > > Hi Coly, > > > > I am not able to see this issue on 5.18.0-051800-generic > > (https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.18/amd64/), so I am > > guessing its been fixed since 5.15.50 sometime. If you are aware of > > which commit may be the fix please let me know. I will keep testing on > > 5.18.0-051800-generic to see if I see it ever. > > > Hi Nikhil, > > Thanks for the update. Recently there are several fixes go into stable 5.15.51 (but still one patch I just submitted to Greg yesterday, maybe it will show up in future 5.15.52). > > Maybe you can try stable tree 5.15.51 with appying one more patch, > commit 7d6b902ea0e0 (“bcache: memset on stack variables in bch_btree_check() and bch_sectors_dirty_init()”) > > BTW, the journal no-space deadlock fix is in 5.15.51 now. > > Coly Li > > > > > On Wed, 29 Jun 2022 at 16:03, Nikhil Kshirsagar <nkshirsagar@xxxxxxxxx> wrote: > >> > >> Hi Coly, > >> > >> please use this as a reference, this is the source code to refer to > >> for the kernel i was able to reproduce this on > >> (https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.15.50/) > >> > >> > >> commit 18a33c8dabb88b50b860e0177a73933f2c0ddf68 (tag: v5.15.50) > >> Author: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> > >> Date: Sat Jun 25 15:18:40 2022 +0200 > >> > >> Linux 5.15.50 > >> > >> Link: https://lore.kernel.org/r/20220623164322.288837280@xxxxxxxxxxxxxxxxxxx > >> Tested-by: Florian Fainelli <f.fainelli@xxxxxxxxx> > >> Tested-by: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx> > >> Tested-by: Bagas Sanjaya <bagasdotme@xxxxxxxxx> > >> Tested-by: Jon Hunter <jonathanh@xxxxxxxxxx> > >> Tested-by: Sudip Mukherjee <sudip.mukherjee@xxxxxxxxxxxxxxx> > >> Tested-by: Ron Economos <re@xxxxxxxx> > >> Tested-by: Guenter Roeck <linux@xxxxxxxxxxxx> > >> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> > >> > >> I am able to reproduce this bug on older kernels too, like 5.4.0-121. > >> I will also test the latest upstream kernel soon. > >> > >> Regards, > >> Nikhil. > >> > >> On Wed, 29 Jun 2022 at 13:47, Coly Li <colyli@xxxxxxx> wrote: > >>> > >>> > >>> > >>>> 2022年6月29日 16:09，Nikhil Kshirsagar <nkshirsagar@xxxxxxxxx> 写道： > >>>> > >>>> Hi Coly, > >>>> > >>>> Note I used partitions for the bcache as well as the hdd, not sure if > >>>> that's a factor. > >>>> > >>>> the kernel is upstream kernel - > >>>> > >>>> # uname -a > >>>> Linux bronzor 5.15.50-051550-generic #202206251445 SMP Sat Jun 25 > >>>> 14:51:22 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux > >>> > >>> > >>> Hi Nikhil, > >>> > >>> I don’t find the commit id 767db4b286c3e101ac220b813c873f492d9e4ee8 fro neither Linus tree nor stable tree. > >>> > >>> The tree you mentioned at https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.15.50/cod/mainline/v5.15.50 (767db4b286c3e101ac220b813c873f492d9e4ee8), I am not sure whether it is a clone of Linus tree or stable tree. Maybe you may try v5.15.50 from the stable tree (git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git), then we can focus on identical code base. > >>> > >>> Thanks. > >>> > >>> Coly Li > >>> > >>> > >>>>>>>>> > >>> > >>> [snipped] > >>>>>>>>> > >>>>>>>>> Is this a bug? It's in writeback mode. I'd setup the cache and run stuff like, > >>>>>>>>> > >>>>>>>>> echo 0 > /sys/block/bcache0/bcache/sequential_cutoff > >>>>>>>>> > >>>>>>>>> I had also echoed 0 into congested_read_threshold_us, > >>>>>>>>> congested_write_threshold_us. > >>>>>>>>> > >>>>>>>>> echo writeback > /sys/block/bcache0/bcache/cache_mode > >>>>>>>> > >>>>>>>> Where do you get the kernel? If this is stable kernel, could you give me the HEAD commit id? > >>>>>>>> > >>>>>>>> Coly Li > >>>>>>>> > >>>>>> > >>> >
