https://bugzilla.novell.com/show_bug.cgi?id=220839 olh@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED ------- Comment #1 from olh@novell.com 2006-11-16 04:23 MST ------- something for 10.3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=220839 olh@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |chuller@novell.com ------- Comment #2 from olh@novell.com 2006-11-24 10:23 MST ------- *** Bug 223454 has been marked as a duplicate of this bug. *** -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=220839 ------- Comment #4 from olh@novell.com 2007-03-01 07:32 MST ------- will be implemented this year. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=220839 agruen@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |kkaempf@novell.com Status|RESOLVED |REOPENED Resolution|WONTFIX | ------- Comment #6 from agruen@novell.com 2007-05-03 04:48 MST ------- Linux is well aware of the EFI partitions: all our kernels have CONFIG_EFI_PARTITION=y enabled. Otherwise, the kernel wouldn't be able to access those partitions, and I would surely have noticed that already ... It would be very nice if parted could preserve the backward-compatibility DOS partitions if at all possible. Failing that, we would need to recognize when the boot disk is GPT in yast, and choose a GPT aware boot loader -- but which would work well on this kind of hardware? If we don't address this problem, machines with OSX/Linux dual-boot setups will still fail in 10.3, which would be pretty bad. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=220839 ------- Comment #8 from olh@novell.com 2007-05-03 05:13 MST ------- zeroing the mbr smells like a bug. I talked to Alexander Graf today. He confirmed: The boot concept is still fucked up on intel systems. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c10 --- Comment #10 from Felix Möller 2007-07-21 02:48:06 MST --- (In reply to comment #6 from Andreas Gruenbacher)

If we don't address this problem, machines with OSX/Linux dual-boot setups will still fail in 10.3, which would be pretty bad.

(In reply to comment #4 from Olaf Hering)

will be implemented this year.

Olaf is there any chance for this to be implemented in 10.3? I would happily test it. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c12 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|Enhancement |Major --- Comment #12 from Ruediger Oertel 2007-07-31 09:43:41 MST --- well, just ran into the same thing during an update. before the update I had GPT and MBR nicely in sync during the update the MBR is zeroed and you end up with an error from the bootloader: grub: grub> setup --stage2=/boot/grub/stage2 (hd0,2) (hd0,2) Error 22: No such partition grub> quit so this behaviour does not only fail installation, it also breaks working setups. BTW: the dual scheme with MBR and another native partitioning scheme is also used on other platforms. In this case it's MacOS and GPT, on AXP/Alpha there was the BSD-disklabel together with MBR. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c14 Olaf Hering changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|olh@novell.com |fehr@novell.com Status|REOPENED |NEW --- Comment #14 from Olaf Hering 2007-07-31 10:37:12 MST --- Thomas, work with the community to get this change accepted. why was it assigned to me anyway? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c17 Andreas Gruenbacher changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |gp@novell.com --- Comment #17 from Andreas Gruenbacher 2007-07-31 11:05:11 MST --- I would expect us to have access to the EFI specification -- Gerald, do you know who could know about it? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c19 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|ro@novell.com | --- Comment #19 from Ruediger Oertel 2007-07-31 12:00:05 MST --- Created an attachment (id=154551) --> (https://bugzilla.novell.com/attachment.cgi?id=154551) yast2 logfiles tarball -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c21 --- Comment #21 from Ruediger Oertel 2007-07-31 12:05:17 MST --- BTW: refit.sf.net implements the sync between GPT and MBR as well, could maybe be used as reference. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c24 Thomas Fehr changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|fehr@novell.com |odabrunz@novell.com --- Comment #24 from Thomas Fehr 2007-08-01 03:56:56 MST --- There seem to be two different things mix in this patch: First, there seems to be demand for a new feature, namely supporting this strange "dual GPT/MSDOS" partition table. This is clearly a feature request and needs to be done via Fate and not bugzilla. There will be no progress in implementing this feature unless it is in fate and approved. Second, something messes up Rudis system during update. During update, yast2-storage does only calls "parted .... print" which certainly does not do any harm to the content of a partition. After looking through the log I find some thing strange, there are two partitions (sda1 and sda3) with the boot flag set. So far this was not excpected by libstorage and it claims all partitions on GPT as partitions of type "EFI boot". The wrong classification of sda3 as "EFI boot" probably leads bootloader to treat sda3 as boot partition instead of sda1. I fixed libstorage now not to mis-detect any partitions on GPT with boot flag set as EFI unless they also contain a fat filesystem. Reassigned to maintainer of bootloader to check if there are any other problems with bootloader and such a strange disk setup. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c26 --- Comment #26 from Thomas Fehr 2007-08-01 04:28:52 MST --- Then we need to find out which exact call destroys the DOS partition table. yast2-storage only calls "parted ... print" and this definitely does not touch the partition table. Bootloader does such things as "/usr/sbin/parted -s /dev/sda set 3 boot on". In theory, this only sets a flag in the gpt partition by maybe this also destroys synchronizity between DOS and GPT disklabel. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c29 --- Comment #29 from Thomas Fehr 2007-08-01 04:44:58 MST --- What is wrong? The detection of /dev/sda3 as partition type "EFI boot" was certainly incorrect and is fixed now. If this is not the reason for your breakage (as I suspected) we need to find out, what was and try to avoid that. Maybe simply calling "/usr/sbin/parted -s /dev/sda set 3 boot on" would be a first step to check of this is really the problem. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c33 Thomas Fehr changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|fehr@novell.com |odabrunz@novell.com --- Comment #33 from Thomas Fehr 2007-09-04 03:24:30 MST --- The parted call to set the boot flag is not done by the storage related code. Therefore I certainly cannot answer Rudis question. The parted call to set the boot flag is done in bootloader code and therefore the changes (regarding this call) need also to be done there: bootloader/grub/misc.ycp:382 Running command /usr/sbin/parted -s /dev/sda set 3 boot on Since this is too late for 10.3 anyway I reassign this back to Olaf. So he can come back to this until next release. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c35 Stefan Fent changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |stefan.fent@novell.com AssignedTo|stefan.fent@novell.com |fehr@novell.com --- Comment #35 from Stefan Fent 2007-10-24 13:26:26 MST --- This clearly is a parted bug - even if, like in this case, bootloader does not necessarily need to write anything to MBR, it may have to in other situations. reassigning -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839#c37 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ro@novell.com Status|RESOLVED |REOPENED Component|Basesystem |YaST2 Product|openSUSE 10.2 |openSUSE 11.0 Resolution|INVALID | Version|Beta 2 plus |unspecified --- Comment #37 from Ruediger Oertel 2007-11-06 04:14:57 MST --- reopening for 11.0 I'll open a fate request for the parted patch, but this bug needs to be reassigned to the new maintainer of yast2-bootloader, please see comment #31: the bootloader should not blindly call parted to set flags that are already set perfectly. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839 Arvin Schnell changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|fehr@novell.com |dfiser@novell.com Status|REOPENED |NEW -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839 Hannes Reinecke changed: What |Removed |Added ---------------------------------------------------------------------------- Attachment #154541|text/x-patch |text/plain mime type| | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839 User agraf@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=220839#c41 --- Comment #41 from Alexander Graf 2008-07-17 17:02:12 MDT --- Created an attachment (id=228615) --> (https://bugzilla.novell.com/attachment.cgi?id=228615) Parted patch to sync GPT->MBR All right, new approach so we need fewer changes to the rest of the toolchain but still improve the situation compared to erasing the MBR: Always sync the MBR to the first GPT partitions. This also enables us to use partitions > 2TB with BIOSes, as long as the user has /boot within the first 4 partitions and within 2TB (all of which restrictions we had before too). Could anyone try this patch please and see if it works properly on a Mac? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=220839 User agraf@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=220839#c43 Alexander Graf changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #43 from Alexander Graf 2008-07-18 02:25:58 MDT --- The patch is on its way to STABLE ... Please reopen this bug if you're still having trouble booting from GPT with BIOS. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.