Hello, It looks like due to a few ugly bugs with sparc assembly and a tg3.c typo, it's looking like we'll need a kernel-source-2.4.27-4 soon. Currently I'm not quite sure if tg3 will work with the firmware, due to a stray & that caused an invalid pointer type warning.
For the rest, this revision will not affect anyone but sparc. But I think I will wait for -3 to enter sarge first - it is not critical if -4 does not make sarge. However, if binutils 2.15-1 goes in, -4 will need to enter as well (the new patch, 065_copy_user_parens.diff, fixes the build against binutils 2.15-1 - see #266772.) So just so we can really make this upload count, I ask everyone to add anything that should be merged into kernel-source-2.4.27 into the SVN tree at /kernel/kernel-2.4/source/kernel-source-2.4.27-2.4.27/debian/patches and add it to the debian/patches/series/2.4.27-4 file. I'm quite tired of uploading kernel-source-2.4.27, which is why i'm making this special request. :) So please test your shiny new kernels and make 2.4.27-4 the best and hopefully last kernel-source-2.4.27 upload before sarge.. A brief note on how the series files work. They define what patches to add or subtract for a given release. The single downside of this is that you may not reap patches from debian/patches, because they are needed to unpatch existing trees of older versions. The feature described here is preserving functionality from Herbert's old monolithic patches. So, for example, Norbert told me the 018_core_cia_1G.diff in -1 and -2 was bogus, so series/2.4.27-3 only (unfortunately) contains: - 018_core_cia_1G.diff. series/2.4.27-4 currently contains: + 065_copy_user_parens.diff + 066_dep_fw_loader.diff + 067_tg3_ampersand.diff This allows for a monolithic grouping of patches for a release, without the need to lump them all together. Cool, isn't it? Also better than dpatch in that you don't need to source some file providing the boilerplate functions.. -- Joshua Kwan
signature.asc
Description: Digital signature
