| Message ID | 20180226142957.6549-1-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Quoting Chris Wilson (2018-02-26 14:29:56) > For large oops dump, for example if ftrace is included, we can easily > exceed the storage buffer and lose the most important bit of > information: where the OOPS occurred. So repeat the location information > just before the end marker. > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Marta Lofstedt <marta.lofstedt@intel.com> I've plonked this in topic/core-for-CI to see if it is effective in helping us nail a few bugs. -Chris
OK cool I'll look out for it landing, CI_DRM_3838 or CI_DRM_3839 I guess. > -----Original Message----- > From: Chris Wilson [mailto:chris@chris-wilson.co.uk] > Sent: Tuesday, February 27, 2018 12:50 PM > To: intel-gfx@lists.freedesktop.org > Cc: Lofstedt, Marta <marta.lofstedt@intel.com> > Subject: Re: [RFC] kernel/panic: Repeat the line and caller information at the > end of the OOPS > > Quoting Chris Wilson (2018-02-26 14:29:56) > > For large oops dump, for example if ftrace is included, we can easily > > exceed the storage buffer and lose the most important bit of > > information: where the OOPS occurred. So repeat the location > > information just before the end marker. > > > > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> > > Cc: Marta Lofstedt <marta.lofstedt@intel.com> > > I've plonked this in topic/core-for-CI to see if it is effective in helping us nail a > few bugs. > -Chris
Quoting Lofstedt, Marta (2018-02-27 11:12:52)
> OK cool I'll look out for it landing, CI_DRM_3838 or CI_DRM_3839 I guess.
Bah, wrong path. This is only called to handle WARN() and we need the
extra information in BUG(). In fact, the information we need is in a raw
printk and not given to WARN, so we'll have to try a different tack.
-Chris
diff --git a/kernel/panic.c b/kernel/panic.c index 2cfef408fec9..9049168d9445 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -517,14 +517,9 @@ struct warn_args { va_list args; }; -void __warn(const char *file, int line, void *caller, unsigned taint, - struct pt_regs *regs, struct warn_args *args) +static void show_location(const char *file, int line, void *caller, + struct warn_args *args) { - disable_trace_on_warning(); - - if (args) - pr_warn(CUT_HERE); - if (file) pr_warn("WARNING: CPU: %d PID: %d at %s:%d %pS\n", raw_smp_processor_id(), current->pid, file, line, @@ -535,6 +530,17 @@ void __warn(const char *file, int line, void *caller, unsigned taint, if (args) vprintk(args->fmt, args->args); +} + +void __warn(const char *file, int line, void *caller, unsigned taint, + struct pt_regs *regs, struct warn_args *args) +{ + disable_trace_on_warning(); + + if (args) + pr_warn(CUT_HERE); + + show_location(file, line, caller, args); if (panic_on_warn) { /* @@ -554,6 +560,8 @@ void __warn(const char *file, int line, void *caller, unsigned taint, else dump_stack(); + show_location(file, line, caller, args); + print_oops_end_marker(); /* Just a warning, don't kill lockdep. */
For large oops dump, for example if ftrace is included, we can easily exceed the storage buffer and lose the most important bit of information: where the OOPS occurred. So repeat the location information just before the end marker. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Marta Lofstedt <marta.lofstedt@intel.com> --- git add ftw -Chris --- kernel/panic.c | 22 +++++++++++++++------- 1 file changed, 15 insertions(+), 7 deletions(-)