| 1 |
On Mon, Jan 7, 2013 at 7:18 AM, Grant Edwards <grant.b.edwards@×××××.com> wrote: |
| 2 |
> On 2013-01-07, Dale <rdalek1967@×××××.com> wrote: |
| 3 |
> |
| 4 |
>> I'm not sure that is a bug. As I posted earlier, this was changed a |
| 5 |
>> good while back. There was a reason for it but I can't recall what it |
| 6 |
>> was. The new devices for CD/DVDs is /dev/sr*. |
| 7 |
> |
| 8 |
> It's been something like 6-8 years hasn't it? |
| 9 |
|
| 10 |
IIRC the SATA interface has always labeled them as /dev/sdX /dev/srX. |
| 11 |
Everything I've built using new hardware in the last 5 years has been |
| 12 |
SATA based and I've not had a new machine with /dev/hdX in longer than |
| 13 |
I can remember. |
| 14 |
|
| 15 |
However, best I can tell, that has _nothing_ to do with why /dev/cdrom |
| 16 |
/dev/dvd disappeared in the last couple of months. Remember, my |
| 17 |
machines have all had /dev/srX. |
| 18 |
|
| 19 |
Going back to my post with one of many solutions to this issue: |
| 20 |
|
| 21 |
First - the old way that udev was recognizing the cdrw/dvd drive on my |
| 22 |
system was via an ID_PATH value for the pci device: |
| 23 |
|
| 24 |
#SUBSYSTEM=="block", ENV{ID_CDROM}=="?*", |
| 25 |
ENV{ID_PATH}=="pci-0000:00:1f.2-scsi-0:0:0:0", SYMLINK+="cdrom", |
| 26 |
ENV{GENERATED}="1" |
| 27 |
|
| 28 |
However you will note that ID_PATH (the key used by udev) doesn't exist any more |
| 29 |
|
| 30 |
c2stable ~ # udevadm info --query=all --name=/dev/sr0 | grep ID_PATH |
| 31 |
c2stable ~ # |
| 32 |
|
| 33 |
|
| 34 |
Best guess I have is that ID_PATH may have been changed to DEVPATH |
| 35 |
|
| 36 |
c2stable ~ # udevadm info --query=all --name=/dev/sr0 | grep DEVPATH |
| 37 |
E: DEVPATH=/devices/pci0000:00/0000:00:1f.2/ata11/host10/target10:0:0/10:0:0:0/block/sr0 |
| 38 |
c2stable ~ # |
| 39 |
|
| 40 |
What I did was ask udev to identify by the drive's model number using ID_MODEL: |
| 41 |
|
| 42 |
New way: |
| 43 |
SUBSYSTEM=="block", ENV{ID_CDROM}=="?*", |
| 44 |
ENV{ID_MODEL}=="Optiarc_DVD_RW_AD-7241S", SYMLINK+="cdrom", |
| 45 |
ENV{GENERATED}="1" |
| 46 |
|
| 47 |
A little playing around suggest you can use anything unique to the device. |
| 48 |
|
| 49 |
Now, my point is that change to /dev/srX was the root cause is FUD. It |
| 50 |
isn't the root cause of this change because it didn't change on my |
| 51 |
systems. All I know is that ID_PATH (from the old file) used to work |
| 52 |
and no longer does. Whatever is responsible for creating that, likely |
| 53 |
some portion of the kernel, changed the value and created a need to |
| 54 |
modify how udev looks at the system. Is it a bug? I don't know. It's |
| 55 |
just the way it is. |
| 56 |
|
| 57 |
Just my views, |
| 58 |
Mark |
Archives