Re: TOI problems
Nigel Cunningham <[email protected]> Mon, 10 Mar 2014 21:50:28 +1100
| Newsgroups | gmane.linux.swsusp.devel |
|---|---|
| Message-ID | <[email protected]> |
This is a multi-part message in MIME format. --===============1206199670== Content-Type: multipart/alternative; boundary="------------000106030909050207000205" This is a multi-part message in MIME format. --------------000106030909050207000205 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hi again. This was fixed in a git commit tonight. It was a simple error, easily fixed. We use scan_swap_map to find the first page in a range. This call does everything necessary to allocate that first page. We then use find_first_to_unuse to find the next page after that which is already allocated. We were then incorrectly calling inc_cluster_info_page for the first page as well as for the remaining pages in the range being allocated, resulting in the VM_BUG_ON being hit. The solution is to not call inc_cluster_info_page for that first page. Regards, Nigel On 05/03/14 22:47, Nigel Cunningham wrote: > Hi again. > > I've reproduced it by simply enabling the "Memory debugging" option in > the kernel config, so that's half the problem solved. It's 10.40pm > here, so I'll leave doing more until tomorrow. > > Nigel > > On 05/03/14 22:20, Nigel Cunningham wrote: >> Sorry again for the further delay. >> >> The screenshot is helpful; since you're running 64 bit, it's less >> likely to be an overflow issue. You're hitting a VM_BUG_ON, so I'll >> try modifying my set up to see if I can trigger it here. Perhaps >> you're using more paranoid settings than everyone else :) >> >> Regards, >> >> Nigel >> >> On 19/02/14 19:13, Matthias Hensler wrote: >>> Hi, >>> >>> On Wed, Feb 19, 2014 at 09:24:02AM +1100, Nigel Cunningham wrote: >>>> 64 bit kernel? >>> Yes, it is a 64bit-kernel. >>> >>> >>> Regards, >>> Matthias >>> >>> >>> _______________________________________________ >>> TuxOnIce-devel mailing list >>> [email protected] >>> http://lists.tuxonice.net/listinfo/tuxonice-devel >> >> >> >> _______________________________________________ >> TuxOnIce-devel mailing list >> [email protected] >> http://lists.tuxonice.net/listinfo/tuxonice-devel > > > > _______________________________________________ > TuxOnIce-devel mailing list > [email protected] > http://lists.tuxonice.net/listinfo/tuxonice-devel --------------000106030909050207000205 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit <html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body text="#000000" bgcolor="#FFFFFF"> Hi again.<br> <br> This was fixed in a git commit tonight. It was a simple error, easily fixed. We use scan_swap_map to find the first page in a range. This call does everything necessary to allocate that first page. We then use find_first_to_unuse to find the next page after that which is already allocated. We were then incorrectly calling inc_cluster_info_page for the<br> first page as well as for the remaining pages in the range being allocated, resulting in the<br> VM_BUG_ON being hit. The solution is to not call inc_cluster_info_page for that first page.<br> <br> Regards,<br> <br> Nigel<br> <br> <div class="moz-cite-prefix">On 05/03/14 22:47, Nigel Cunningham wrote:<br> </div> <blockquote cite="mid:[email protected]" type="cite"> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> Hi again.<br> <br> I've reproduced it by simply enabling the "Memory debugging" option in the kernel config, so that's half the problem solved. It's 10.40pm here, so I'll leave doing more until tomorrow.<br> <br> Nigel<br> <br> <div class="moz-cite-prefix">On 05/03/14 22:20, Nigel Cunningham wrote:<br> </div> <blockquote cite="mid:[email protected]" type="cite"> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> Sorry again for the further delay.<br> <br> The screenshot is helpful; since you're running 64 bit, it's less likely to be an overflow issue. You're hitting a VM_BUG_ON, so I'll try modifying my set up to see if I can trigger it here. Perhaps you're using more paranoid settings than everyone else :)<br> <br> Regards,<br> <br> Nigel<br> <br> <div class="moz-cite-prefix">On 19/02/14 19:13, Matthias Hensler wrote:<br> </div> <blockquote cite="mid:[email protected]" type="cite"> <pre wrap="">Hi, On Wed, Feb 19, 2014 at 09:24:02AM +1100, Nigel Cunningham wrote: </pre> <blockquote type="cite"> <pre wrap="">64 bit kernel? </pre> </blockquote> <pre wrap="">Yes, it is a 64bit-kernel. Regards, Matthias </pre> <br> <fieldset class="mimeAttachmentHeader"></fieldset> <br> <pre wrap="">_______________________________________________ TuxOnIce-devel mailing list <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:[email protected]">[email protected]</a> <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://lists.tuxonice.net/listinfo/tuxonice-devel">http://lists.tuxonice.net/listinfo/tuxonice-devel</a></pre> </blockquote> <br> <br> <fieldset class="mimeAttachmentHeader"></fieldset> <br> <pre wrap="">_______________________________________________ TuxOnIce-devel mailing list <a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:[email protected]">[email protected]</a> <a moz-do-not-send="true" class="moz-txt-link-freetext" href="http://lists.tuxonice.net/listinfo/tuxonice-devel">http://lists.tuxonice.net/listinfo/tuxonice-devel</a></pre> </blockquote> <br> <br> <fieldset class="mimeAttachmentHeader"></fieldset> <br> <pre wrap="">_______________________________________________ TuxOnIce-devel mailing list <a class="moz-txt-link-abbreviated" href="mailto:[email protected]">[email protected]</a> <a class="moz-txt-link-freetext" href="http://lists.tuxonice.net/listinfo/tuxonice-devel">http://lists.tuxonice.net/listinfo/tuxonice-devel</a></pre> </blockquote> <br> </body> </html> --------------000106030909050207000205-- --===============1206199670== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ TuxOnIce-devel mailing list [email protected] http://lists.tuxonice.net/listinfo/tuxonice-devel --===============1206199670==--