Re: need (pmac owner help) with powerpc package.
On Mon, Dec 22, 2003 at 01:43:10PM +0100, Michael Schmitz wrote:
> > > could you please tell us which bugs actually need help? I did not find
> > > any bugs where there are unconfirmed problems. The older two packages
> > > have no open bug.
> > I believe many of those are upstream issues or so. The older packages
> > had no problems, since i asked for their removal, and reassigned all
> > their packages to 2.4.22. The bugs are at :
> > http://bugs.debian.org/cgi-bin/pkgreport.cgi?src=kernel-patch-2.4.22-powerpc
> Please specify which bugs you need help with; the majority of the bugs
> seems to be against long obsolete kernel versions (I'd just close those
> where you didn't have a confirmation that the bug still exists as of
> 2.4.22). More below.
Well, i need help with all those, i don't have the hardware to test
this, and often no clear grasp of what is at issue, not always though.
The immediate help i need is in a classifying effort of those bugs, to
know which i can close without effort, and which i need to look into. I
doubt that i will get reply from those bugs open since more than a year,
even though the problem does still remain, so i don't think simply
closing them is the right solution.
> > The issues i see are mostly of the kind : scsi driver not working,
> The aic7xxx and NCR drivers I'd build as modules (the newpmac models all
> have IDE boot disks) and close the 'builtin SCSI driver breaks boot' bugs.
> > probably worth merging, and checking with a newer kernel if this is
> > still the case, the brightness issue we talked here some time ago, and
> > the various xfree86, atyfb and m3mirror issues.
> > These are often stuff coming from the 2.2.x kernels i never maintained,
> > and maybe from the potato time.
> Again, please provide a list of bug numbers (and the type of hardware you
> need testing on). From a look at the bug titles I could see nothing
> related to atyfb (which is what I personally could test).
> Bugs I do have an opinion on:
> The m3mirror bug (#217778) needs a Pismo powerbook to test (IIRC BenH had
> one, and perhaps one of the Michel-s).
> The 'xfree crashes on wallstreet with 2.2 kernel' (#112867) definitely got
> fixed in 2.4 kernels (I backported the PCI resource re-allocation from 2.4),
> you can close that one for 2.4 for sure, or mark it wontfix.
Ok, thanks, will close with a 'we believe that this is fixed in 2.4.22'
> The 'mac-fdisk hangs on writing partition table' (#137575) I'd close for
> the same reasons: I could never reproduce the bug, plus I specifically
> changed mac-fdisk to replace sync with a device specific flush.
> The nano lockup (#119081) got cured by 2.4.12, you can close it (unless we
> still have 2.2 kernels in the distribution, in that case mark it as
> wontfix, perhaps). Eh, you tried to close it but that wouldn't stick for
> some reason?
Well, there are 2.2.x kernels for woody, but those got removed from
sarge/sid, or soon will be. I closed it already, but i think i remember
it was reopened, not sure though, couldn't find the mail about it in my
mail archive. I will close it, and a new bug can be opened. This one is
only gathering spam ATM anyway.
> The HFS cdrom bug (#149639) was a transient 2.4 problem IIRC, should be
> fixed now. At least I haven't had any more problems. Should be closed.
> The cube bug (#122228) looks like a 'hardware too new for kernel 2.4.16'
> type problem. That's two years back; Ben's kernel worked, so I take it for
> granted it works fine today. Should be closed.
> That's all I could see from the BTS logs, and you'd have made my task
> easier by listing the bug numbers including short descriptions. Maybe next
> time ??
Ok. I will already close those, and it will clear somewhat the bug page.
> Regarding the upload stall - if it's not just the 'package needs manual
> override, and ftp-masters are notioriously busy cleaning up the rootkit
> mess' reason, please tell me more (by PM).
Well, it is mostly that, + James Troup had reservations about the
multiplications of packages, but he never told me personnally, just some
rumor i got trough Joey Hess, so i don't really know.
I think the main point here is that it really sucks having
uncomunicative ftp-masters, not that they are busy right now. I have
some separate strategies i could do in case, but i can do nothing unless
the ftp-masters start working on the NEW queue again, and give me
feedback about the package.
I could for example, only ship the vmlinux kernel, and generate the
zImage.(chrp|chrp-rs6k|prep) ni a postinst using the stuff in
kernel-build or something such, same as how i intent to generate
builtin initrds. Don't think that is optimal though, and the way
kernel-image handles bootloader issues is also not nice (you need to
choose between yaboot, quik or nothing right now, but this is done per
subarch, which doesn't make sense, since the newworld and oldworld pmac
share the same subarch).
> Comment: several of these bugs should _not_ have been reassigned to 2.4
> kernels, as they do not happen there (as was obvious from the BTS
> messages). Trigger-happy QA ....
Well, i agreed to have all of them reassigned to 2.4 when the 2.2.x
kernels where removed from sarge/sid, since i needed to look over them
before i closed them, which i never really had the time to do, and which
is why i am asking help here.
Anyway, thanks for your help on this.