Hi Tejun, Kay,
I'm investigating a customer report which manifests itself all the way
up in gnome-session when a BMC hotplug-adds a simulated DVD device. The
user logs into their server's BMC and enables "media redirection", an
emulated DVD device + .iso is dynamically added to the bus... in the
past this has worked well, however, they are now noticing a timing
condition on RHEL7 that prevents gnome from successfully auto-mounting
the DVD media.
With Harald's help, I've done some debugging and we've found out that on
hotplug-add, the kernel sends two uevents (ADD, CHANGE) in short
succession.
(Example with an ordinary, physical USB DVD device on my laptop, is
very similar):
% udevadm monitor -k -e
KERNEL[2409061.130338] add
/devices/pci0000:00/0000:00:14.0/usb3/3-9/3-9.3/3-9.3:1.0/host20/target20:0:0/20:0:0:0/block/sr1
(block)
ACTION=add
DEVNAME=/dev/sr1
DEVPATH=/devices/pci0000:00/0000:00:14.0/usb3/3-9/3-9.3/3-9.3:1.0/host20/target20:0:0/20:0:0:0/block/sr1
DEVTYPE=disk
MAJOR=11
MINOR=1
SEQNUM=5885
SUBSYSTEM=block
...
KERNEL[2409061.134076] change
/devices/pci0000:00/0000:00:14.0/usb3/3-9/3-9.3/3-9.3:1.0/host20/target20:0:0/20:0:0:0/block/sr1
(block)
ACTION=change
DEVNAME=/dev/sr1
DEVPATH=/devices/pci0000:00/0000:00:14.0/usb3/3-9/3-9.3/3-9.3:1.0/host20/target20:0:0/20:0:0:0/block/sr1
DEVTYPE=disk
> DISK_MEDIA_CHANGE=1
MAJOR=11
MINOR=1
SEQNUM=5889
SUBSYSTEM=block
(Both of these events trigger a call out to the 'cdrom_id' userspace
program, the latter of which interferes with the gnome-session
auto-mounting feature.)
With a systemtap probe, I can also see that there are four userspace
openers of the cdrom when it is added:
(parent) -> (child) : system-tap probe-point
-----------------------------------------------------------------------
> systemd-udevd(849) -> systemd-udevd(6783) :
> module("cdrom").function("cdrom_open@drivers/cdrom/cdrom.c:980")
systemd-udevd(6783) -> cdrom_id(6791) :
module("cdrom").function("cdrom_open@drivers/cdrom/cdrom.c:980")
systemd-udevd(849) -> systemd-udevd(6783) :
module("cdrom").function("cdrom_open@drivers/cdrom/cdrom.c:980")
systemd-udevd(6783) -> cdrom_id(6794) :
module("cdrom").function("cdrom_open@drivers/cdrom/cdrom.c:980")
where on the first opener, the kernel eventually invokes
sr_mod::sr_check_events() and gets a DISK_EVENT_MEDIA_CHANGE return
code.
In the case of my USB DVD -> laptop example, there is no media in my
device, however I still see the DISK_EVENT_MEDIA_CHANGE event. This is
a bit confusing, and I was wondering if there was an explanation for
the following:
drivers/scsi/sr.c :: sr_probe()
disk->events = DISK_EVENT_MEDIA_CHANGE | DISK_EVENT_EJECT_REQUEST;
...
cd->media_present = 1;
DISK_EVENT_MEDIA_CHANGE events will pass through to userspace and
for some reason cd->media_present defaults to 1? More on that below...
drivers/scsi/sr.c :: sr_check_events()
...
do_tur:
/* let's see whether the media is there with TUR */
last_present = cd->media_present;
ret = scsi_test_unit_ready(cd->device, SR_TIMEOUT, MAX_RETRIES, &sshdr);
/*
* Media is considered to be present if TUR succeeds or fails with
* sense data indicating something other than media-not-present
* (ASC 0x3a).
*/
cd->media_present = scsi_status_is_good(ret) ||
(scsi_sense_valid(&sshdr) && sshdr.asc != 0x3a);
if (last_present != cd->media_present)
cd->device->changed = 1;
if (cd->device->changed) {
events |= DISK_EVENT_MEDIA_CHANGE;
cd->device->changed = 0;
cd->tur_changed = true;
}
...
sr_check_events() compares the previous (in this case, default)
media_present value with what the TUR returns. If it has changed, then
turn on the DISK_EVENT_MEDIA_CHANGE event bit.
In my laptop USB DVD case, !scsi_status_is_good and sshdr.asc == 0x3a,
so last_present (1) and cd->media_present (0) mis-compare and the change
event is set. That does not seem intuitive to me, is this a bug?
Bringing this back to the reported BMC case, which presumably does have
"media" present in the virtual device... is it reasonable to expect a
DISK_EVENT_MEDIA_CHANGE even for a new device that contains media? (I
haven't verified, but in this case GET_EVENT_STATUS_NOTIFICATION might
be enough to set media present.)
If there is documentation that explains DISK_EVENT_MEDIA_CHANGE conditions
somewhere, feel free to point me there.
Thanks,
-- Joe
