Re: [dm-devel] dm-crypt hard lockup
On 6/28/20 6:50 PM, Bart Van Assche wrote: > Since considerable time I'm doing kernel builds (make -j8) on an > openSUSE Tumbleweed system on top of dm-crypt and an NVMe SSD but I have > not yet encountered any kind of lockup. Maybe another driver, e.g. an > I/O scheduler, is responsible for the lockups? It could by another driver. I tried with a different filesystem (xfs), without any I/O scheduler and even with the sources on a different drive without dm-crypt, and the results were more or less the same - lockups involving dm-crypt. Maybe I'm just abusing the system - I'm actually running "make -j", so the number of tasks is unlimited. With a limited number it does not hang. I know it's not a good idea to do this but it used to work, and I think it shouldn't cause the kernel to die like this. Artur -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
Re: [dm-devel] dm-crypt hard lockup
On 2020-06-26 02:07, Artur Paszkiewicz wrote: > I'm getting regular lockups which seem to be caused by dm-crypt. I > reproduced it on vanilla v5.8-rc2, but I started regularly seeing this > some time ago on openSUSE Tumbleweed kernels. It's easily reproducible > (every time, after about a minute) when I run "make -j" on the linux > kernel sources, sometimes it occurs also when doing other IO intensive > tasks on multiple CPUs. I'm using LVM and ext4 on dm-crypt devices, > Intel SSDSC2KW010X6 and SSDSC2BA200G3 SSDs. Since considerable time I'm doing kernel builds (make -j8) on an openSUSE Tumbleweed system on top of dm-crypt and an NVMe SSD but I have not yet encountered any kind of lockup. Maybe another driver, e.g. an I/O scheduler, is responsible for the lockups? Bart. -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel