With the transition to the new kernel rpm versioning scheme, we ended up having to call things such as the pre-2.6.23-rc1 git kernels, which upstream are 2.6.22-gitX, 2.6.23-rc0-gitX. However, I missed a tiny detail, which leaves the kernel package showing 2.6.23, but the resulting kernel's uname 2.6.22. Makes Bad Things happen with any kernel in the 2.6.23-0.x.rc0.git1.fc8 range, where x is . Fixing now (in theory, this is a simple fix by twiddling the kernel Makefile's SUBLEVEL), will push a fixed build ASAP...
Jarod Wilson wrote:
With the transition to the new kernel rpm versioning scheme, we ended up having to call things such as the pre-2.6.23-rc1 git kernels, which upstream are 2.6.22-gitX, 2.6.23-rc0-gitX. However, I missed a tiny detail, which leaves the kernel package showing 2.6.23, but the resulting kernel's uname 2.6.22. Makes Bad Things happen with any kernel in the 2.6.23-0.x.rc0.git1.fc8 range, where x is . Fixing now (in theory, this is a simple fix by twiddling the kernel Makefile's SUBLEVEL), will push a fixed build ASAP...
Of course, Linus merged a ton of stuff, and now the latest git snapshots are currently not liking to actually compile, so we don't have a repaired kernel package out just yet. Sorry for the delay folks. The Man (davej) is working through the compile issues though, and hopes to have something relatively soon. Stay tuned... :)
Jarod Wilson wrote:
With the transition to the new kernel rpm versioning scheme, we ended =
up
having to call things such as the pre-2.6.23-rc1 git kernels, which upstream are 2.6.22-gitX, 2.6.23-rc0-gitX. However, I missed a tiny detail, which leaves the kernel package showing 2.6.23, but the resulting kernel's uname 2.6.22. Makes Bad Things happen with any kern=
el
in the 2.6.23-0.x.rc0.git1.fc8 range, where x is . Fixing now (in theory, this is a simple fix by twiddling the kernel Makefile's SUBLEVEL), will push a fixed build ASAP...
=20 Of course, Linus merged a ton of stuff, and now the latest git snapshot=
s
are currently not liking to actually compile, so we don't have a repaired kernel package out just yet. Sorry for the delay folks. The Ma=
n
(davej) is working through the compile issues though, and hopes to have=
something relatively soon. Stay tuned... :)
kernel 2.6.23-0.28.rc0.git6.fc8 is built and boot tested on a few systems. The uname mismatch is indeed fixed, though the hpet hang on some ich systems is still present (workaround by booting w/nohpet). Should hit rawhide trees tomorrow...
--=20 Jarod Wilson jwilson@redhat.com