On Thu, 2012-11-29 at 12:36 +0100, Bas Wijnen wrote: > tags 693663 + confirmed > thanks > > Hello Ben, > > Thank you for your work to document all these issues. If there would > be an alternative, I would not hesitate to get rid of this module. > Unfortunately, there isn't.
Thanks for accepting this despite my provocative summary. > This module was originally written for Linux 2.6.8, including (almost) > all the bugs, by the manufacturer of the hardware. I adjusted it so it > uses dkms and can compile on newer kernels. I noticed the bad coding, > but didn't start fixing all the bugs so far. > > If I would give priority to this package, I would try to rewrite it so > that it conforms to v4l2. If you're not already aware of it then see the series beginning at <http://lwn.net/Articles/203924/>. Of course the kernel internal API for V4L2 has changed since then; I'm not sure how much. > As it is now, the only reasonable (and > documented) way to use it is through a non-free library provided by > the hardware producer. Now that I think about it, this means it really > belongs in contrib. I think you're right. > I am using this module, and I (manually) make sure that I don't have > any races. I agree that this should be done by technical means. Also, > I think I see some of the other problems you describe (in particular > the memory leakage). I'm interested to get this fixed, but not so > interested that I've started doing it already. > > Summarizing: > - I agree that this module should not be in a stable release. > - I think it does deserve to be in unstable, because people with this > hardware want to use it on Debian; this support is better than no support. > - I very much invite people to help and fix the issues. This would be difficult for someone who doesn't have the hardware and software. > - I'll make some changes to the description to make clear that the > module is buggy. [...] Thanks. Ben. -- Ben Hutchings Never attribute to conspiracy what can adequately be explained by stupidity.
signature.asc
Description: This is a digitally signed message part
