On 27 May 2008, at 9:56 am, dann frazier wrote:
etch's aboot doesn't permit booting recent 2.6 kernels. While backporting a fix for this issue, I ran into a couple of other build issues that prevent etch's aboot from being rebuilt properly. The proposed patch for this update and more detailed explanations are available here: http://lists.debian.org/debian-alpha/2008/05/msg00026.html This package is currently orphaned, but the previous maintainer did comment: http://lists.debian.org/debian-alpha/2008/05/msg00027.htmlOther than the testing I mentioned above, I've also had successful testreports from Bdale and Tim Cutts (on albeniz.d.o, iirc). This is obviously a requirement for alpha support in etchnhalf since 2.6.24 adds the ELF header incompatability. (Though releasing w/o alpha support is an option as well).
2.6.24 has been a right royal PITA to get working on albeniz. Once the aboot stuff was sorted (thanks Dann and Steve for your help with that), I found that it couldn't find its root filesystem because of a race condition in SCSI enumeration and software RAID, which I solved by adding a manual 10 second rootdelay parameter to the kernel command line.
Once I was past that, I was bitten by the removal of large numbers of drivers from 2.6.24 because of firmware licence issues (specifically, the acenic driver, which is needed for this machine's GigE NIC). I'm currently rebuilding the kernel package with that driver re-enabled, which should get albeniz.d.o back for porter use.
Tim --The Wellcome Trust Sanger Institute is operated by Genome Research Limited, a charity registered in England with number 1021457 and a company registered in England with number 2742969, whose registered office is 215 Euston Road, London, NW1 2BE.