Bug 1072294
| Summary: | should disable host_cdrom device pass-through via scsi-generic with correct warning message | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Sibiao Luo <sluo> |
| Component: | qemu-kvm | Assignee: | Markus Armbruster <armbru> |
| Status: | CLOSED NOTABUG | QA Contact: | Virtualization Bugs <virt-bugs> |
| Severity: | high | Docs Contact: | |
| Priority: | high | ||
| Version: | 7.0 | CC: | acathrow, armbru, chayang, hhuang, juzhang, michen, pbonzini, qzhang, virt-maint, xfu |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2014-06-24 13:05:14 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Sibiao Luo
2014-03-04 10:24:35 UTC
(In reply to Sibiao Luo from comment #0) > 2.pass-through host_cdrom device via scsi-generic. > e.g:...-drive > file=/dev/sg1,if=none,id=drive-data-disk,format=raw,media=cdrom,cache=none, > aio=native,werror=stop,rerror=stop -device > virtio-scsi-pci,id=scsi1,indirect_desc=off,event_idx=off,bus=pci.0,addr=0x8 > -device scsi-generic,drive=drive-data-disk,id=data-disk > Warning: option deprecated, use lost_tick_policy property of kvm-pit instead. > qemu-kvm: -drive > file=/dev/sg1,if=none,id=drive-data-disk,format=raw,media=cdrom,cache=none, > aio=native,werror=stop,rerror=stop: could not open disk image /dev/sg1: > Could not open file: Invalid argument Hmm, according bug 920096#c1, I should not specify cache=none for scsi-generic, i also tried the cache=writeback/writethrough which can boot up guest successfully and we can see the passthrough host_cdrom in guest correctly. # sg_map /dev/sg0 /dev/sda /dev/sg1 /dev/sr0 e.g:...-drive file=/dev/sg1,if=none,id=drive-data-disk,format=raw,*cache=writeback*,aio=native -device virtio-scsi-pci,id=scsi1,bus=pci.0,addr=0x7 -device scsi-generic,drive=drive-data-disk,bus=scsi1.0,id=data-disk Expected results: QEMU will quit with warning message prompt: Driver 'host_cdrom' is not whitelisted. Base on above, i will raise the priority for it, thanks. Best Regards, sluo We disabled host CD-ROM passthrough via block backend host_cdrom (bug 760885). Therefore, you can't back a scsi-cd or ide-cd device model with a host CD-ROM device in RHEL-7. We didn't disable SCSI host device passthrough via device model scsi-generic. Therefore, you *can* back a scsi-generic device model with a host SCSI CD-ROM device in RHEL-7. Maybe we should've disabled that kind of host CD-ROM passthrough as well, maybe not. Question is moot, as 7.0 shipped with it enabled. Closing NOTABUG. Please reopen if you think we should disable use of host SCSI CD-ROM via scsi-generic even though we shipped it in 7.0. (In reply to Markus Armbruster from comment #2) > We disabled host CD-ROM passthrough via block backend host_cdrom (bug > 760885). Therefore, you can't back a scsi-cd or ide-cd device model > with a host CD-ROM device in RHEL-7. > > We didn't disable SCSI host device passthrough via device model > scsi-generic. Therefore, you *can* back a scsi-generic device model > with a host SCSI CD-ROM device in RHEL-7. > > Maybe we should've disabled that kind of host CD-ROM passthrough as > well, maybe not. Question is moot, as 7.0 shipped with it enabled. > > Closing NOTABUG. Please reopen if you think we should disable use of > host SCSI CD-ROM via scsi-generic even though we shipped it in 7.0. I think we should disable pass-through the host CD-ROM via scsi-generic as well, or we should implement it to work well, thanks. Best Regards, sluo Comment#1 makes me believe scsi-generic backed by host CD-ROM works: "can boot up guest successfully and we can see the passthrough host_cdrom in guest correctly". If that's not the case, please describe how exactly it doesn't work. (In reply to Markus Armbruster from comment #4) > Comment#1 makes me believe scsi-generic backed by host CD-ROM works: "can > boot up guest successfully and we can see the passthrough host_cdrom in > guest correctly". If that's not the case, please describe how exactly it > doesn't work. Yes, thanks for your reminds, I just forgot the comment #1, it worked well indeed. (In reply to Markus Armbruster from comment #2) > We disabled host CD-ROM passthrough via block backend host_cdrom (bug > 760885). Therefore, you can't back a scsi-cd or ide-cd device model > with a host CD-ROM device in RHEL-7. > > We didn't disable SCSI host device passthrough via device model > scsi-generic. Therefore, you *can* back a scsi-generic device model > with a host SCSI CD-ROM device in RHEL-7. > > Maybe we should've disabled that kind of host CD-ROM passthrough as > well, maybe not. Question is moot, as 7.0 shipped with it enabled. Hi Markus, Thanks for your inputs first. Back to QE POV, do we need to test scsi-generic & CD-ROM passthrough scenario in RHEL7.1? Do not disable scsi-generic & CD-ROM passthrough is by design? or just ignore it? Look forward to your feedback. Best Regards, Junyi > > Closing NOTABUG. Please reopen if you think we should disable use of > host SCSI CD-ROM via scsi-generic even though we shipped it in 7.0. SCSI host device passthrough (device models scsi-generic and scsi-block) is an intentional RHEL-7 feature. Evidence: bug 721250. General rule for RHEL-7: qemu-kvm can do it, libvirt can configure it, we support it. Exceptions are possible, but should be documented. Regarding whether we need to test this: excellent point! I figure our testing of SCSI host device passthrough should cover CD-ROM devices, too. (In reply to Markus Armbruster from comment #7) > SCSI host device passthrough (device models scsi-generic and scsi-block) is > an intentional RHEL-7 feature. Evidence: bug 721250. > > General rule for RHEL-7: qemu-kvm can do it, libvirt can configure it, we > support it. Exceptions are possible, but should be documented. > > Regarding whether we need to test this: excellent point! I figure our > testing of SCSI host device passthrough should cover CD-ROM devices, too. Thanks Markus. Best Regards, Junyi |