[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-block] [PATCH] Report error when opening device with locked tr
Re: [Qemu-block] [PATCH] Report error when opening device with locked tray
Tue, 7 Jun 2016 16:00:37 -0400
Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.0
On 06/07/2016 06:28 AM, Kevin Wolf wrote:
> Am 06.06.2016 um 21:40 hat Colin Lord geschrieben:
>> This commit causes qmp_blockdev_change_medium to report an error if an
>> attempt is made to open a device with a locked tray.
> The old behaviour is that the command seemingly succeeds, but the medium
> isn't actually changed. Correct?
Close. Old "change" command also fails, but with a confusing error.
> Should this be mentioned in the commit message? You just describe what
> you change, but not why.
- Change uses qmp_blockdev_open_tray, which "succeeds."
- Change then tries to use qmp_x_blockdev_remove_medium, but receives
potentially confusing error "Tray is locked."
- Moments later, the tray is likely now open.
- Change uses do_open_tray, which returns -EINPROGRESS.
- Change can propagate this error upwards without attempting to remove
- User gets "Device <foo> is locked and force was not specified, wait
for tray to open and try again" error.
Why: "The new error tries to inform the user that there is an action
pending and that the command, if run again, may succeed."
>> Signed-off-by: Colin Lord <address@hidden>
>> This is based off my previous patch regarding the do_open_tray function
>> (currently at v3). Probably should have been submitted as a patch set
>> but I wasn't thinking that far ahead when I submitted the first patch.
>> blockdev.c | 7 +++++--
>> 1 file changed, 5 insertions(+), 2 deletions(-)
> Yes, would probably have made sense as a series, but as long as it's
> only two patches, it's not really a problem.
> Please make sure to put such comments below the "---" line, though, i.e.
> comments that make sense for the review, but not as part of the commit
> log. Then git-am automatically removes that part from the commit message
> while applying the patch. I did it manually for this one now.