[Bug 934640] New: YaST2 crashes in installer on ASUS X205TA
http://bugzilla.suse.com/show_bug.cgi?id=934640 Bug ID: 934640 Summary: YaST2 crashes in installer on ASUS X205TA Classification: openSUSE Product: openSUSE Factory Version: 201505* Hardware: x86-64 OS: Other Status: NEW Severity: Critical Priority: P5 - None Component: YaST2 Assignee: yast2-maintainers@suse.de Reporter: kukuk@suse.com QA Contact: jsrain@suse.com Found By: --- Blocker: --- Created attachment 637740 --> http://bugzilla.suse.com/attachment.cgi?id=637740&action=edit yast2-logs.tgz I have a netbook with a eMMC disk. YaST2 installer crashes during installation when creating the installation proposal. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c1
Arvin Schnell
http://bugzilla.suse.com/show_bug.cgi?id=934640
Arvin Schnell
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c2
--- Comment #2 from Thorsten Kukuk
The YaST log shows that parted reports IO errors for the device"
2015-06-13 14:45:37 <2> (none)(2859) [libstorage] CmdParted.cc(probe):58 THROW: parted complains: Warning: Error fsyncing/closing /dev/mmcblk0rpmb: Input/output error: "/usr/sbin/parted -s '/dev/mmcblk0rpmb' unit cyl print unit s print"
/dev/mmcblk0rpmb is "Replay Protected Memory Block - eMMC" It is a specific memory area which is in an authenticated and replay protected manner. You shouldn't touch that as long as we don't use that for DRM Content protection keys or EM provided/Specific security Keys, for example. I think you get this errors since we didn't authenticate, but no idea how this would work. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c3
--- Comment #3 from Stefan Hundhammer
2015-06-13 14:45:37 <2> (none)(2859) [libstorage] CmdParted.cc(probe):58 THROW: parted complains: Warning: Error fsyncing/closing /dev/mmcblk0rpmb: Input/output error: "/usr/sbin/parted -s '/dev/mmcblk0rpmb' unit cyl print unit s print"
/dev/mmcblk0rpmb is "Replay Protected Memory Block - eMMC" It is a specific memory area which is in an authenticated and replay protected manner.
You shouldn't touch that as long as we don't use that for DRM Content protection keys or EM provided/Specific security Keys, for example.
I think you get this errors since we didn't authenticate, but no idea how this would work.
AFAICS libstorage always accessed that device like this, probably always receiving that error. Until recently, however, it just silently ignored the error; there was simply no output on stdout to parse, so no partitions or other technical data were ever reported for that device. Now this is reported as an exception - which seems to be too strict error checking in this case. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c4
--- Comment #4 from Thorsten Kukuk
AFAICS libstorage always accessed that device like this, probably always receiving that error. Until recently, however, it just silently ignored the error; there was simply no output on stdout to parse, so no partitions or other technical data were ever reported for that device.
Up to now our distributions didn't contain kernel support for this kind of devices, thus this /dev/mmcblk0rpmb didn't exist. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c5
--- Comment #5 from Arvin Schnell
http://bugzilla.suse.com/show_bug.cgi?id=934640
Martin Pluskal
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c6
Stefan Hundhammer
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c7
--- Comment #7 from Thorsten Kukuk
Now setting disks to "read only" when probing with "parted" reports an I/O error.
I hope you don't set the whole disk to read-only if this single partition reports an I/o error? Because this would prevent us from installing on any eMMC device. Better is to filter devices with "/dev/*rpmb" and ignore them.
Notice that the expert partitioner might still not prevent users from trying to create a new disk label (ms-dos or GPT) on that device.
Which is absolute fine. In this case, the problem is not the disk itself, but the special "partition" /dev/mmcblk0rpmb -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c8
--- Comment #8 from Stefan Hundhammer
I hope you don't set the whole disk to read-only if this single partition reports an I/o error?
If this is a single partition, the kernel reports it wrong. libstorage would not try to probe it with that parted command if it had not been reported as a disk. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c9
--- Comment #9 from Thorsten Kukuk
If this is a single partition, the kernel reports it wrong. libstorage would not try to probe it with that parted command if it had not been reported as a disk.
I'm no expert on this, but to me it looks like this is not even a single partition, but only a special interface. However, all you need to make sure of is, that if /dev/mmcblk0rpmb reports I/O errors, that you don't set the whole disk (/dev/mmcblk0) read-only. I don't know about the libstorage interna, if you handle both devices as different disks or the same. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c10
Stefan Hundhammer
http://bugzilla.suse.com/show_bug.cgi?id=934640
http://bugzilla.suse.com/show_bug.cgi?id=934640#c11
--- Comment #11 from Stefan Hundhammer
However, all you need to make sure of is, that if /dev/mmcblk0rpmb reports I/O errors, that you don't set the whole disk (/dev/mmcblk0) read-only. I don't know about the libstorage interna, if you handle both devices as different disks or the same.
On that level, a partition cannot throw an I/O error. Only disks can. So we should be on the safe side. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply@novell.com