On Thu, Oct 29, 2026 at 12:34 AM Feng Yang <[email protected]> wrote: > > On Wed, 28 Oct 2026 20:51:47 +0800, Songsen Yang wrote: > > Hi Feng, > > > > I hit the same root cause independently via syzbot and sent a fix a > > few hours after yours, without having seen this patch -- sorry for the > > overlap: > > > > https://lore.kernel.org/bpf/[email protected]/ > > Hi Songsen, > > Please don't worry about the overlap. > > I hadn't noticed the issue Sashiko raised when I prepared my fix; > It looks like a pt_regs on the stack can fix this.: > > BPF_CALL_4(bpf_get_stack_raw_tp_sleepable, struct > bpf_raw_tracepoint_args *, args, > void *, buf, u32, size, u64, flags) > { > struct pt_regs regs; > > perf_fetch_caller_regs(®s); > return bpf_get_stack_sleepable((unsigned long) ®s, > (unsigned long) buf, > (unsigned long) size, flags, > 0);
using sleepable variant of bpf_get_stacks is a better approach, because stack traces might request build id, and in sleepable mode that works reliably instead of bailing early if memory is not paged in, so I think having specialized sleepable version is better. > } > > That said, I think your approach is the simpler one and the better fit > for the fix itself. So if the maintainers prefer your version, feel > free to ignore my patch. > > Thanks, > Feng >
