HAL issues

[Joe Marcus Clarke]

On 8/3/11 6:15 PM, Kevin Oberman wrote:
> On Sun, Jul 31, 2011 at 10:37 PM, Kevin Oberman <kob6558 at gmail.com> wrote:
>> On Sat, Jul 30, 2011 at 8:59 PM, Joe Marcus Clarke <marcus at freebsd.org> wrote:
>>> On 7/27/11 6:45 PM, Kevin Oberman wrote:
>>>> Joe,
>>>>
>>>> It's taken a while, but here is verbose output from hald showing the problems.
>>>> Seems like I need to trigger a re-taste to get hald to proceed, but there are
>>>> still issues with the contents of /media/.hal-mtab.
>>>>
>>>> The attached log is annotated to explain what is triggering the output.
>>>>
>>>> Hope this helps! Let me know if other information would help.
>>>
>>> Try http://www.marcuscom.com/downloads/patch-hald_freebsd_hf-storage.c
>>>
>>> Joe
>>
>> No difference that I can see. When I plug the disk in, nothing
>> happens. lshal only reports da0. If I do a geli attach of da0s3, all
>> three file systems mount. All is as it should be. If I unmount all of
>> them and disconnect/re-connect, nothing mounts, again. If I
>> fsck_msdosfs /dev/da0s2, it mounts all three filesystems, but the file
>> system on da0s1 is not added to .hal-mtab, so I have to edit the file
>> to unmount it.
>>
>> All of this is exactly what was happening prior to the latest patch.
>> The annotated hald verbose output is attached, though it may or may
>> not be useful.
>>
>> If I get a bit of time, I'll build hald for debugging an see if I can
>> find out what is happening. I've read enough of the code that I may be
>> able to figure out what is needed. I assume that you don't have any
>> disks configured to do this.
> 
> Joe,
> 
> Any clues about how to debug hald?
> 
> I built hald with WITH_DEBUG and can start it in gdb, but I let it do
> its initial
> run and then interrupt and set break-points only to get:
> Continuing.
> *** [DIE] hald_runner.c:runner_died():204 : Runner died
> 
> Program exited with code 01.
> 
> Is there a way to avoid this or am I going to be stuck with debug via
> print statements? :-(

I never use WITH_DEBUG.  I think I was able to debug hald, but it's been
a while.  I would build with "-g -O -pipe" and set STRIP= .

I don't see what could be causing the problem you're seeing.  I suppose
there may be a problem seeing all of the file systems on the disk.  I
don't have a removable disk with multiple file systems, and I don't use
geli.  You might be better off using printf's to look at the hf-storage
code at the time the disk attaches.  Walk through the code paths there.

Joe

-- 
Joe Marcus Clarke
FreeBSD GNOME Team	::	gnome at FreeBSD.org
FreeNode / #freebsd-gnome
http://www.FreeBSD.org/gnome