|
From: Benjamin C. <ben...@si...> - 2006-01-24 15:58:19
Attachments:
benjamin.collar.vcf
|
Hello I am using valgrind-3.1.0 within a UML session. It runs fine for about 30 seconds, then does this: map : /proc/mm map failed, err = 12 infinitely. The kernel is 2.4.20 with the SKAS patches installed. The UML kernels are Montavista's CGE 3.1 with the latest UML patches applied. I tried MV Version 4 on a 2.6/UML kernel but it worked less well than the 3.1--valgrind refused to start, running out of memory immediately. Even valgrind --help doesn't work in that configuration. The application is a beast taking hundreds of megabytes of dynamic memory. I am happy to provide more details as necessary--please let me know. Any tips detailing how I can follow/debug/fix this problem would be appreciated. Thanks for any help Benjamin |
|
From: Tom H. <to...@co...> - 2006-01-24 17:10:08
|
In message <43D...@si...>
Benjamin Collar <ben...@si...> wrote:
> I am using valgrind-3.1.0 within a UML session. It runs fine for about
> 30 seconds, then does this:
>
> map : /proc/mm map failed, err = 12
>
> infinitely.
Well that isn't a valgrind message, so the question is, where is it
coming from and what does it mean...
I assume this appears in the window you are running valgrind in, and
not in the kernel log or anything? If so then as it isn't something
that valgrind would print it must be something the program you are
running valgrind is printing.
> The application is a beast taking hundreds of megabytes of dynamic memory.
Well that is probably the problem then - error 12 is ENOMEM so most
likely your program has run out of memory. Don't forget that under
valgrind it will be using a lot more memory than normal.
Is this UML instance 32 or 64 bit? A 32 bit instance will be quite
restricted in how much memory a program can allocate under valgrind.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: Nicholas N. <nj...@cs...> - 2006-01-24 22:12:53
|
On Tue, 24 Jan 2006, Tom Hughes wrote: >> The application is a beast taking hundreds of megabytes of dynamic memory. > > Well that is probably the problem then - error 12 is ENOMEM so most > likely your program has run out of memory. Don't forget that under > valgrind it will be using a lot more memory than normal. If you're running Memcheck (the default tool), try running --tool=none, which will use less memory. Nick |
|
From: Benjamin C. <ben...@si...> - 2006-01-25 14:44:06
Attachments:
benjamin.collar.vcf
|
Hi everyone The solution was simple: throw more hardware at it. Bought an extra gigabyte and now things are as happy as they should be. Thanks for responding Ben Nicholas Nethercote wrote: > On Tue, 24 Jan 2006, Tom Hughes wrote: > >>> The application is a beast taking hundreds of megabytes of dynamic >>> memory. >> >> >> Well that is probably the problem then - error 12 is ENOMEM so most >> likely your program has run out of memory. Don't forget that under >> valgrind it will be using a lot more memory than normal. > > > If you're running Memcheck (the default tool), try running --tool=none, > which will use less memory. > > Nick > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 > _______________________________________________ > Valgrind-users mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-users |
|
From: Benjamin C. <ben...@si...> - 2006-01-25 17:31:37
Attachments:
benjamin.collar.vcf
|
Hello again Funnily enough, throwing more memory at it didn't solve the problem. I was just fooled for a short while. The real problem is that none of UML or Valgrind or my program are running out of memory. Valgrind and my program are running out of memory maps! This fixed it: echo 262144 > /proc/sys/vm/max_map_count And here is where I found the answer: http://uml.harlowhill.com/index.php/Troubleshooting Benjamin Nicholas Nethercote wrote: > On Tue, 24 Jan 2006, Tom Hughes wrote: > >>> The application is a beast taking hundreds of megabytes of dynamic >>> memory. >> >> >> Well that is probably the problem then - error 12 is ENOMEM so most >> likely your program has run out of memory. Don't forget that under >> valgrind it will be using a lot more memory than normal. > > > If you're running Memcheck (the default tool), try running --tool=none, > which will use less memory. > > Nick > > > ------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. Do you grep through log > files > for problems? Stop! Download the new AJAX search engine that makes > searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 > _______________________________________________ > Valgrind-users mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-users |