Jeff Roberson [Sat, 6 Jul 2002 03:46:36 +0000 (03:46 +0000)]
Add a new configuration directive that inserts calls to debugging functions
in the VOP inlines. This is intended to replace the simple locking
specifications for calls that have more complicated behavior such as rename and
lookup.
The syntax of the new entries is:
#! name pre/post function
If the function is marked 'pre' it is executed prior to calling the VOP and
takes a pointer to a struct vop_{name}_args as it's only parameter.
If the function is marked 'post' it is executed after the VOP call and takes
a pointer to a struct vop_{name}_args as it's first parameter and the integer
return value from the vop as the second paramter.
Andrew Gallatin [Fri, 5 Jul 2002 23:33:27 +0000 (23:33 +0000)]
Remove bogus vm_page_wakeup() in vm_page_cowfault() that will cause panics
in the zero-copy send path if a process attempts to write to a page
which is still in flight.
Luigi Rizzo [Fri, 5 Jul 2002 22:43:06 +0000 (22:43 +0000)]
Implement the last 2-3 missing instructions for ipfw,
now it should support all the instructions of the old ipfw.
Fix some bugs in the user interface, /sbin/ipfw.
Please check this code against your rulesets, so i can fix the
remaining bugs (if any, i think they will be mostly in /sbin/ipfw).
Once we have done a bit of testing, this code is ready to be MFC'ed,
together with a bunch of other changes (glue to ipfw, and also the
removal of some global variables) which have been in -current for
a couple of weeks now.
Merge our ELF branding into Binutils 2.12_20020622 snapshot.
(note that this will be the last Binutils update to offer our original
method of ELF branding (WANT_OLD_BRANDELF_METHOD))
o Change almost all magic numbers to the appropriate constants.
o Fix zero payloading, unbreak ping -s 0.
o Increase socket recieve buffer, ping -s 65467 is working now.
Submitted by: anti-magic sweep based on kris's patch
Reviewed by: bde, silence on -audit
MFC after: 2 months
Add a driver for the Broadcom BCM52xx "mini-theta" PHY. This includes the
internal PHY on the 3COM 3C905B and 3C905C parts, however I've rigged it so
that xlphy (aka exphy) takes precedence for the time being.
If people try this with their xl cards and decide that it's a better choice,
we can switch this later.
This is the PHY used in various iMacs and possibly other GMAC-equipped
Macintoshes with 10/100 PHYs (the ones with 10/100/1000 appear to use brgphy).
Jeff Roberson [Fri, 5 Jul 2002 05:51:25 +0000 (05:51 +0000)]
Cleanups for vnode lock debugging.
- Tell IS_LOCKING_VFS to ignore block and character devices. specfs vnodes
aren't locked for io and they just generate lots of false positives.
- Add newlines to the badlock prints.
Peter Wemm [Fri, 5 Jul 2002 01:27:35 +0000 (01:27 +0000)]
Back out proc part of last commit. UMA manages the thread cache only, and
we just have to deal with the kstack when told to. We do not have a
UMA-managed cache for the proc struct and its associated upage yet. So,
go back to the old lazy mechanism. Note that if UMA destroys pages that
used to contain proc structures, we'll lose the corresponding upage
forever. (zones never did this - once a page was allocated, it stayed
attached to the proc zone forever)
Peter Wemm [Fri, 5 Jul 2002 00:42:09 +0000 (00:42 +0000)]
Copy from sparc64/pmap.c rev 1.64 (Retrofit changes from i386/pmap.c
rev 1.328-1.331.) but for uarea only. We still have our own broken
kstack code here.
Peter Wemm [Fri, 5 Jul 2002 00:06:08 +0000 (00:06 +0000)]
Turn on BPF_ALIGN for all non-i386 platforms, instead of having an
ifdef list that currently lists all the non-i386 platforms that bpf
currently works on.
Brian Somers [Thu, 4 Jul 2002 23:33:35 +0000 (23:33 +0000)]
Back out the previous revision
Objected to by: Andre Oppermann <oppermann@pipeline.ch>
After Andre's objection, I've re-examined rfc 2759 and noted that it
says that the domain name shouldn't be used when generating the
NT-Response field. So it looks like the bug is in freeradius rather
than in ppp.
John Polstra [Thu, 4 Jul 2002 23:19:20 +0000 (23:19 +0000)]
Disable hardware checksum offloading of IP fragments in the bge
driver. I tried a few obvious experiments, but was unable to make
the 3c996B-T generate correct UDP checksums for transmitted fragmented
packets. I'm not so sure the device is even capable of it.
Alan Cox [Thu, 4 Jul 2002 22:07:37 +0000 (22:07 +0000)]
o Resurrect vm_page_lock_queues(), vm_page_unlock_queues(), and the free
queue lock (revision 1.33 of vm/vm_page.c removed them).
o Make the free queue lock a spin lock because it's sometimes acquired
inside of a critical section.
Remove vestiges of old code...
These functions are always called on new memory so they can
not already be set up, so don't bother testing for that.
(This was left over from before we used UMA (which is cool))
Johan Karlsson [Wed, 3 Jul 2002 19:19:54 +0000 (19:19 +0000)]
Remove -p flag.
OpenBSD's implementation lacks -p, and we don't want to
support the option now, only to lose it if/when we later
switch to OpenBSD's implementation.
This functionality is provided by which(1).
Alan Cox [Wed, 3 Jul 2002 19:16:37 +0000 (19:16 +0000)]
o Make the reservation of KVA space for kernel map entries a function
of the KVA space's size in addition to the amount of physical memory
and reduce it by a factor of two.
Under the old formula, our reservation amounted to one kernel map entry
per virtual page in the KVA space on a 4GB i386.
Tim J. Robbins [Wed, 3 Jul 2002 14:38:27 +0000 (14:38 +0000)]
Avoid checking whether each line is the last line of the file when we don't
need to know. Instead, check when we are trying to match a "$" address.
This does not change the way sed processes regular files, but makes it behave
more sensibly when used interactively.
Fix another POLA breakage in a new tar: in diff mode presence of file
in the archive, but its absence in the file system should be reported
as difference, not as error.
When we are extracting directory with --unlink option and the directory
already exists don't spew bogus "Cannot unlink" error and exit with
non-zero code.
Ruslan Ermilov [Wed, 3 Jul 2002 12:28:03 +0000 (12:28 +0000)]
Try really hard to fix parallel installs. Add a bunch of .ORDER
directives to ensure that all realinstall sub-tasks are executed
after beforeinstall, similarly ensure that all afterinstall sub-
tasks are executed after realinstall. Demonstration: