|
|
 Up |
|
|
  |
Author: BradBrad
Date: Feb 13, 2008 21:44
The following diffs adds support for the Intel ICH9 Ethernet chipsets.
There is also a small change in here that affects the ICH8 chipsets.
Please test this with any em(4) adapters but especially with ICH8 and
ICH9. Please provide a dmesg.
Index: if_em.c
===================================================================
RCS file: /cvs/src/sys/dev/pci/if_em.c,v
retrieving revision 1.176
diff -u -p -r1.176 if_em.c
--- if_em.c 4 Feb 2008 00...
|
| Show full article (24.15Kb) |
|
| |
no comments
|
|
|
Author: Jeremy EvansJeremy Evans
Date: Feb 13, 2008 16:23
On 11/14 05:31, Jeremy Evans wrote:
> Here is a patch that lets nc use -U and -u simultaneously. When you use
> -U and -u together, it connects or listens (with -l) on a unix socket of
> type SOCK_DGRAM. This appears to work from my testing, but I have some
> questions:
>
> 1) Am I using mktemp securely? I'm creating a socket and not a file
> with the result, but I'm not sure if that is secure or not.
> 2) I couldn't get connecting to a unix datagram socket working without
> creating a temporary socket for the connecting program, so the receiving
> program has some place to send the data back. It makes sense that a
> connectionless protocol would need a named socket to return data on, but
> I'm not positive about that.
> 3) The man page doesn't mention that you can't use -U with -u, so...
|
| Show full article (3.22Kb) |
|
| |
no comments
|
|
|
Author: Graeme LeeGraeme Lee
Date: Feb 13, 2008 16:13
I'm following -current on my machine, and am experiencing a kernel panic
with ahci. This is with the controller set to both Normal, and
AHCI->SATA modes in the bios. If it's left as normal, I can disable
AHCI and it works as a standard ide, but it is tragically slow ( <
1.5MB/s peak transfer). Same deal for SP kernels.
OpenBSD 4.2-current (GENERIC.MP) #4: Tue Feb 12 11:46:13 EST 2008
root@ mapg.omniconnect.com.au:/usr/src/sys/arch/amd64/compile/GENERIC.MP
real mem = 2011688960 (1918MB)
avail mem = 1941430272 (1851MB...
|
| Show full article (4.91Kb) |
|
no comments
|
|
|
Author: Owain AinsworthOwain Ainsworth
Date: Feb 13, 2008 13:33
On Wed, Feb 13, 2008 at 05:38:22PM -0300, Gleydson Soares wrote:
> when i tried to run the cwm with unknown option, the getopt return a undefined char '?'.
> to fix, i added a new function usage().
Commited, thanks.
--
There is no time like the pleasant.
|
| |
|
no comments
|
|
|
Author: Henning BrauerHenning Brauer
Date: Feb 13, 2008 13:14
* Marc Balmer openbsd.org> [2008-02-13 21:57]:> Henning Brauer wrote:
>
>> * Mark Kettenis [2008-02-12 21:27]:
>>> Anyhow, you cannot use this argument to keep people from adding new
>>> watchdog devices. You're just as unsure whether any of the devices
>>> for which we already have a driver will work.
>> exactly. the argument is dumb.
>
> heya, yI did not tell anybody to not commit a driver.. just mentioning
> a problem.
"I did myself not commit a watchdog driver for this chip for this exact
reason."
doesn't parse that way for me, but anyway. doesn't really matter as
long as that argument is out of the way for a new driver.
|
| Show full article (1.33Kb) |
|
no comments
|
|
|
Author: Marc BalmerMarc Balmer
Date: Feb 13, 2008 12:59
Henning Brauer wrote:
> * Mark Kettenis [2008-02-12 21:27]:
>> Anyhow, you cannot use this argument to keep people from adding new
>> watchdog devices. You're just as unsure whether any of the devices
>> for which we already have a driver will work.
>
> exactly. the argument is dumb.
heya, yI did not tell anybody to not commit a driver.. just mentioning
a problem.
> if this is really much of a concern, teach the watchdoig framework about
> multiple watchdogs. guess that'll has to happen anyway, maybe
> timecounter-alike.
I suggest multiple independent watchdogs, not selecting a specific one
as the single one.
|
| |
|
no comments
|
|
|
Author: Gleydson SoaresGleydson Soares
Date: Feb 13, 2008 12:45
when i tried to run the cwm with unknown option, the getopt return a undefined char '?'.
to fix, i added a new function usage().
Index: calmwm.c
===================================================================
RCS file: /cvs/xenocara/app/cwm/calmwm.c,v
retrieving revision 1.8
diff -u -r1.8 calmwm.c
--- calmwm.c 13 Feb 2008 12:09:47 -0000 1.8
+++ calmwm.c 13 Feb 2008 19:24:50 -0000
@@ -71,7 +71,7 @@
DefaultFontName = xstrdup(optarg);
break;
default:
- errx(1, "Unknown option '%%c'", ch);
+ usage();
}
}
argc -= optind;
@@ -320,4 +320,13 @@
while ((pid = waitpid(-1, &status, WNOHANG)) > 0 || ...
|
| Show full article (1.34Kb) |
|
no comments
|
|
|
Author: Henning BrauerHenning Brauer
Date: Feb 13, 2008 05:09
* Mark Kettenis [2008-02-12 21:27]:
> Anyhow, you cannot use this argument to keep people from adding new
> watchdog devices. You're just as unsure whether any of the devices
> for which we already have a driver will work.
exactly. the argument is dumb.
if this is really much of a concern, teach the watchdoig framework about
multiple watchdogs. guess that'll has to happen anyway, maybe
timecounter-alike.
--
Henning Brauer, hb@bsws.de, henning@ openbsd.org
BS Web Services, http://bsws.de
Full-Service ISP - Secure Hosting, Mail and DNS Services
Dedicated Servers, Rootservers, Application Hosting - Hamburg & Amsterdam
|
| |
|
no comments
|
|
|
|
|
|
|
|
|
|
|
