1 |
On Fri, Nov 30, 2012 at 3:09 AM, Walter Dnes <waltdnes@××××××××.org> wrote: |
2 |
> On Fri, Nov 30, 2012 at 12:28:01AM +0000, Jorge Almeida wrote |
3 |
> |
4 |
>> When a pen is inserted, this is what is set: |
5 |
>> |
6 |
>> DEVNAME=sdd |
7 |
> |
8 |
> This is almost exactly what I remember from when I was |
9 |
> writing/testing/debugging my automount scripts. The one difference I |
10 |
> remember is that when mdev handled it, there was no "DEVNAME" variable, |
11 |
> but rather it was "MDEV". But otherwise identical. Two important notes... |
12 |
|
13 |
Strange. The variable name is set by the kernel. Maybe you were using a |
14 |
patched kernel? |
15 |
|
16 |
> |
17 |
> 1) For a USB mass storage device (pen or external hard drive) with N |
18 |
> partitions, the hotplug handler will get N+1 events when inserting and |
19 |
> also when removing. E.g. if your pen drive has 3 partitions, you'll get |
20 |
> 4 events... |
21 |
|
22 |
Right. The helper program is launched 4 times and will create (or remove, |
23 |
depending on $ACTION) 4 symlinks. |
24 |
> |
25 |
> Would you be OK if the devices were always mounted in /media ? The |
26 |
> reason I ask is that my scripts use pmount, which can take an optional |
27 |
> label argument. E.g. if MDEV is "sdd1" |
28 |
> |
29 |
|
30 |
pmount looks a good idea for a many-users workstation. For a 1-user computer, |
31 |
there's no loss of functionality with sticking with fstab. |
32 |
|
33 |
> |
34 |
> Actually, if I was writing it, I would add a few lines to my script |
35 |
> for the "add" ACTION |
36 |
|
37 |
Sure, I just didn't mention it. |
38 |
|
39 |
> * label=${MDEV} |
40 |
> * look for a "serial" file in the PCI path of the newly-inserted device |
41 |
> * if found; then |
42 |
> grep through a textfile to match the contents of the "serial" file |
43 |
|
44 |
I thought of keeping data-- (key,value) = (serial_number, |
45 |
custom_name_of_device)-- in a cdb database. I think it is faster than reading |
46 |
a text file, and no parsing needed. It is also scalable (although this is |
47 |
probably irrelevant on a custom-workstation scale...). My (untested) program is |
48 |
about 11k, statically compiled. |
49 |
|
50 |
By the way, I don't suppose there is a mailing list to talk about these |
51 |
matters (mdev/ udev-alternative/ udev-fork related)? This is really |
52 |
distro-agnostic stuff... |
53 |
|
54 |
Thanks |
55 |
|
56 |
Jorge Almeida |