My issues with kernel development [LWN.net]

My issues with kernel development

Posted Jun 12, 2008 2:21 UTC (Thu) by pr1268 (subscriber, #24648)
Parent article: Andrew Morton on kernel development

First of all, I wish to thank Andrew for his thoughts and time in responding. Such discussion relating to kernel development is refreshing.

As for reporting bugs, I've two in particular with 2.6.25[.x] that I've been loath to report:

make oldconfig doesn't appear to work anywhere near what I expected. After doing so, I had a .config for a Summit subarch with drivers enabled for all sorts of devices I don't have in my 32-bit x86 PC.
Inserting a CD/DVD into either of my DVD-RW drives and subsequently issuing an appropriate mount(8) command had the shell gripe at me stating that no media was present. However, waiting 15-20 seconds between closing the drive door and issuing the mount command works fine. This behavior is different than 2.6.24 (and earlier).

Granted, I was hesitant to report either of these (until now) because I was unsure whether (1) was operator error, and whether (2) was expected new behavior given the patches submitted for 2.6.25. Plus, I didn't want to add yet another message to the (already crowded) LKML. But, I'm curious: would reporting either of these be appropriate?

I would certainly love to contribute to the kernel development project--I even subscribed to the LKML (having been inspired by our editor's eagerness to help with the Kill-The-BKL project)--but being a newbie, I could use a little guidance.

Thanks again to Andrew for his candor.

(Log in to post comments)

My issues with kernel development

Posted Jun 12, 2008 5:00 UTC (Thu) by grundler (subscriber, #23450) [Link]

Yes - Andrew definitely deserves the kudos he gets.

(1) isn't really bug since "make oldconfig" is "expected behavior". Try "make menuconfig" and
see if a menu driven config tool works better for you. Too often, I find the "Help" text
useless and I'm not a kernel newbie. Updating those to be meaningful (e.g. spelling out
uncommon acronyms) would help alot of people.

(2) is a regression and sounds like it's bisect'able. In fact, recent bug on linux-scsi sounds
similar to this though might not be the same:
    http://marc.info/?t=121229388800003&r=1&w=2

So reporting the problem to linux-ide and/or linux-scsi might be a good starting point. You
don't have to report problems to LKML since there are plenty of topic specific mailing lists
that have less traffic. See linux-2.6.25/MAINTAINERS file for the various mailing lists. If
you post to the wrong list, people generally will redirect you to the right one. As Andrew
suggests, be persistent.

Lastly, regarding "being a new newbie" try http://kernelnewbies.org which is one of many
starting points. Usually any help with documentation, code review, or testing is something
anyone with a computer can do - especially if you are finding bugs, willing to report them and
test out (likely bad) theories on the bug. This interaction will lead to learning lots of new
stuff.

Please report it in bugzilla

Posted Jun 12, 2008 6:41 UTC (Thu) by mingo (subscriber, #31122) [Link]

(1) isn't really bug since "make oldconfig" is "expected behavior". Try "make menuconfig" and see if a menu driven config tool works better for you. Too often, I find the "Help" text useless and I'm not a kernel newbie. Updating those to be meaningful (e.g. spelling out uncommon acronyms) would help alot of people.

It is a serious upstream kernel regression if "make oldconfig" (used on a .config that worked with a previous version of the kernel) suddenly breaks a working setup. Please report it if you get hit by such a bug/regression and it will be fixed.

We'd be shooting ourselves in the foot if we made it harder to test new kernels.

Ditto for the second bug - if mounting CDs worked well before and it suddenly starts producing spurious "no media" mount failures that's a plain bug/regression.

Please report them on bugzilla.kernel.org.

Also, if you test new kernels, make sure you run the kerneloops client which automatically reports crashes to kerneloops.org.

Please report it in bugzilla

Posted Jun 12, 2008 7:51 UTC (Thu) by pr1268 (subscriber, #24648) [Link]

Thank you both for the replies.

I'm beginning to wonder if the two issues I have with 2.6.25[.x] are related in a weird way:

make oldconfig wasn't working right, so I had to make menuconfig and manually select the kernel options. I may have [un]selected a bunch of options I [don'] use for my hardware, thus...
...the CD/DVD drives may have been acting flaky due to various IDE/SCSI/SATA drivers [not] in the kernel, or functioning differently based on my userspace configuration (I was running Slackware 12.0 when I tried 2.6.25.4 and noted these anomalies) and the odd config used when building the kernel.

Another reason why I was loath to report the CD/DVD not mounting issue was because I have some unusual IDE/SATA hardware in my system (a Promise PDC20271 ATA controller card, a Silicon Image SATA controller card, one of the two DVD burners is IDE whilst the other is SATA), but Linux has ordinarily given me no grief whatsoever for running this odd configuration (I also have a mix of IDE and SATA hard drives and a software RAID-0, but that's another story). Again, I must stress that this could all be a silly case of operator error (I'm good at finding these kinds of bugs ;-) ), or maybe it is a defect that needs the attention of the kernel developers...

I will admit that I'm somewhat of an informal kernel tester; I've compiled and run recent (-stable) kernels for the past 3 1/2 years now (thus explaining why I like Slackware--it works well with vanilla kernels), and I've only had to report one show-stopper (Oops in 2.6.15 due to NULL dereference in usbhid.c).

Thanks again for your replies; I'll look into reporting the make oldconfig issue.

Please report it in bugzilla

Posted Jun 12, 2008 14:25 UTC (Thu) by iabervon (subscriber, #722) [Link]

It's possible to have user error with "make oldconfig" on the first try (like getting the
wrong config into it), but if you can reproduce it, it's worth reporting. (And if you can't
reproduce it, you'll have a correct config...)

There was someone recently reporting problems with mounting optical media if he waited more
than 30 seconds after inserting it. It might be related, or it might be a coincidence, but you
might want to look into http://lkml.org/lkml/2008/6/6/170. The thread is kind of inconclusive,
but you might be able to help if you've got a different failure pattern (you need to wait,
while other people need to hurry), but also have a problem with timing and optical media
insertion that came up between 2.6.24 and 2.6.25. It's got things to try, anyway.

Make oldconfig issue has been reported in Bugzilla

Posted Jun 12, 2008 15:23 UTC (Thu) by pr1268 (subscriber, #24648) [Link]

I opened a bugzilla bug (#10898) on the make oldconfig issue. Apparently this is a regression reported by Linus himself, and a patch is in the works. Make oldconfig worked fine for 2.6.24.x (as I mentioned in the bugzilla description).

As for the mount(8) CD/DVD issue, well, I'll test that later this afternoon... Time to go to work... I'm still not discounting the possibility that a funky config kernel build combined with my strange mix of hardware (see above--yes, that's all one PC!) might have caused this anomaly.