New issue
Advanced search Search tips

Issue 658705 link

Starred by 1 user
Status: WontFix
Owner:
mark@chromium.org
Closed: Oct 2016
Cc:
scottmg@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 3
Type: Bug



Sign in to add a comment

Crashpad spam in debug.log

Project Member Reported by grt@chromium.org, Oct 24 2016 
Where by "spam" I mean messages that are meaningless to end-users. Are these appearing in my canary's debug.log because of a bug somewhere that needs to be fixed, or should the logging be removed? Thanks.

[1014/100330:ERROR:process_info.cc(608)] range at 0xab18262600000000, size 0x24a fully unreadable
[1014/100330:ERROR:process_info.cc(608)] range at 0xab18264600000000, size 0x24a fully unreadable
[1014/100330:ERROR:process_info.cc(608)] range at 0x0, size 0x24a fully unreadable
[1014/100331:WARNING:in_range_cast.h(38)] value 2520103820140 out of range
[1014/100331:WARNING:in_range_cast.h(38)] value 2520103821508 out of range
[1014/100331:WARNING:in_range_cast.h(38)] value 2520103821848 out of range
[1014/100331:WARNING:in_range_cast.h(38)] value 2520103818878 out of range
[1014/100331:WARNING:in_range_cast.h(38)] value 2519770311520 out of range

Comment 1 by mark@chromium.org, Oct 24 2016

Status: WontFix (was: Assigned)
They’re valuable in the context of the crash report that Crashpad was producing, in that they say “hey, this part of the crash report I was trying to grab? Here’s why I couldn’t do it.” They probably point to memory corruption in the process but they serve as a signal that the crash report may be incomplete, and may contain information about how the process got sick.

Bug crashpad:26 is about carrying these sorts of messages along in the crash dump itself.

Comment 2 by scottmg@chromium.org, Oct 24 2016 

In Canary, these could be walking indirect possible-pointers from the stack as opposed to memory that necessarily is expected to be pointable-to. If that's the case, we probably shouldn't log during that part of memory gathering.

(That applies to the "range at" ones, not sure about "value X out of range".)

Comment 3 by grt@chromium.org, Oct 24 2016 

Bug crashpad:26 sounds like the right thing. Since there's no general way to discover the messages in a user's debug.log, how about not emitting them? There's no need to use up disk space all over the world that you'll never see, is there?
►

Sign in to add a comment