Control: tags -1 + pending On Tue, 2015-08-11 at 18:55 +0200, Cyril Brulebois wrote: > Control: tag -1 confirmed > > Hi Benjamin, > > Benjamin Kaduk <ka...@mit.edu> (2015-08-08): > > The security update to the linux kernel (3.2.65-1+deb7u2) changed the > > KPI by moving d_alias into the d_u union, which causes subsequent > > (re)build attempts of the openafs kernel module to fail. (Existing > > openafs kernel modules continued to work due to clever work by the > > kernel maintainer.) This situation was maintained without great > > incident until the upload of openafs 1.6.1-3+deb7u3 to the security > > repo, which cause most people to attempt to build a new openafs kernel > > module and fail. Two grave bugs (780865 and 794224) reflect that > > situation. > > It would be slightly better to have those filed against the proper > binary package or the source package. They're currently recorded in > the BTS against (binary) openafs which doesn't exist. > > (For the record I've spotted this because the first thing I look for is > the bug graph, to see what the status of the bug report is for all > suites, as we usually like unstable/testing to have fixed versions in > the first place, or as not affected.) > > > We had to deal with this problem for jessie back in February, but I > > didn't have enough time to fix it for wheezy then. The jessie > > updates also included a fix for a bug causing potential data > > corruption in mmapped files; I included the fix for that (778851) in > > this w-p-u as well. > > All of this seems rather reasonable. The first bullet point of the > changelog entry seems like it has suffered from an unedited copy/paste > after the bug closures though? Quoting it: > > > diff -Nru openafs-1.6.1/debian/changelog openafs-1.6.1/debian/changelog > > --- openafs-1.6.1/debian/changelog 2015-07-30 06:27:51.000000000 -0400 > > +++ openafs-1.6.1/debian/changelog 2015-08-08 14:23:47.000000000 -0400 > > @@ -1,3 +1,12 @@ > > +openafs (1.6.1-3+deb7u4) wheezy-proposed-updates; urgency=medium > > + > > + * Fix the kernel module build when d_alias is in the d_u union > > + (Closes: #780865, Closes: #794224) > > + to d_alias being a member of the d_u union. > > Feel free to upload your package, with or without a possible edition of > this entry.
Uploaded and flagged for acceptance. Regards, Adam -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org