This bug is for tracking regressions in portage packages (except kernel sources themselves) due to 2.6.22 upgrades. Please don't mark bugs as fixed until the fix is in the stable tree. Assuming everything goes to plan, we'll be pushing for 2.6.22 stable on 30th October.
This bug is about 2.6.23, right? (not 2.6.22, as the text says)
yes it's about 2.6.23, sorry. I somehow manage to screw up that text every time...
There's a gcc bug, which might prevent kernel 2.6.23 from going stable: Bug 135745 It causes sandbox violations when compiling external modules.
As mentioned on the other bug, that's a bug in gcc, sandbox, or some other low level *userspace* component. If you are only seeing it now rather than later then it means you recently upgraded gcc or something like that. It's totally unrelated to kernel version, and not a kernel bug, so does not block 2.6.23 going stable.
To clarify, I meant "now rather than earlier" and am simply trying to say that you get this bug with ALL kernel versions (yes! even 2.6.22, the current stable!) when you have the affected choice of compiler/userspace-stuff installed. Please don't reply to this comment, it's offtopic for this bug.
How about this bug: https://bugs.gentoo.org/show_bug.cgi?id=195381 Does it count as a blocker?
Here's another one, this time related to udev-114: https://bugs.gentoo.org/show_bug.cgi?id=195384
(In reply to comment #4) > It's totally unrelated to kernel version, and not a kernel bug, so does not > block 2.6.23 going stable. No. The problem comes and goes when swiching the kernel version. If i set my /usr/src/linux symlink to a 2.6.22 kernel, compiling modules works. If i set my /usr/src/linux symling to a 2.6.23 kernel, compiling modules does not work. I assume, they really changed something in the buildsystem of the kernel. (Maybe they call gcc without -pipe or something like it? I have no clue, what these gcda files are for)
(In reply to comment #8) > (In reply to comment #4) > > No. The problem comes and goes when swiching the kernel version. > > If i set my /usr/src/linux symlink to a 2.6.22 kernel, compiling modules works. > If i set my /usr/src/linux symling to a 2.6.23 kernel, compiling modules does > not work. > > > I assume, they really changed something in the buildsystem of the kernel. > (Maybe they call gcc without -pipe or something like it? I have no clue, what > these gcda files are for) > The same happens to me. If I switch back the symlink to pinto to linux-2.6.22-gentoo-r8 the nvidia-drivers package builds fine. But if the symlink points to linux-2.6.23-gentoo i have the access violation error.
Please stop discussing bug #135745 on this bug, and instead discuss it on bug #135745 instead. Thanks! :)
I just added 196001 - only thing that changed from one toolchain/system/world rebuild to the next was linux-headers, which caused net-misc/iputils-20070202 to fail to build.
(In reply to comment #11) > I just added 196001 Very sorry to bugspam, I found #195861 with a fix for this problem only after filing this.
How's about this one: https://bugs.gentoo.org/show_bug.cgi?id=197264 NFS data corruption. That's relatively serious 8-).
This bug is for problems in packages outside of the kernel. Please don't pollute this with kernel problems...
Oops sorry 8-).
Closing as 2.6.23 is stable
