HAL causing slab cache overflow

[Fischer, Anna]

I am running hald 0.5.6 on a SLES10 Linux box with a 2.6.16.13-4 SMP kernel.

I have attached a log file showing the slab cache statistics on my system which shows that there is a huge number of socket buffers and their 1024 byte data chunks hanging around. When analysing the data information I have seen that those socket buffers are including HAL messages that seem to be pending on the d-bus somewhere. Over time the number of these objects in the slab cache is increasing consistently until the system runs out of memory. Has anyone ever seen this before? Why are there so many HAL messages pending on the system?

I don't know much about HAL and I don't have a specific HAL configuration. I have attached some more details on my running system. Where can I look to find out why the HAL daemon doesn't seem to work properly?

Many thanks,
Anna

-------------- next part --------------
A non-text attachment was scrubbed...
Name: lshal.log
Type: application/octet-stream
Size: 192013 bytes
Desc: lshal.log
Url : http://lists.freedesktop.org/archives/hal/attachments/20080520/d4e8e66b/attachment-0003.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ps-hal.log
Type: application/octet-stream
Size: 220 bytes
Desc: ps-hal.log
Url : http://lists.freedesktop.org/archives/hal/attachments/20080520/d4e8e66b/attachment-0004.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: slabinfo.log
Type: application/octet-stream
Size: 15891 bytes
Desc: slabinfo.log
Url : http://lists.freedesktop.org/archives/hal/attachments/20080520/d4e8e66b/attachment-0005.obj