in base built against libpthread.19.0 cannot load
the mysql.so shared object linked against libpthread.18.1
Philip Guenther
-- Forwarded message --
From: Philip Guenther guent...@cvs.openbsd.org
Date: Mon, Apr 6, 2015 at 6:27 PM
Subject: CVS: cvs.openbsd.org: src
To: source-chan
otherwise have found.
This will require more invasive work, but will to the Right Thing for
both device and DUID users and will silently and correctly handle the
device-DUID transition.
Philip Guenther
On Mon, Mar 30, 2015 at 9:19 AM, Gregory Edigarov ediga...@qarea.com wrote:
After discussing with Sven Falempin changed the message to Packet timed
out to be more exact
updated diff:
What's the output when ping packets or their replies are reordered?
Philip Guenther
looking for
trouble. Maybe we should stop scheduling non-kernel threads before
that point?
Philip Guenther
On Fri, Mar 27, 2015 at 2:50 AM, Carlos Martín Nieto c...@dwim.me wrote:
A call to opendir thread-safe and the readdir calls only share the buffer
within the same directory stream, which is local to this function. Therefore
this lock does not buy us anything.
committed, thanks
Philip
. To quote
rfc5321:
Stamp = From-domain By-domain Opt-info [CFWS] ;
FWS date-time
...and there are no other explicit semicolons in the syntax.
Philip Guenther
On Sat, Apr 4, 2015 at 6:03 PM, Mike Burns mike+open...@mike-burns.com wrote:
...
I agree with your analysis of the grammar, but I believe that is what
the code is doing. That line in question is the one that puts the bare
semicolon before the date-time. The newlines have been re-arranged so
On Sat, Apr 4, 2015 at 10:35 AM, Jérémie Courrèges-Anglas
j...@wxcvbn.org wrote:
just to make things clear, I have no strong opinion towards Philip's
proposal. I merely had kept __BSD_VISIBLE mostly because those
functions had appeared in BSD before they were standardized. Updated
diff,
On Sat, Apr 4, 2015 at 11:31 AM, Jérémie Courrèges-Anglas
j...@wxcvbn.org wrote:
Here's some nitpicking.
Matching the .h good. ok guenther@
On Fri, Apr 3, 2015 at 3:24 PM, Todd C. Miller
todd.mil...@courtesan.com wrote:
On Fri, 03 Apr 2015 15:17:59 -0700, Philip Guenther wrote:
** it would only help someone building with something like cc
-D_POSIX_C_SOURCE=199506 -D_BSD_SOURCE=1 which, IMHO, is insane.
Why is it insane to want
On Fri, Apr 3, 2015 at 1:29 PM, Mark Kettenis mark.kette...@xs4all.nl wrote:
From: j...@wxcvbn.org (=?utf-8?Q?J=C3=A9r=C3=A9mie_Courr=C3=A8ges-Anglas?=)
Date: Fri, 03 Apr 2015 18:02:37 +0100
A few months ago, I spotted warnings in a port, related to the decl of
one of those functions not
, in case the Windows opendir/readdir emulation
isn't thread-safe (unlikely).
Philip Guenther
the user code's position in the generated .c
file. Let's not make these less portable for no gain.
Philip Guenther
On Sat, 28 Mar 2015, Carlos Mart?n Nieto wrote:
I?ve been looking into making libcrypto automatically thread-safe. The
obvious solution is to use pthread to perform the locking instead of
relying on the user to set locking callbacks, as the final user
shouldn?t need to care that LibreSSL is
On Sun, Mar 8, 2015 at 2:52 PM, Christian Weisgerber na...@mips.inka.de wrote:
On 2015-03-08, Stuart Henderson st...@openbsd.org wrote:
! needs to be escaped in tab-completion otherwise this happens.
$ ls
Season 4 UPDATE!-FFdJhCSr8ng.mp4
$ ls Season\ 4\ UPDATE!-FFdJhCSr8ng.mp4
ls: Season 4
the
flag word against a word with everything set except the one you want
to clear.
I.e, this
flags |= F_DOWN;
sets the F_DOWN bit and
flags = ~F_DOWN;
clears it.
Philip Guenther
a bug: TAPE=- tar cv won't behave correctly.
Philip Guenther
On Fri, Feb 27, 2015 at 2:25 PM, Todd C. Miller
todd.mil...@courtesan.com wrote:
I suppose it should be obvious but I feel better have this stated
explicitly.
ok, though part of me would prefer to merge it into the sentence
before the list, ala
Further, either of the tv_nsec fields can be set
On Wed, Feb 25, 2015 at 8:38 AM, Todd C. Miller
todd.mil...@courtesan.com wrote:
On Wed, 25 Feb 2015 14:57:56 +0100, Michael Felt wrote:
Lastly, the app certhash.c does not compile. AIX does not have O_DIRECTORY
defined, nor do I expect it is an acceptable argument to open (.,...)
You should
On Thu, Feb 5, 2015 at 3:19 AM, Todd C. Miller
todd.mil...@courtesan.com wrote:
Older versions of POSIX had SIZE_MAX in limits.h but C99 standardized
it in stdint.h. We should be including stdint.h to get SIZE_MAX
for portability's sake.
These all look good. ok guenther@
(A few more turned
On Wed, Jan 7, 2015 at 1:26 AM, Helg xx...@msn.com wrote:
Anyone?
Took me a couple tries, but it's in.
On Tue, Jan 13, 2015 at 11:38 PM, Mike Larkin mlar...@azathoth.net wrote:
i386 is next, but that requires a PAE paging model and compatible CPU.
I've got the PAE mode booting but it's not ready for prime time yet.
On Tue, Jan 13, 2015 at 11:36 PM, Philip Guenther guent...@gmail.com wrote:
On Tue, Jan 13, 2015 at 11:38 PM, Mike Larkin mlar...@azathoth.net wrote:
i386 is next, but that requires a PAE paging model and compatible CPU.
I've got the PAE mode booting but it's not ready for prime time yet.
Hmm
On Sun, Jan 11, 2015 at 9:36 AM, Ted Unangst t...@tedunangst.com wrote:
env won't run a command with an = in its name. This is documented as a
bug, but it's easily fixed in a backwards compatible way.
...
+ if (strcmp(*argv, --) == 0)
+ argv++;
No, env foo=bar -- baz=qux
On Sun, 11 Jan 2015, Ted Unangst wrote:
Even more awesome.
How about enforcing the full rule?
Index: env.c
===
RCS file: /cvs/src/usr.bin/env/env.c,v
retrieving revision 1.15
diff -u -p -r1.15 env.c
--- env.c 8 Mar 2014
On Mon, 5 Jan 2015, Helg wrote:
The man page for LOCK(9) says that if the kernel option LOCKDEBUG is
enabled, additional facilities are provided to assist in determining
deadlock occurrences.
I created a copy of /sys/arch/amd64/conf/GENERIC and added option
LOCKDEBUG. Executing config
[(@#*$(*# control-enter keybinding]
On Mon, Jan 5, 2015 at 7:15 PM, Philip Guenther guent...@gmail.com wrote:
On Mon, Jan 5, 2015 at 11:01 AM, Ted Unangst t...@tedunangst.com wrote:
...
In the regular timeout case, I'm not sure what you're changing. There
is a problem here though. If we're
On Mon, Jan 5, 2015 at 11:01 AM, Ted Unangst t...@tedunangst.com wrote:
...
In the regular timeout case, I'm not sure what you're changing. There
is a problem here though. If we're already close to the timeout
expiring, we shouldn't sleep the full timeout, only the time left
since we began the
the manpages!?!?). If you want to waste your own time in this
particular way, you can bear the cost of doing so by keeping that diff
in your sandbox instead of cluttering the tree.
Philip Guenther
On Sat, Jan 3, 2015 at 8:38 AM, Alexander Bluhm alexander.bl...@gmx.net wrote:
My goal is to make logging via syslog reliable. At least I want
to see when a message gets lost.
So my idea is to write a kernel log message if sendsyslog(2) cannot
deliver a message. Then you see the problem on
On Fri, Jan 2, 2015 at 1:28 PM, Boris Goldberg bo...@twopoint.com wrote:
Philip is right. It's embarrassing I haven't checked the case with hard
quota myself.
We'll look into it. Philip is actually suggesting the way. I suppose he
doesn't have time to implement it himself, does he? Don't
On Sun, Dec 28, 2014 at 6:02 AM, Ted Unangst t...@tedunangst.com wrote:
...
Agreed. Here's a different approach. If we need more memory, allocate
the required amount. No realloc since we aren't interested in the
previous contents. Continue caching the previous allocation, but don't
worry about
the C layout, so that the next
time it changes they'll get the updated format.
Philip Guenther
()
and loop again. Well, and if the timeout is set, we should tsleep()
for the set timeout.
No, that's wrong. This is inside a while() loop; if that loop's
condition holds, do you want to sleep for the original length of time
(again) or only for the remaining time?
Philip Guenther
leaving that on. We ain't using it *right now*, but,
well, the source tree on my laptop is, and more than ever. :-)
Philip Guenther
On Mon, Dec 15, 2014 at 6:04 AM, Jonathan Gray j...@jsg.id.au wrote:
So now time is printed by default afl has found that time_t values
such as -9223372035438150153 will cause localtime() to fail and
return NULL. strftime() can't deal with this and will at some point
dereference tm without
Some time ago, I had added support for using the MWAIT instruction in the
idle loop. Various people found that made their boxes run hot, to the
point that several developers diked it out of their own builds; I've
committed one of those yesteryad pending a proper fix.
So, to start on that:
On Sat, Dec 6, 2014 at 1:25 PM, Kaspars Bankovskis
kasp...@bankovskis.net wrote:
One flag was undocumented + some cleanup.
Looks good to me!
Philip Guenther
On Thu, Dec 4, 2014 at 5:53 PM, Todd C. Miller
todd.mil...@courtesan.com wrote:
On Thu, 04 Dec 2014 22:32:39 +0100, Ingo Schwarze wrote:
That's beside the point. The HISTORY section generally wants to
convey two pieces of information: Where an interface was originally
invented (in this
On Mon, Nov 24, 2014 at 11:24 AM, Florian Obser flor...@openbsd.org wrote:
...
Since we are probably not supposed to send a Content-Type header I
think it makes sense to duplicate the httpmsg generating code in this
case;
If a GET of that resource would have a Content-Type, then the HEAD of
it
On Wed, 19 Nov 2014, enh wrote:
On Tue, Nov 18, 2014 at 10:55 PM, Philip Guenther guent...@gmail.com wrote:
enh, this look as good in the perf measurement cited in that googlesource
link?
yes. performance seems roughly identical. thanks!
Committed, which leads to the next diff: use
On Wed, Nov 19, 2014 at 1:00 PM, Ted Unangst t...@tedunangst.com wrote:
The sys/file.h include is likely not the correct include for any
userland code. Most cases it's unneeded; in other cases fcntl.h is the
correct header.
Yep. Diff looks good; ok guenther@
On Tue, 18 Nov 2014, enh wrote:
On Tue, Nov 18, 2014 at 10:01 PM, Philip Guenther guent...@gmail.com wrote:
...
Maybe it would be better to do
if (!__sdidinit)
__sinit();
like other places?
good point. i hadn't considered that stdin/stdout/stderr aren't
.)
Philip Guenther
but is
initialized as follows
reverse = calloc(1, sizeof(struct systrace_alias));
Committed. Thanks!
Philip Guenther
On Thu, Nov 13, 2014 at 10:31 PM, Brad Smith b...@comstyle.com wrote:
ping. Discussion died out. Can we please move this issue forward?
It's in
Philip
is wrong, sorry. The code where a ctype function is being
called on a char from a char * pointer should be casting to (unsigned
char); the code where 'c' is set from lgetc() should declare c as an
int, as lgetc() does the necessary cast.
Philip Guenther
with the new command...)
Philip Guenther
in the man page so it fits into the otherwise
alphabetical order of the escape sequences.
This probably makes sense, though maybe \e is mis-sorted because it's
an extension? Nah.
Philip Guenther
both of
them. Or is there some reason this extension is required?
Philip Guenther
On Wed, Oct 22, 2014 at 1:25 AM, Ingo Schwarze schwa...@usta.de wrote:
OK?
Ingo
Looks good to me
Philip
. OpenBSD 5.6 was cut as a release a ways before that.
Philip Guenther
/etc/hosts? It's been removed from POSIX and I'm at
loss to think of a use of it that wouldnt make further nonportable
assumptions.
Maybe we just fix getent(1) to return an error like it does for ethers?
Philip Guenther
On Sat, Oct 18, 2014 at 5:24 PM, Ingo Schwarze schwa...@usta.de wrote:
Index: getent.c
===
RCS file: /cvs/src/usr.bin/getent/getent.c,v
retrieving revision 1.6
diff -u -p -r1.6 getent.c
--- getent.c30 Mar 2013 14:03:08 -
:
function rm {
local rmopts arg
for arg
do
case $arg in
-*) set -A rmopts -- ${rmopts[@]} $arg;;
*) command rm ${rmopts[@]} $arg || return;;
esac
done
}
Philip Guenther
and CLOCK_PROF, modeled after what FreeBSD does.
out of curiousity, what program needs these?
Possibly none, but we currently define CLOCK_VIRTUAL but don't
actually support it.
IMO we should just delete CLOCK_VIRTUAL from sys/_time.h and clock_gettime(2).
Philip Guenther
and working and replace it with something
non-standard or using more ifdefs? Umm, why?
Maybe this is a good idea, but you let out all the here's why! part
in your message.
Philip Guenther
after pulling in
sys/types.h or sys/param.h, so it's just noise and wasted compiler
cycles. The manpages for the kernel-only interfaces (bemtoh*, etc)
should presumably follow that: recommending sys/types.h until someone
removes the indirect #include.
Philip Guenther
on.
Have you actually managed to make it crash? I've already committed a
check for this when this first came out, mapping out of bounds pri
values to LOG_USER, and at that time no one was able to crash the code
without the check...
Thoughts? Okays?
Meh, seems like overkill.
Philip Guenther
On Wed, Oct 8, 2014 at 10:22 PM, David Gwynne da...@gwynne.id.au wrote:
deraadt points out i suck at grep.
ok?
Sure. Even has the same semantics if a signal is received.
/boot /boot
installboot -v /boot /usr/mdec/biosboot sd0
That should be
installboot -v sd0 /usr/mdec/biosboot /boot
(and if your boot drive isn't sd0 then change the line as appropriate,
of course.)
Philip Guenther
POSIX and does not follow the POSIX utility argument
guidelines multiple ways. Changing OpenRCS to follow those guidelines
would make it incompatible with GNU RCS and the the deployed base of
scripts...and fingers...
Philip Guenther
and assemble
the release out of that? To absolutely guarantee that it cannot pick
up a file left over in the running install from a previous version.)
Philip Guenther
the direct message.
Philip Guenther
. We've tried to do more minimal
#including/exporting generally, but SysV IPC is enough of a corner
that just pulling in sys/types.h from sys/ipc.h is fine by me.
Philip Guenther
On Sat, Sep 20, 2014 at 6:52 AM, Matti Karnaattu mkarnaa...@gmail.com wrote:
Should you add EBADMSG and EPROTO too?
IMO, it's a bug in POSIX that those weren't marked OB XSR.
Note that adding errno defines requires a libc major bump.
On Thu, Sep 18, 2014 at 3:33 PM, Matti Karnaattu mkarnaa...@gmail.com wrote:
This doesn't compile:
It's broken code: PTHREAD_ONCE_INIT is only required to be an
initializer, not a value.
Philip Guenther
about signal.h you're referring to there...)
Philip Guenther
On Sun, Sep 14, 2014 at 9:33 PM, Jean-Philippe Ouellet
jean-phili...@ouellet.biz wrote:
We use async-signal-safe (not async-signal safe) elsewhere,
and so does POSIX.
Yep. Committed to there and signal(3). Thanks!
On Mon, 8 Sep 2014, Jona Joachim wrote:
On 2014-09-06, Gr?goire Duch?ne gduch...@awhk.org wrote:
Hi,
Although CIRCLEQ_* macros are deprecated, queue(3) does not say so.
perhaps it would be interesting to mention why circle queues got
deprecated.
Let's go a step further and remove them
On Fri, 12 Sep 2014, Todd C. Miller wrote:
No objection here. I'd go a step farther and nuke all the FOO_END
macros as well. They add nothing and just make code harder to read.
Better?
Index: queue.3
===
RCS file:
was used
in the check, resulting the ALLOCATED FRAG # MARKED FREE warning never
being shown!
This was happy on a clean filesystem (I haven't intentionally corrupted a
filesystem for testing) and I've installed this on my laptop.
oks?
Philip Guenther
Index: dir.c
On Fri, Sep 5, 2014 at 3:54 PM, David Vasek va...@fido.cz wrote:
On Fri, 5 Sep 2014, Philip Guenther wrote:
While di_blocks is unsigned, daddr_t is signed so we cast and print blocks
as long long's.
Noted, thanks.
quad_t id_filesize; /* for DATA nodes, the size of the directory
On Wed, Sep 3, 2014 at 10:51 AM, Stefan Fritsch s...@sfritsch.de wrote:
On Tuesday 02 September 2014 15:22:16, Philip Guenther wrote:
From physio(9):
minphys
A device specific routine called to determine the
maximum transfer size that the device's strategy routine can
for, no?
Not quite what I hoped, but ok. Do you know a good place where this
should be documented?
minphys is currently documented on physio(9)
Philip Guenther
On Tue, Sep 2, 2014 at 2:07 PM, Stefan Fritsch s...@sfritsch.de wrote:
On Tuesday 02 September 2014 13:15:19, Philip Guenther wrote:
On Tue, Sep 2, 2014 at 12:51 PM, Stefan Fritsch s...@sfritsch.de
wrote:
On Monday 01 September 2014 22:24:16, David Gwynne wrote:
i haven't found
On Wed, Aug 27, 2014 at 11:50 AM, Miod Vallat m...@online.fr wrote:
Why not keep the softdep flag when updating rw-ro?
E.g. via mount -ur /usr/obj
fstab(5) already allows ro+softdep.
Because if we were to apply this diff, there would be no way to switch
from rw+softdep to rw.
And what
On Wed, Aug 20, 2014 at 5:23 AM, Visa Hankala v...@hankala.org wrote:
On Tue, Aug 19, 2014 at 10:14:59PM -0700, Philip Guenther wrote:
Can you describe what you're using that needs it?
Well, I am not using it. The code has not been sent to the Attic yet and
I happened to take a look
On Mon, Aug 25, 2014 at 2:01 PM, Kent R. Spillner kspill...@acm.org wrote:
...
Here's an updated diff. This adds a new testcase covering the situation
when b has fewer lines than a (the minus sign also appeared in the wrong
location).
Ok?
Looks good to me. Thanks! ok guenther@
On Mon, Aug 18, 2014 at 8:02 AM, Mike Belopuhov m...@belopuhov.com wrote:
I don't know if we recycle them somehow, but just in case...
--- sys/sys/file.h
+++ sys/sys/file.h
...
-#defineDTYPE_CRYPTO5 /* crypto */
+/* was define DTYPE_CRYPTO5 */
When it goes, this
it does
anything useful.
Maybe the [ -f /sbin/ldconfig ] test should go away completely...
Philip Guenther
On Wed, Jul 9, 2014 at 12:26 PM, Stefan Fritsch s...@sfritsch.de wrote:
On Tuesday 08 July 2014 23:53:21, Mark Kettenis wrote:
...
If we're serious about supporting OpenBSD on (KVM) hypervisors,
something like this makes sense. We tend to try and have a single
kernel that runs on the
On Sun, Aug 17, 2014 at 9:36 PM, Justus Baumgartner medde...@gmail.com
wrote:
Add mention of execvpe() setting errno on failure.
Ooops, yep. Committed, thanks!
Philip Guenther
wrong: the -w
option shouldn't break the output format.
Philip Guenther
On Wed, Jul 30, 2014 at 7:21 PM, Justus Baumgartner medde...@gmail.com
wrote:
On Sun, Jul 27, 2014 at 6:38 AM, Philip Guenther guent...@gmail.com
wrote:
...
That phrasing seems wrong: the environment pointer in execvpe() doesn't
follow the null pointer that terminates the list
On Sun, Aug 10, 2014 at 5:58 PM, Philip Guenther guent...@gmail.com wrote:
...
So, complete diff. This also zaps the unnecessary casts in the .c files.
This has been committed; thank for the patch.
Philip Guenther
On Sun, 10 Aug 2014, Miod Vallat wrote:
If the queue is to be used as a linear list, invoking insque(element,
NULL), where element is the initial element of the queue, shall
initialize the forward and backward pointers of element to null
pointers.
Hmm. Do the vax ASM versions
On Thu, Aug 7, 2014 at 1:07 PM, Maurice Janssen maur...@z74.net wrote:
The -t option for ldattach doesn't seem to work when using an msts
time receiver. The patch below fixes this. I guess this part was
forgotten when msts support was added in r1.4.
Committed.
Philip Guenther
in a header
to define NULL), _do_ please follow up to your original email... ;-)
Philip Guenther
what to do at that point.
Of course, if the OpenSSL security release in checks watch 34 minutes
touches this area
Philip Guenther
which changed different conditions to the same
condition; I missed that those two ended up the same. No semantic
shurg
I'll commit this after unlock...
Philip Guenther
be a better choice, but changing that behavior seems unwise.
Philip Guenther
in execvpe() doesn't
follow the null pointer that terminates the list of arguments. Something
more is needed to cover both.
Philip Guenther
will be available
That is, it would move the default CA loading from the specified either or
both option and nothing failed case to the specified neither option
case. Yes?
Philip Guenther
On Sat, Jul 26, 2014 at 4:56 AM, Mark Kettenis mark.kette...@xs4all.nl
wrote:
Thanks to the evil Intel AMT serial, I've been able to figure out what
made my x220 hang when suspending. Turns out that the fix matthew@
committed almost two weeks ago uncovered another bug that made
is
called. It checked a list of functions that are typically used in error
handling and may set errno: close, fclose, unlink, rmdir, fflush and
kill.
Nice job! Committed (except for the microcode/bwi file where your diff
revealed another existing bug that we can fix at the same time...)
Philip
On Sat, Jul 19, 2014 at 2:36 PM, Doug Hogan d...@acyclic.org wrote:
Strings changed in the warning message.
Committed; thanks!
On Thu, Jul 17, 2014 at 7:27 PM, Doug Hogan d...@acyclic.org wrote:
Ran into this while running the regression suite for another patch.
The regression test for open_memstream causes a SIGBUS because it closes
the stream too early.
Replaced fclose with fflush since they both update pbuf and
On Sat, Jul 19, 2014 at 6:57 PM, Doug Hogan d...@acyclic.org wrote:
On Sat, Jul 19, 2014 at 06:40:54PM -0700, Philip Guenther wrote:
Nice job! Committed (except for the microcode/bwi file where your diff
revealed another existing bug that we can fix at the same time...)
Thanks! Here's
of the declaration specifiers in a declaration is
an obsolescent feature.
and -Wextra complains:
warning: 'static' is not at beginning of declaration
Yep; (revised) diff committed.
Philip Guenther
601 - 700 of 983 matches
Mail list logo
