Bug#866306: patch for haveged
Control: tag -1 + patch upstream moreinfo Control: severity -1 important Hi Natanael ! Thanks for providing the patch. Did you forward this patch upstream, and do you know whether it was included in the latest upstream release? Hi Vagrant ! Could you confirm whether the patch solves the bug on your hardware? Best, nicoo On Fri, May 18, 2018 at 04:58:10PM +0200, Natanael Copa wrote: > Tag: Patch > > Hi, > > I bumped into this issue on Alpine Linux. It appears that the sysfs not > always report the cpu cache size on arm (both 32bit and 64 bit). It > reports size to be -1 and haveged does not handle that case. > > I have seen the case on xgene and thunderx machines. > > I have attached a patch that should fix the issue. > > Thanks! > > -nc signature.asc Description: PGP signature
Bug#866306: patch for haveged
On 2018-09-03, Nicolas Braud-Santoni wrote: > Hi Natanael ! > > Thanks for providing the patch. > > Did you forward this patch upstream, and do you know whether it was > included in the latest upstream release? > > > Hi Vagrant ! > > Could you confirm whether the patch solves the bug on your hardware? I'd be happy to try it out! Unfortunately, the patch sent to the debian BTS is some xml representation of a patch; I'm not sure how to apply it. Would it be possible to resend the patch in a more conventional patch format? It looks like there's a diff somewhere in here: diff --git a/src/havegetune.c b/src/havegetune.c index f1a99f2..de39c53 100644 --- a/src/havegetune.c +++ b/src/havegetune.c @@ -795,6 +795,8 @@ static int vfs_configInfoCache( ctype = vfs_configFile(pAnchor, path, vfs_configType); strcpy(path+plen, "size"); size = vfs_configFile(pAnchor, path, vfs_configInt); + if (size == -1) + size = ctype == 'I' ? GENERIC_ICACHE : GENERIC_DCACHE; cfg_cacheAdd(pAnchor, SRC_VFS_INDEX, pArgs[1], level, ctype, size); } } live well, vagrant signature.asc Description: PGP signature
Bug#866306: patch for haveged
Tag: Patch Hi, I bumped into this issue on Alpine Linux. It appears that the sysfs not always report the cpu cache size on arm (both 32bit and 64 bit). It reports size to be -1 and haveged does not handle that case. I have seen the case on xgene and thunderx machines. I have attached a patch that should fix the issue. Thanks! -nc fix-cpu-cache-size-detection.patch\haveged\main - aports - Main aports tree index : aports 1.10-stable 1.9 2.0-stable 2.1-stable 2.2-stable 2.3-stable 2.4-stable 2.5-stable 2.6-stable 2.7-stable 3.0-stable 3.1-stable 3.2-stable 3.3-stable 3.4-stable 3.5-stable 3.6-stable 3.7-stable master Main aports tree gitolite aboutsummaryrefslogtreecommitdiffstats log msg author committer range path: root/main/haveged/fix-cpu-cache-size-detection.patchblob: 07da2b46d94bbda540d74592615cae6a91b8 (plain) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Some ARM cpus does not report the cache size or say it is -1 diff --git a/src/havegetune.c b/src/havegetune.c index f1a99f2..de39c53 100644 --- a/src/havegetune.c +++ b/src/havegetune.c @@ -795,6 +795,8 @@ static int vfs_configInfoCache( ctype = vfs_configFile(pAnchor, path, vfs_configType); strcpy(path+plen, "size"); size = vfs_configFile(pAnchor, path, vfs_configInt); + if (size == -1) + size = ctype == 'I' ? GENERIC_ICACHE : GENERIC_DCACHE; cfg_cacheAdd(pAnchor, SRC_VFS_INDEX, pArgs[1], level, ctype, size); } } var gaJsHost = (("https:" == document.location.protocol) ? "https://ssl."; : "http://www.";); document.write(unescape("%3Cscript src='" + gaJsHost + "google-analytics.com/ga.js' type='text/javascript'%3E%3C/script%3E")); try { var pageTracker = _gat._getTracker("UA-2804446-12"); pageTracker._trackPageview(); } catch(err) {}
Bug#866306: patch for haveged
On Mon, 3 Sep 2018 21:54:08 +0200 Nicolas Braud-Santoni wrote: > Control: tag -1 + patch upstream moreinfo > Control: severity -1 important > > Hi Natanael ! > > Thanks for providing the patch. > > Did you forward this patch upstream, and do you know whether it was > included in the latest upstream release? Hi, At the time i didn't know where to send it upstream, but it seems like the place is github now. I have forwarded this upstream: https://github.com/jirka-h/haveged/pull/7 Thanks! -nc pgpSoq8ogxJPo.pgp Description: OpenPGP digital signature
Bug#866306: patch for haveged
On Mon, 03 Sep 2018 13:57:17 -0700 Vagrant Cascadian wrote: > On 2018-09-03, Nicolas Braud-Santoni wrote: > > Hi Vagrant ! > > > > Could you confirm whether the patch solves the bug on your hardware? > > I'd be happy to try it out! Unfortunately, the patch sent to the debian > BTS is some xml representation of a patch; I'm not sure how to apply > it. Would it be possible to resend the patch in a more conventional > patch format? You can fetch[1] it from upstream pull request[2]. It would also be nice if you can give a comment there if it works or not. Thanks! [1]: https://patch-diff.githubusercontent.com/raw/jirka-h/haveged/pull/7.patch [2]: https://github.com/jirka-h/haveged/pull/7 -nc pgpfUYm8OZpC5.pgp Description: OpenPGP digital signature
Bug#866306: patch for haveged
On 2018-09-19, Natanael Copa wrote: > On Mon, 03 Sep 2018 13:57:17 -0700 > Vagrant Cascadian wrote: >> On 2018-09-03, Nicolas Braud-Santoni wrote: >> > Could you confirm whether the patch solves the bug on your hardware? >> >> I'd be happy to try it out! Unfortunately, the patch sent to the debian >> BTS is some xml representation of a patch; I'm not sure how to apply >> it. Would it be possible to resend the patch in a more conventional >> patch format? > > You can fetch[1] it from upstream pull request[2]. It would also be > nice if you can give a comment there if it works or not. > > Thanks! > > [1]: https://patch-diff.githubusercontent.com/raw/jirka-h/haveged/pull/7.patch > [2]: https://github.com/jirka-h/haveged/pull/7 Unfortunately, it didn't fix the issue for me. The workaround of passing the --data=16 argument still does. Maybe this is an entirely different issue you were experiencing? I'll follow up on the pull request, too. live well, vagrant signature.asc Description: PGP signature