On Fri, May 22, 2020 at 06:51:20PM +0200, Petr Mladek wrote:
> On Fri 2020-05-15 11:44:30, Kees Cook wrote:
> > From: Pavel Tatashin <pasha.tatashin@xxxxxxxxxx>
> >
> > kmsg_dump() allows to dump kmesg buffer for various system events: oops,
> > panic, reboot, etc. It provides an interface to register a callback call
> > for clients, and in that callback interface there is a field "max_reason"
> > which gets ignored unless always_kmsg_dump is passed as kernel parameter.
>
> Strictly speaking, this is not fully true. "max_reason" field is not
> ignored when set to KMSG_DUMP_PANIC even when always_kmsg_dump was not set.
>
> It should be something like:
>
> "which gets ignored for reason higher than KMSG_DUMP_OOPS unless
> always_kmsg_dump is passed as kernel parameter".
>
> Heh, I wonder if anyone will be able to parse this ;-)
Ah yeah, good point. I've reworded things like this:
kmsg_dump() allows to dump kmesg buffer for various system events: oops,
panic, reboot, etc. It provides an interface to register a callback
call for clients, and in that callback interface there is a field
"max_reason", but it was getting ignored when set to any "reason"
higher than KMSG_DUMP_OOPS unless "always_kmsg_dump" was passed as
kernel parameter.
Allow clients to actually control their "max_reason", and keep the
current behavior when "max_reason" is not set.
> Otherwise, it looks good to me. With the updated commit message:
>
> Reviewed-by: Petr Mladek <pmladek@xxxxxxxx>
Thanks!
--
Kees Cook
