Spec URL: http://domsch.com/linux/fedora/sreadahead/sreadahead.spec SRPM URL: http://domsch.com/linux/fedora/sreadahead/sreadahead-0.02-1.fc9.src.rpm Description: sreadahead speeds up booting by pre-reading disk blocks used during previous boots. Part of the 5-second bootup project.
I have improved the original makefile a bit to support using custom CFLAGS and installing, so the autotools patch can be dropped. You can see the results (patch, new spec file and new srpm) at http://fedora.danny.cz/sreadahead/
It doesn't matter to me if it uses autotools, or if it uses Dan's (nice and small) Makefile. I had sent the autotools patch to Arjan, but whatever upstream wishes to use is fine by me. Arjan, see the patch and spec Dan provides in a comment to this bug and decide for yourself. Thanks, Matt
when ever this comes in rawhide I volunteering to test it and provide feedback.
Few questions from fedora-devel that should be answered - https://www.redhat.com/archives/fedora-devel-list/2008-October/msg00273.html
on how to generate the file list: find / -type f > filelist.txt on the kernel patch; that's still in upstream discussion.
(In reply to comment #5) > on how to generate the file list: > find / -type f > filelist.txt shouldn't be better to explicitly define (somewhere in /etc/sysconfig/sreadahed) relevant directories (/bin, /sbin, /usr, ..)? "find /" could be real overkill for people with large full disks (I guess we needn't porn pictures and mp3 files during the boot... :-)
any progress on the kernel patch?
It's an entirely different kernel patch now.
This one: http://lkml.org/lkml/2009/1/27/297 together with sreadahead-1.0 http://lkml.org/lkml/2009/1/27/296
http://kyle.fedorapeople.org/sreadahead/ |`- http://kyle.fedorapeople.org/sreadahead/sreadahead.spec `-- http://kyle.fedorapeople.org/sreadahead/sreadahead-1.0-1.fc10.src.rpm Please let me know if I made a hash of the specfile. I'd like us to get this and the hooks for it into F11. I'm not willing to merge the patch for the open tracing yet, as we would effectively be "guaranteeing" a userspace interface (not that debugfs really provides these assurances.) and I'd like to wait and see what actually gets merged upstream...
Ping? Is this review stuck?
It still depends on non-upstream kernel changes.
I've ported this to use the utrace-ftrace support we currently have in Fedora... Is that acceptable for rawhide/F-12? I'll submit a patched srpm when I've done a bit more testing to make it somewhat more bulletproof. Thanks! Kyle
(In reply to comment #13) > I've ported this to use the utrace-ftrace support we currently have in > Fedora... Is that acceptable for rawhide/F-12? If that means that instead of relying on kernel patch for the proper function, you have reworked readahead itself, then it is fine. > I'll submit a patched srpm when I've done a bit more testing to make it > somewhat more bulletproof. OK
(In reply to comment #13) > I've ported this to use the utrace-ftrace support we currently have in > Fedora... Is that acceptable for rawhide/F-12? Is patch is available somewhere or this is an abandoned piece of software?
Ping? Any progress here? Or we can close this review?
I've got no interest anymore. I think this was a good idea 5 years ago that never really came to fruition. Closing - if I'm wrong, someone else can re-open and take it. Thanks, Matt
This actually got folded into systemd, so as a separate item it is obsolete.