26 and 27 kernels was released. http://wiki.openvz.org/Download/kernel/2.6.26/2.6.26-chekhov.1 http://wiki.openvz.org/Download/kernel/2.6.27/2.6.27-aivazovsky.1 Also this mean, that it's time to add 2.6.27.9999 (in masked state) ebuild. Reproducible: Always Steps to Reproduce:
Created attachment 179199 [details] sys-kernel/openvz-sources-2.6.27.9999 ebuild Simple renaming 2.6.26.9999 with description change
Created attachment 179217 [details] sys-kernel/openvz-sources-2.6.27.1-r1 [aivazovsky] OpenVZ team seems to have changed their versioning scheme. I'm using forth version component to difference codenamed kernels, specifying codename in ebuild itself. This is less portable, but seems to be better evil than change atom name on every new codename =). Resulting kernel name is 2.6.27-openvz-aivazovsky.1 I have also specify SLOT especially, to use codename in it. Now it looks in eix like that: "(2.6.27-aivazovsky.1) [M](~)2.6.27.1-r1!b!s"
Created attachment 179225 [details] sys-kernel/openvz-sources-2.6.26.1-r1 [chekhov] Same for 2.6.26-chekhov.1 release. I suppose that warning about different from upstream versions needed, but my English is too bad to write it myself. kernel dir name: linux-2.6.26-openvz-chekhov.1 kernel name: v2.6.26-openvz-chekhov.1 slot: (2.6.26-chekhov.1) [M](~)2.6.26.1-r1!b!s
Also, IMHO, package.mask a little bit outdated...
Please don't CC maintainers when you cannot assign bugs.
Sorry, I just don't know how to act in such situation. Next time I'll be more accurate, thanks for note. 2herd: Is there any problem with comming it? Maybe I can help somehow... Only known problem with this kernel (still unconfirmed) - it maybe unstable with started VEs (in some, still unclear, conditions) on uniprocessor x86 systems. I've already issued upstream bug for this issue, but, taking into account, that this kernels rather not adopted for such low-perfomance enviroment and any post-18 kernels hardmasked anyway, that mustn't be a big problem. On multicore amd64 systems, following other users reactions, it works flawlessly (status unknown for single-core ia32_64 or multi-core ia32 systems). If you need any futher information or improvements to ebuilds, just make us know ;).
*** Bug 256058 has been marked as a duplicate of this bug. ***
Should these not be 2 bugs? One for 2.6.26 and one for 2.6.27? And is there any news on the state of either one?
Mantainer know about them, but have no time to add. Also they are consisted as useless - we have git-ebuilds in tree, so use them. Both of snapshots suffers from it_real_fn BUG, which can hit low-perfomance machines (currently, both git kernels suffers from it too - you can apply a patch http://bugzilla.openvz.org/attachment.cgi?id=905). So, just use git ebuild, you can check new features and fixes at http://git.openvz.org/?p=linux-2.6.27-openvz;a=summary
The upsream bug associated to the it_real_fn issue: http://bugzilla.openvz.org/show_bug.cgi?id=1160 (It is RESOLVED/FIXED, so I guess the patch was already merged into Git?) Currently there is no 2.6.27.9999 ebuild in the portage tree, "just" a 2.6.26.9999 one.
0o... Really? I don't notice that, I supposed that all git ebuilds are in tree... At least pva said something like this... maybe =) Bug resolved/fixed by me - patch contents shows that it is comming from developer's local repository, I suppose we just must wait untill he will finish some project and will push all changes. That's not too hard to apply it by hands =). Or you can use my overlay - https://svn.niifaq.ru/portage/
night nord: your ebuilds for chekhov and aivazovsky look good, except that one shouldnt use the revision to indicate a different upstream release. see http://devmanual.gentoo.org/general-concepts/ebuild-revisions/index.html but openvz-sources-2.6.27.1.1 would work.
(In reply to comment #9) > Both of snapshots suffers from it_real_fn BUG, which can hit > low-perfomance machines (currently, both git > kernels suffers from it too - you can apply a patch > http://bugzilla.openvz.org/attachment.cgi?id=905). Patch apparently hit Git now: http://git.openvz.org/?p=linux-2.6.26-openvz;a=commit;h=0328e3d32c6915650b14dd40fcd7598a420b1364
Comment on attachment 179199 [details] sys-kernel/openvz-sources-2.6.27.9999 ebuild openvz-sources-2.6.27.9999 is now in the tree. thanks.
Thanks. I've fixed -r* issue in releases, thank you for notice, but both kernels suffers from courple of serious bugs, so they are unusable on production. 2.6.27-openvz.git-dc7245a just ok.
Comment on attachment 179217 [details] sys-kernel/openvz-sources-2.6.27.1-r1 [aivazovsky] Buggy
Comment on attachment 179225 [details] sys-kernel/openvz-sources-2.6.26.1-r1 [chekhov] Buggy
Marking fixed until next release.
i've added 2.6.26.2.1 and 2.6.27.1.1 to my dev overlay - as templates for when new non-broken versions will be released by upstream...
