| Philip Guenther | correct mxcsr+mxcsr_mask handling (revised)
Here's a revised diff to correct the handling on amd64 and i386 of the
MXCSR register in frames that could be altered by users, so that a user
can't trick the kernel into faulting by trying to load an invalid MXCSR
value during the return to userspace.
Originally worked out by joshe@; this corrects the timing of the call to
fpuinit() for the primary CPU on amd64 so that the saved mask value is
initialized correctly. It also uniformly uses
__attribute__((aligned(16))) do get a buffer of ...
| Dec 25, 1:02 pm 2010 |
| Ted Unangst |
| Dec 25, 12:35 pm 2010 |
| Philip Guenther | amd64 atomic macro naming cleanup
Fix the naming of the atomic macros on amd64: right now, the
x86_atomic_*_l() macros actually operate on unsigned 32bit integers
instead of longs, so:
1) change the callers to use the _u32 versions
2) update the _ul definitions to bew the 64bit versions
3) remove the _l macros
Open question: perhaps the _ul macros should be removed too (they're
unused after this diff) and future code should just explicitly use the
_u32 or _u64 macros?
Also, complete remove the x86_multicast_ipi() ...
| Dec 25, 12:12 pm 2010 |
| Alexandr Shadchin | cleanup directory sys/dev/pckbc
remove file sys/dev/pckbc/Makefile
This file is accidentally imported from NetBSD, and probably never used.
--
Alexandr Shadchin
Index: Makefile
===================================================================
RCS file: Makefile
diff -N Makefile
--- Makefile 13 Nov 2000 20:12:34 -0000 1.1
+++ /dev/null 1 Jan 1970 00:00:00 -0000
@@ -1,8 +0,0 @@
-# $OpenBSD: Makefile,v 1.1 2000/11/13 20:12:34 aaron Exp $
-# $NetBSD: Makefile,v 1.1 1998/06/12 23:22:54 cgd Exp $
-
-INCSDIR= ...
| Dec 25, 10:57 am 2010 |
| Marc Espie | patch: register-plist should force package names to increase
A bit McGyverish, since the tools we have usually deal with actual packages,
so OpenBSD::Search is out of the question.
This makes register-plist error out if it finds a newer packing-list with
the exact same pkgpath in the PLIST_DB.
So, if you use PLIST_DB (by default), it will notice if you break things
by having version numbers go backwards, instead of only checking that in
dependencies.
In case you end up with bogus packing-list, you have all the information
to remove it anyways ...
| Dec 25, 4:46 am 2010 |
| Ted Unangst | add back tcp sysctls
As I mentioned previously, the auto recv space scaling algorithm isn't
optimized for all links. At list in my case, with the proverbial
satellite link (high bandwidth, high latency), the window never appears to
grow.
Manually setting the default recv space allows me to download at high
speed again. This diff brings back the two relevant sysctls.
Index: tcp_var.h
===================================================================
RCS file: ...
| Dec 24, 8:50 pm 2010 |
| Ted Unangst | Re: add back tcp sysctls
On Sat, Dec 25, 2010 at 6:03 AM, Mark Kettenis <mark.kettenis@xs4all.nl>
Everything is relative. It's 6 Mbit (not *that* fast), but has ping
times reminiscent of dial up.
| Dec 25, 10:54 am 2010 |
| Mark Kettenis | Re: add back tcp sysctls
Anyway, it'd be nice if the algorithm could be improved such that it
works for a broader range of links. For example, you probably want to
scale more aggressively if the RTT is large.
That said, my experience with Linux, which supposedly has had window
auto-tuning for quite some time, tells me that there will always be
corner cases where some manual tuning is necessary. So this diff
| Dec 25, 4:03 am 2010 |
| Marco Peereboom |
| Dec 25, 7:19 am 2010 |
| Philip Guenther | Re: compat_bsdos leftover
On Fri, Dec 24, 2010 at 12:45 PM, Mark Kettenis <mark.kettenis@xs4all.nl>
wrote:
Yep, the default ldt should be able to go away completely and the ldt
register explicitly set to zero except for processes that use the
i386_set_ldt() sysarch call. I'm currently separated from my i386
test box, so I won't hazard a diff for this right now, but it deserves
a blow torch to clean it out.
Philip Guenther
| Dec 24, 5:11 pm 2010 |
| Clint Pachl | Re: pkg_delete recursive uninstall option
I usually remove the diff between manually installed pkgs and all
top-level pkgs, then repeat until there is no more diff. But this always
seemed so ad-hoc.
This feature Useche submitted would make my life slightly easier and
keep my system cleaner while being less error prone.
| Dec 25, 12:20 am 2010 |
| Nima Hoda |
| Dec 24, 11:00 pm 2010 |
| Philip Guenther | Re: Make sys/user.h safe for multiple includes
On Mon, Dec 20, 2010 at 4:06 PM, Christian Ludwig
You have a header that needs the full struct user definition and can't
make do with a forward declaration of it?
(Note that the struct user definition is the _only_ correct reason to
#include <sys/user.h>)
This has been handled so far by simply including sys/user.h in each
file that needs the header. For example, machine/freebsd_machdep.h
would seem to need the struct user definition for its offsetof usage,
but instead the one .c file that ...
| Dec 24, 5:19 pm 2010 |
