keeleysam / tenfourfox

Automatically exported from code.google.com/p/tenfourfox
0 stars 0 forks source link

More stack overflow madness #113

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
Reported at http://tenfourfox.blogspot.com/2011/11/beta-delay.html

Stack signature is mangled and at upper edge of stack, so it looks like even 
256MB of stack may not be sufficient. Disabling tracejit helps, so this is our 
old friend issue 37. We'll try to sneak another stack bump into 9b1.

Original issue reported on code.google.com by classi...@floodgap.com on 1 Dec 2011 at 12:49

GoogleCodeExporter commented 9 years ago
It will not be possible for us to expand the stack without relocating it. 384MB 
seems a reasonable jump.

Original comment by classi...@floodgap.com on 1 Dec 2011 at 1:40

GoogleCodeExporter commented 9 years ago
That didn't work for this site (!!!): 
http://hypervocal.com/news/2011/supernanny-state-200-pound-ohio-boy-taken-from-f
amily-placed-into-foster-care/

We had to run the stack up to a full 1GB. This is as high as we go with stack; 
there ain't no more. Hopefully methodjit fixes it. FTR, we adjusted stack_addr 
by adding 0xb0000000 to the stack_size, so for 0x40000000, the address is 
0xf0000000.

Future bugs like this will be wontfix because we cantfix.

Original comment by classi...@floodgap.com on 1 Dec 2011 at 2:12

GoogleCodeExporter commented 9 years ago
Is it possible to blacklist known sites or scripts so the jit gets dynamically 
disabled when they're visited? 

Or to utilize the unresponsive script warning (warning = tracejit aborts)? If I 
set dom.max_script_run_time = 1, it warns me and doesn't crash on the test site 
when I click stop script. Also it tells me which script is the culprit which 
may be useful for debugging.

No idea if any of this is technicaly feasable, just brain-storming. 

Once the full GB of stack is reached in Macs that have only 512 MB of RAM, 
people will force quit TFF anyway because the HD will thrash endlessly. 

Original comment by chtru...@web.de on 1 Dec 2011 at 10:56

GoogleCodeExporter commented 9 years ago

Original comment by chtru...@web.de on 1 Dec 2011 at 10:57

Attachments:

GoogleCodeExporter commented 9 years ago
Let's farm this into a separate bug (issue 114) so I can track the larger 
problem specifically.

Original comment by classi...@floodgap.com on 1 Dec 2011 at 2:27

GoogleCodeExporter commented 9 years ago

Original comment by classi...@floodgap.com on 1 Dec 2011 at 2:35

GoogleCodeExporter commented 9 years ago
Reported fixed by multiple users on selected testcase.

Original comment by classi...@floodgap.com on 3 Dec 2011 at 12:59