| Summary: | sys-apps/sreadahead removal (was sys-apps/sreadahead-1.0 version bump) | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Auke Booij (tulcod) <auke> |
| Component: | New packages | Assignee: | Jeremy Olexa (darkside) (RETIRED) <darkside> |
| Status: | RESOLVED FIXED | ||
| Severity: | enhancement | CC: | treecleaner |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| URL: | http://code.google.com/p/sreadahead/issues/detail?id=5 | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
sys-apps/sreadahead-1.0.ebuild
sreadahead-manual.patch sreadahead_iter_ctrl.patch Patch for 2.6.29 |
||
|
Description
Auke Booij (tulcod)
2009-02-01 13:04:02 UTC
Created attachment 180532 [details]
sys-apps/sreadahead-1.0.ebuild
needs two patches i will attach in a few seconds (shamelessly stolen from upstream), and the two *.rc files already in the tree for 0.04
Created attachment 180534 [details, diff]
sreadahead-manual.patch
this is the kernel patch, which is to be placed in /usr/share/doc/sreadahead-1.0/
Created attachment 180535 [details, diff] sreadahead_iter_ctrl.patch this fixes some bit of code. apparently, it "fixes the problem of an empty pack file". also see http://code.google.com/p/sreadahead/issues/detail?id=5 okay, so it doesn't. first of all, it needs a directory /var/lib/sreadahead/debugfs to exist, but then still /etc/init.d/sreadahead fails with "unable to select tracer". apologies. /etc/init.d/sreadahead-pack fails with "unable to select tracer", and /etc/init.d/sreadahead fails with "device or resource busy". okay, so sreadahead-pack is no longer needed. in fact, it's not even included. mental note: read the changelog more carefully next time you spot a new version of something. anyhow, it seems like the tracing API changed quite a bit for 2.6.28, so further work upstream will be needed. that, or i am doing something awfully wrong. follow http://code.google.com/p/sreadahead/issues/detail?id=5 Thanks for your work here. I kinda want to wait until that bug is resolved. I'll check back (maybe this is why I didn't see any results when I initially worked on sreadahead) Upstream bug is still open =/ yeah, i'm subscribed to it and will report as soon as they have an update. not much seems to be moving. 2.6.29 is released, gentoo-sources will likely soon follow up. while the upstream still doesn't provide a patch for <=2.6.28, maybe this ebuild can be put in the tree anyway now? i haven't tested it yet, but i will as soon as gentoo-sources-2.6.29 is released. (In reply to comment #10) > 2.6.29 is released, gentoo-sources will likely soon follow up. while the > upstream still doesn't provide a patch for <=2.6.28, maybe this ebuild can be > put in the tree anyway now? i haven't tested it yet, but i will as soon as > gentoo-sources-2.6.29 is released. > There still is no patch for >=2.6.28, why would it work with 2.6.29? Maybe I am confused.. (In reply to comment #11) > > There still is no patch for >=2.6.28, why would it work with 2.6.29? Maybe I am > confused.. > parallellepipedum says in the upstream bug: > The patch currently only works on 2.6.29-rc1 and newer kernel versions. if i understood all this correctly, the patch doesn't work on <=2.6.28 Guess not. %% head Makefile VERSION = 2 PATCHLEVEL = 6 SUBLEVEL = 29 EXTRAVERSION = %% patch -p1 < sreadahead-manual.patch patching file fs/open.c patching file kernel/trace/Kconfig Hunk #1 succeeded at 150 (offset 31 lines). patching file kernel/trace/Makefile Hunk #1 succeeded at 20 (offset 5 lines). patching file kernel/trace/trace.h Hunk #1 FAILED at 22. 1 out of 1 hunk FAILED -- saving rejects to file kernel/trace/trace.h.rej i was able to use the 0001-kernel-trace-open.patch with gentoo-sources-2.6.29. it doesn't seem that sreadahead is completely working, yet, but at least the kernel patch applies. (In reply to comment #14) > i was able to use the 0001-kernel-trace-open.patch with gentoo-sources-2.6.29. > it doesn't seem that sreadahead is completely working, yet, but at least the > kernel patch applies. > ah, thx for the hint. So sreadahead-manual.patch is obsolete then? I committed it with the trace-open patch in /usr/share/... Thanks for your help here. We will leave this bug open for further discussion if needed. I have no idea if the uploaded patches should be used, and the upstream developers don't seem to be making any progress either. One more thing: the init.d script should be modified. sreadahead-pack is no longer needed, nor is the check for existence of /etc/readahead.pack (sreadahead automatically rebuilds that file if it doesn't exist, and it actually uses /var/lib/sreadahead/pack instead (which reminds me of the fact that the ebuild needs to be updated to create /var/lib/sreadahead/debugfs/)). (In reply to comment #16) > I have no idea if the uploaded patches should be used, and the upstream > developers don't seem to be making any progress either. You are right, I masked it because we really don't know what is going on ;) > > One more thing: the init.d script should be modified. sreadahead-pack is no > longer needed, nor is the check for existence of /etc/readahead.pack > (sreadahead automatically rebuilds that file if it doesn't exist, and it > actually uses /var/lib/sreadahead/pack instead (which reminds me of the fact > that the ebuild needs to be updated to create /var/lib/sreadahead/debugfs/)). > fixed init script and missing dir in -r1 Created attachment 186608 [details, diff]
Patch for 2.6.29
In kernel version 2.6.29 iter_ctrl was renamed to trace_options. This patch will allow the "ftrace_printk" option to be enable on 2.6.29 (as well as earlier kernels).
http://code.google.com/p/sreadahead/issues/detail?id=5 Upstream is planning to release a patcheless sreadahead. I suggest we wait for that, instead of fiddling around with patches. (In reply to comment #19) > http://code.google.com/p/sreadahead/issues/detail?id=5 > > Upstream is planning to release a patcheless sreadahead. I suggest we wait for > that, instead of fiddling around with patches. > Good call, thx for the update. in the meantime, the project is sort of dead, are we going to keep waiting for the patchless version? (In reply to comment #21) > in the meantime, the project is sort of dead, are we going to keep waiting for > the patchless version? > You are right, it is kinda silly to keep around. I vote for removal, do you agree? I am definitely not an official developer, so I don't know the policy. I gotta agree, though, that the current package is sort of broken - at least it is when built againts the recent kernel versions, so I guess it is useless for most people. Please note that some guy made a patch for 2.6.30, but that's of course not an official patch. That said, this project needs a kick in the butt, and it's not getting any, so I assume it's dead. What a shame. Thanks, it is too hard for us to maintain some non-official patches for Gentoo. Too bad upstream failed at this project. IMO. @treecleaners: If anyone gets to it before I get internet at home again, please mask/last-rite this package. # Jeremy Olexa <darkside@gentoo.org> (4 Oct 2009) # Removed in 30 days. Dead project, needs kernel patches. bug 257207 sys-apps/sreadahead Removed from tree |
