> However, could you explain what failure(s) are caused by the current patch, > and steps to recreate? > (We'll need this info if we want to do an SRU for this issue.)
Sorry, I didn't wait for failure to happen - I reviewed the changes in Feisty and noticed hard coded values, which made me curious. I was able to track down the history of __syscall_return in the kernel mailing lists up to the point where MAX_ERRNO was introduced to __syscall_return oin include/asm/unistd.h (upon release of 2.6.19). As I see it, all kind of strange behaviour might happen - or the driver might just work anyway (because the higher error returns are never triggered). > In any case, testable debs for the new fglrx for l-r-m-2.6.22 are posted, if > you're interested: > http://people.ubuntu.com/~bryce/Testing/ Cool. However, 2.6.22 doesn't work too well in Feisty (which I'm using) - e.g. udev fails (iirc). But I always keep up with ATI's fglrx releases. As 8.37.6 doesn't need any changes to work with 2.6.22 everything should be fine (or better "as good as it gets" with ATI's horrible driver) for l-r-m-2.6.22 as well. How are you going to proceed with Feisty? Thinkwiki.org links a patch for 8.34.8 (http://www.thinkwiki.org/wiki/Problems_with_fglrx#fglrx_8.34.8) which looks OK to me. But be cautious - their 8.35.5 patch has the same bugs as your patch. Alternatively, you could introduce 8.37.6 in Feisty. -- firegl_public.c: Ubuntu modification uses obsolete __syscall_return https://bugs.launchpad.net/bugs/101984 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
