Re: [PATCH] RFC: debugobjects: capture stack traces at _init() time

2018-03-23 Thread Daniel Vetter
On Tue, Mar 20, 2018 at 8:44 PM, Thomas Gleixner wrote: > On Tue, 20 Mar 2018, Daniel Vetter wrote: > >> Sometimes it's really easy to know which object has gone boom and >> where the offending code is, and sometimes it's really hard. One case >> we're trying to hunt down is when module unload cat

Re: [PATCH] RFC: debugobjects: capture stack traces at _init() time

2018-03-20 Thread Thomas Gleixner
On Tue, 20 Mar 2018, Daniel Vetter wrote: > Sometimes it's really easy to know which object has gone boom and > where the offending code is, and sometimes it's really hard. One case > we're trying to hunt down is when module unload catches a live debug > object, with a module with lots of them. >

[PATCH] RFC: debugobjects: capture stack traces at _init() time

2018-03-20 Thread Daniel Vetter
Sometimes it's really easy to know which object has gone boom and where the offending code is, and sometimes it's really hard. One case we're trying to hunt down is when module unload catches a live debug object, with a module with lots of them. Capture a full stack trace from debug_object_init()