On Fri, Jan 13, 2023 at 03:53:44PM +0000, Matthew Wilcox wrote: > On Fri, Jan 13, 2023 at 09:14:15PM +0530, Naresh Kamboju wrote: > > [ 15.980869] Hardware name: Supermicro SYS-5019S-ML/X11SSH-F, BIOS > > 2.0b 07/27/2017 > > [ 15.988336] RIP: 0010:do_wp_page (memory.c:?) > Uh, are you compiling your kernels without debuginfo? The results > from syzbot & 0day are much more useful: > https://lore.kernel.org/linux-mm/Y8FnAwWOxLrfoWTN@xxxxxxxxxxxxxxxxxxxx/T/#u > for an example. This will be a consequence of x86 not including DEBUG_INFO in defconfig - a lot of the testing systems cover unmodified upstream configs as a baseline, it tends to make life easier and be an indication of what people want covering. A quick survey says most architectures are setting at least DEBUG_INFO_REDUCED, it'd probably be a good idea for x86 to do so as well. The logs from arm64 and probably also arm (depending on exactly which config was being covered) should be more usefully decodable, Naresh you mentioned that those were also failing?
Attachment:
signature.asc
Description: PGP signature
