On Thu, 17 Jan 2002, Michael Reifenberger wrote:
On Thu, 17 Jan 2002, Bruce Evans wrote:
...
Oops. There should be no alias for md10c. Try this version. It fixes
the may want an alias case in dkmodminor() and moves all the dk inlines
to subr_diskslice.c.
This patch works better.
On Fri, 18 Jan 2002, Bruce Evans wrote:
...
It somehow worked here, but after fixing the non-removal of the
/dev/md11*, I got back what is probably the original panic too. At
least with my fixed changes, the panic is caused by not clearing
pointers in destroy_dev(). The pointers come back
On Fri, 18 Jan 2002, Michael Reifenberger wrote:
On Fri, 18 Jan 2002, Bruce Evans wrote:
...
devfs still leaks 2 DEVFS allocations and 6 devbuf allocations for every
iteration of the script.
...
Does this mean that your latest patch removes the panic but leaks some
resources.
No :).
On Thu, 17 Jan 2002, Bruce Evans wrote:
...
Oops. There should be no alias for md10c. Try this version. It fixes
the may want an alias case in dkmodminor() and moves all the dk inlines
to subr_diskslice.c.
This patch works better.
This brings us back to the original problem :-)
Now it
On Mon, 14 Jan 2002, Michael Reifenberger wrote:
On Tue, 15 Jan 2002, Bruce Evans wrote:
...
Try this version. Only disklabel.h has many changes. The code for
avoiding creation of bogus 'c' partitions didn't work at all.
This works during startup but the following commands cases a
Hi,
more food:
The kernel startup is:
...
vga: ...
makedev(116,65538)
make_dev(maj=116,min=65538,name=ad0)
ad0: ...
makedev(116,65546)
make_dev(maj=116,min=65546,name=ad1)
ad1: ...
acd0: ...
Mounting root from ufs:/dev/ad0s3a
makedev(116,65538) (?!?!? Why, where)
On Mon, 14 Jan 2002, Michael Reifenberger wrote:
Hi,
more food:
The kernel startup is:
...
vga: ...
makedev(116,65538)
make_dev(maj=116,min=65538,name=ad0)
ad0: ...
makedev(116,65546)
make_dev(maj=116,min=65546,name=ad1)
ad1: ...
acd0: ...
Mounting root from ufs:/dev/ad0s3a
On Tue, 15 Jan 2002, Bruce Evans wrote:
...
Yes, it is a bit confusing. sys/types.h gives the userland version.
Only as confusing as the rest of the dk* code :-)
Try this version. Only disklabel.h has many changes. The code for
avoiding creation of bogus 'c' partitions didn't work at
On Sat, 12 Jan 2002, Michael Reifenberger wrote:
Hi,
more input:
The panic occurs in dsname() while dereferencing devsw(dev)-d_name.
devsw(dev) gives NULL.
This was fairly clear from the panic message (the low fault address
is almost certainly for a null pointer + offset, and it's fairly
Hi,
more input:
The panic occurs in dsname() while dereferencing devsw(dev)-d_name.
devsw(dev) gives NULL.
The call to dsname() occurs from dkmodminor() defined in disklabel.h
I've seen two calls to dsname, only the second one panics.
Any thoughts?
Bye!
Michael Reifenberger
^.*Plaut.*$, IT,
Hi,
Hmm. I used an old set of patches to avoid filtering out local changes
again. Try the enclosed up to date patches.
Hmmm.
Fresh -current, fresh patch.
(Hunk 2 still doesnt apply cleanly, but this seems to be irrelevant
for the problem)
Old symptom:
...
mounting root from ufs:/dev/ad0s3a
On Tue, 8 Jan 2002, Michael Reifenberger wrote:
Hi,
the following script panics my system when executed twice:
(uncomment the dd lines for the first time or do it manually)
##
#!bin/sh
block_size=512
blocks=$((120 * 1024 * 1024 / $block_size))
heads=16
sectors=63
In message [EMAIL PROTECTED], Bruce Evans writes:
This seems to be just another null pointer panic caused by the dk macros
creating half-baked devices with null devswitches. I sent the following
quick fixes for this to the devfs maintainer a couple of weeks ago. They
also fix the non-creation
On Wed, 9 Jan 2002, Bruce Evans wrote:
...
This seems to be just another null pointer panic caused by the dk macros
creating half-baked devices with null devswitches. I sent the following
quick fixes for this to the devfs maintainer a couple of weeks ago. They
also fix the non-creation of
On Wed, 9 Jan 2002, Michael Reifenberger wrote:
On Wed, 9 Jan 2002, Bruce Evans wrote:
...
This seems to be just another null pointer panic caused by the dk macros
creating half-baked devices with null devswitches. I sent the following
quick fixes for this to the devfs maintainer a
15 matches
Mail list logo