Created attachment 606831 [details] boot error Description of problem: When F18 Alpha TC3 x86_64 netinst is booted on a Thinkpad T420s in UEFI mode, the following message is displayed: > Secure boot not enabled > error: failure reading sector 0x40 from `cd0'. > > Press any key to continue... The message is not always the same, sometimes there was `hd1' instead of `cd0', or something similar. The key-press is not really needed, after some time (15-30 seconds?) it continues to boot automatically. It ends up with kernel panic, but that might be a different bug. When the same ISO is converted to USB stick using livecd-iso-to-disk, it still prints "Secure boot not enabled", but it doesn't print the error and it doesn't ask for pressing a key, it continues to boot right away. USB boot also doesn't show the traceback, but that's probably influenced by bug 848641 which ends the boot prematurely. Version-Release number of selected component (if applicable): F18 Alpha TC3 x86_64 netinst Thinkpad T420s http://www.smolts.org/client/show/pub_a0f8be89-b530-4af5-a50a-e1d06c11b3cb How reproducible: always Additional info: Jan Sedlak is the owner of the machine, adding to CC. We weren't able to verify this problem on a different UEFI machine, because we don't have another UEFI machine with optical drive.
Created attachment 606832 [details] kernel panic during boot
The error message doesn't directly prevent installer boot (even though it present a good user experience), the kernel panic does. If this is deemed to be two different issues, I'll split the bug and propose the error message as an Alpha-NTH blocker and the kernel panic as an Alpha blocker. In the meantime, marking the whole bug as an Alpha blocker.
The kernel doesn't print out "press any key to continue..." on reading CD errors. That is coming from grubx64.efi.
Same problem was discussed on http://lists.fedoraproject.org/pipermail/devel/2012-July/170434.html . Mjg gives some hints but apparently no solution.
(In reply to comment #4) > Same problem was discussed on > http://lists.fedoraproject.org/pipermail/devel/2012-July/170434.html . Mjg > gives some hints but apparently no solution. Actually that is a different bug - bug 849389. This one is about the error message, that one is about boot hang.
(In reply to comment #5) The mail report also had sector 0x40 issues and thus seems somehow related. It indicates that we probably can't blame it on a defect media. It also hints that it is an "old" problem, so when it haven't been fixed or reported widely yet then it must be something that only shows up under certain circomstances.
Discussed at 2012-08-27 QA meeting, functioning as a blocker review meeting. Accepted as NTH: our understanding of the impact of this bug is that it affects Intel AHCI-based systems when installing UEFI from slow optical drives. That's a small enough group to be NTH rather than blocker. (Booting from USB is considered an acceptable workaround).
It's believed that https://admin.fedoraproject.org/updates/kernel-3.6.0-0.rc2.git2.1.fc18,grub2-2.00-5.fc18,pesign-0.10-4.fc18 should have fixed this. That's been pushed stable but isn't in a compose yet. Next compose ought to confirm whether it's truly resolved.
Unfortunately Jan and his computer are away for the whole week. Does somebody else have a UEFI machine with a DVD drive and can test this? Thanks.
I can try.
unfortunately I couldn't reproduce the bug; tc4 booted fine via UEFI from DVD for me. tc5 did too, but that doesn't prove anything regarding this bug.
I can confirm that it's fixed in Alpha TC6. Neither "failure reading sector" nor kernel panic was displayed during boot. Since patches are in stable, I'm closing this bug.