[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [bluetooth-dev] destroyed proc filesystem
First of all, please use the -u option to diff when creating
a patch as that provides a context and makes it sooo much
easier to read.
Second, the code needs to support creating the /proc files
for 2.0, 2.2 as well as 2.4, and most of your patch seems to
be about removing support for Linux < 2.4. This also makes
it very hard to see what you actually changed in the 2.4 case.
Could you please provide a cleaner patch?
And btw, the /proc files look perfectly fine on my 2.4.10
system (little endian).
> -----Original Message-----
> From: Matthias Fuchs [mailto:email@example.com]
> Sent: Friday, October 12, 2001 18:39
> To: bluetooth-dev
> Subject: [bluetooth-dev] destroyed proc filesystem
> After loading the bt.o lkm I wondered about my /proc directory. All
> entries that should come from the OpenBT stack are unreadable. Here
> is how it nearly looks like:
> -r-------- 1 0 0 16781312 kcore
> -rw-r--r-- 1 0 0 0 ppc_htab
> -r--r--r-- 1 0 0 0 misc
> -r--r--r-- 1 0 0 0
> -r--r--r-- 1 0 0 0 ribe_bt_buf: Only %d available (non - fragmented)
> -r--r--r-- 1 0 0 0 ilable (non - fragmented)
> -rw-rw-rw- 1 0 0 0 sdp_srv
> I seems to be a problem with the way the proc files are setup
> in openbt. I took a look in one of my own drivers that uses
> the proc filesystem. Where I used a much simpler way to
> generate proc entries. I made some changes to sdp.c to see if
> it makes any differences ... and it does. The above screenshot
> is made AFTER I changed sdp.c to use a different way to register
> the sdp_srv proc entry. The others are still untouched. SO that's
> why you see an uncorrupted sdp_srv entry and all other entries
> are unusable. I am not sure what happened to the /proc directory.
> When I unload bt.o, everything looks fine again !
> My target is (still) a PowerPC 405 (IBM) with Linux kernel
> 2.4.2. I get the same result when I use 2.4.10 on the same
> target. Did anybody else get these phenomenon ? Is there
> something wrong with proc implementation in openbt ? It was
> working fine on my board for a long time but now this
> happens :-(
> I attached a diff between the original sdp.c and my version
> that creates an usable /proc/sdp_srv entry. The code is not
> clean ... it's my (dirty) working version. Perhaps anybody
> know what is wrong.
To unsubscribe from this list: send the line "unsubscribe bluetooth-dev" in
the body of a message to firstname.lastname@example.org