Author: des
Date: Mon Sep 9 06:02:30 2013
New Revision: 255413
URL: http://svnweb.freebsd.org/changeset/base/255413
Log:
Remove unneeded mappings from libmap32.conf. Move it up one level and
install it on powerpc64 in addition to amd64.
Reviewed by: kib
Approved by: re (blanket)
Ad
Konstantin Belousov writes:
> Dag-Erling Smørgrav writes:
> > How about this, which eliminates the unneeded mappings and installs
> > libmap32.conf on powerpc64 as well?
> > [...]
> This looks fine.
Thanks.
> Are you going to remove COMPAT_32BIT conditional from the rtld.c ?
I wouldn't want to
Author: delphij
Date: Mon Sep 9 05:01:18 2013
New Revision: 255412
URL: http://svnweb.freebsd.org/changeset/base/255412
Log:
In r243868, the error message buffer errmsg have been changed from
an on-stack array to a pointer and therefore sizeof(errmsg) would
become 4 or 8 bytes depending on
Author: np
Date: Mon Sep 9 04:38:57 2013
New Revision: 255411
URL: http://svnweb.freebsd.org/changeset/base/255411
Log:
Rework the tx credit mechanism between the cxgbe/tom driver
and the card. This helps smooth out some burstiness in the
exchange.
Approved by: re (glebius)
Modified
On Sun, Sep 08, 2013 at 09:59:55PM +0200, Dag-Erling Sm??rgrav wrote:
> How about this, which eliminates the unneeded mappings and installs
> libmap32.conf on powerpc64 as well?
>
> Index: etc/Makefile
> ===
> --- etc/Makefile (r
Author: np
Date: Mon Sep 9 00:16:59 2013
New Revision: 255410
URL: http://svnweb.freebsd.org/changeset/base/255410
Log:
Fix a miscalculation that caused cxgbe/tom to auto-increment
a TOE socket's tx buffer size too aggressively.
Approved by: re (delphij)
Modified:
head/sys/dev/cxgbe/
Author: alc
Date: Sun Sep 8 21:30:53 2013
New Revision: 255409
URL: http://svnweb.freebsd.org/changeset/base/255409
Log:
Prior to r254304, we only began scanning the active page queue when the
amount of free memory was close to the point at which we would begin
reclaiming pages. Now, we co
Author: des
Date: Sun Sep 8 20:48:23 2013
New Revision: 255405
URL: http://svnweb.freebsd.org/changeset/base/255405
Log:
Hook host(1) up to the build in the LDNS case.
Approved by: re (blanket)
Modified:
head/usr.bin/Makefile
Modified: head/usr.bin/Makefile
===
On 8 Sep 2013, at 21:18, Dag-Erling Smørgrav wrote:
> David Chisnall writes:
>> Log:
>> On platforms where clang is the default compiler, don't build gcc or
>> libstdc++.
>
> This uncovers what I think is a bug in bsd.prog.mk:
>
> des@ds4 /usr/src/usr.bin% make
> ===> alias (all)
> ===> app
David Chisnall writes:
> Log:
> On platforms where clang is the default compiler, don't build gcc or
> libstdc++.
This uncovers what I think is a bug in bsd.prog.mk:
des@ds4 /usr/src/usr.bin% make
===> alias (all)
===> apply (all)
===> ar (all)
===> asa (all)
===> at (all)
===> atf (all)
===>
How about this, which eliminates the unneeded mappings and installs
libmap32.conf on powerpc64 as well?
Index: etc/Makefile
===
--- etc/Makefile(revision 255393)
+++ etc/Makefile(working copy)
@@ -50,8 +50,8 @@
2013/9/8 Davide Italiano :
> As a side note, I've already discussed with Kostik how to handle this
> in the future and he thinks (and I agree) that the right action there
> is to remove clones and SI_CHEAPCLONE, or implement the real
> SI_CHEAPCLONE semantic.
Yes. Exactly. I've also noticed that i
On 09/08/13 05:05, Dag-Erling Smørgrav wrote:
Dag-Erling SmXXrgrav writes:
Log:
Add a stock libmap32.conf for amd64. The first two lines have no effect
except to document the hardcoded standard library search path for 32-bit
binaries. The third line performs the equivalent substituti
Author: des
Date: Sun Sep 8 19:40:32 2013
New Revision: 255403
URL: http://svnweb.freebsd.org/changeset/base/255403
Log:
Import Magerya Vitaly's ldns-host, and build it instead of the BIND version
in the WITH_LDNS_UTILS case.
Approved by: re (blanket)
Added:
head/contrib/ldns-host/
Author: des
Date: Sun Sep 8 19:39:18 2013
New Revision: 255402
URL: http://svnweb.freebsd.org/changeset/base/255402
Log:
LDNS needs OpenSSL. This wasn't a problem as long as it was only build
statically, since any program using it would have to link with it anyway.
Approved by: re (bla
Author: des
Date: Sun Sep 8 19:29:45 2013
New Revision: 255401
URL: http://svnweb.freebsd.org/changeset/base/255401
Log:
MFH (r255243): create /var/authpf with correct ownership and permissions.
Approved by: re (glebius)
Modified:
releng/9.2/etc/mtree/BSD.var.dist
Directory Properties:
Author: bryanv
Date: Sun Sep 8 19:13:06 2013
New Revision: 255400
URL: http://svnweb.freebsd.org/changeset/base/255400
Log:
Use correct type for the vmx vlan filter table
Approved by: re (glebius, gjb)
Modified:
head/sys/dev/vmware/vmxnet3/if_vmxvar.h
Modified: head/sys/dev/vmware/vmx
Author: des
Date: Sun Sep 8 18:27:26 2013
New Revision: 255398
URL: http://svnweb.freebsd.org/changeset/base/255398
Log:
MFH (r254684, r254955): fix segfault on reload.
MFH (r255156): silence an error message on non-IPv6 interfaces.
Approved by:re (glebius)
Modified:
releng/9.2/us
Author: trociny
Date: Sun Sep 8 18:19:40 2013
New Revision: 255397
URL: http://svnweb.freebsd.org/changeset/base/255397
Log:
Relese the interface in the last.
Reviewed by: glebius
Approved by: re (kib)
Modified:
head/sys/netinet/ip_carp.c
Modified: head/sys/netinet/ip_carp.c
==
Author: kib
Date: Sun Sep 8 17:51:22 2013
New Revision: 255396
URL: http://svnweb.freebsd.org/changeset/base/255396
Log:
Drain for the xbusy state for two places which potentially do
pmap_remove_all(). Not doing the drain allows the pmap_enter() to
proceed in parallel, making the pmap_remov
Author: trociny
Date: Sun Sep 8 17:47:14 2013
New Revision: 255395
URL: http://svnweb.freebsd.org/changeset/base/255395
Log:
MFC r254776:
Make ipfw nat init/unint work correctly for VIMAGE:
* Do per vnet instance cleanup (previously it was only for vnet0 on
module unload, and led
Author: des
Date: Sun Sep 8 17:28:11 2013
New Revision: 255394
URL: http://svnweb.freebsd.org/changeset/base/255394
Log:
MFH (r255243): create /var/authpf with correct ownership and permissions.
Modified:
stable/9/etc/mtree/BSD.var.dist
Directory Properties:
stable/9/etc/mtree/ (props ch
On Sun, Sep 8, 2013 at 6:26 PM, Ed Schouten wrote:
> 2013/9/7 Davide Italiano :
>> Log:
>> Don't clear the unused SI_CHEAPCLONE flag in tap_create()/tuncreate().
>
> Both the tun and the tap driver still use clonelists, right? I'm not
> sure this fix is all right.
>
> --
> Ed Schouten
Please c
Author: des
Date: Sun Sep 8 16:56:17 2013
New Revision: 255393
URL: http://svnweb.freebsd.org/changeset/base/255393
Log:
MFV (r255387): undo autoprop damage
Approved by: re (glebius)
Modified:
head/contrib/openpam/lib/Makefile.am (contents, props changed)
Directory Properties:
head
On Sun, Sep 8, 2013 at 6:24 PM, Ed Schouten wrote:
> 2013/9/7 Davide Italiano :
>> Log:
>> - Use make_dev_credf(MAKEDEV_REF) instead of the race-prone make_dev()+
>> dev_ref() in the clone handlers that still use it.
>> - Don't set SI_CHEAPCLONE flag, it's not used anywhere neither in devfs
Author: markm
Date: Sun Sep 8 16:48:03 2013
New Revision: 255391
URL: http://svnweb.freebsd.org/changeset/base/255391
Log:
Fix verbose output line; needs
Submitted by: Sean Bruno
Approved by: re (glebius)
Modified:
head/sys/dev/random/random_adaptors.c
Modified: head/sys/dev/rando
2013/9/7 Davide Italiano :
> Log:
> Don't clear the unused SI_CHEAPCLONE flag in tap_create()/tuncreate().
Both the tun and the tap driver still use clonelists, right? I'm not
sure this fix is all right.
--
Ed Schouten
___
svn-src-all@freebsd.org ma
2013/9/7 Davide Italiano :
> Log:
> - Use make_dev_credf(MAKEDEV_REF) instead of the race-prone make_dev()+
> dev_ref() in the clone handlers that still use it.
> - Don't set SI_CHEAPCLONE flag, it's not used anywhere neither in devfs
> (for anything real)
Wait. This piece of code is never
Konstantin Belousov writes:
> Any line in libmap.conf cannot add to the standard search path.
> It only translates existing path.
I know.
> And, the standard path is specified by the ELF standard, which we try to
> follow. I.e., hardcoding STD_LIB_PATH in rtld is basically an external
> requirem
On Sun, Sep 08, 2013 at 01:48:54PM +0200, Dag-Erling Sm??rgrav wrote:
> Konstantin Belousov writes:
> > There is no /lib32 on 64bit platforms.
>
> It doesn't exist, but it's supported:
>
> #ifdef COMPAT_32BIT
> #undef STANDARD_LIBRARY_PATH
> #undef _PATH_ELF_HINTS
> #define _PATH_ELF_HINTS
(r255386)
> @@ -38,6 +38,9 @@
> # xargs -n1 | sort | uniq -d;
> # done
>
> +# 20130908: libssh becomes private
> +OLD_LIBS+=usr/lib/libssh.so.5
> +OLD_LIBS+=usr/lib32/libssh.so.5
> # 20130903: gnupatch is no more
> OLD_FILES+=usr/bin/gnupatch
> OLD_FILES+=usr/sh
Konstantin Belousov writes:
> There is no /lib32 on 64bit platforms.
It doesn't exist, but it's supported:
#ifdef COMPAT_32BIT
#undef STANDARD_LIBRARY_PATH
#undef _PATH_ELF_HINTS
#define _PATH_ELF_HINTS "/var/run/ld-elf32.so.hints"
/* For running 32 bit binaries */
#define STANDARD_LIBR
Author: des
Date: Sun Sep 8 11:35:56 2013
New Revision: 255388
URL: http://svnweb.freebsd.org/changeset/base/255388
Log:
MFH (r254684, r254955): fix segfault on reload.
MFH (r255156): silence an error message on non-IPv6 interfaces.
Modified:
stable/9/usr.sbin/rtadvd/config.c
stable/9/us
Author: des
Date: Sun Sep 8 11:21:08 2013
New Revision: 255387
URL: http://svnweb.freebsd.org/changeset/base/255387
Log:
Damn autoprops!
Modified:
vendor/openpam/NUMMULARIA/lib/Makefile.am (contents, props changed)
vendor/openpam/dist/lib/Makefile.am (contents, props changed)
Directory
On Sun, Sep 08, 2013 at 09:46:23AM +, Dag-Erling Smц╦rgrav wrote:
> +/lib /lib32
> +/usr/lib /usr/lib32
There is no /lib32 on 64bit platforms. I do not understand why these pathes
are neeeded, since compat32 rtld has proper default path built in, but if
anything, /l
Dag-Erling SmXXrgrav writes:
> Log:
> Add a stock libmap32.conf for amd64. The first two lines have no effect
> except to document the hardcoded standard library search path for 32-bit
> binaries. The third line performs the equivalent substitution for the
> private library directory.
p
==
--- head/ObsoleteFiles.inc Sun Sep 8 09:46:22 2013(r255385)
+++ head/ObsoleteFiles.inc Sun Sep 8 10:04:26 2013(r255386)
@@ -38,6 +38,9 @@
# xargs -n1 | sort | uniq -d;
# done
+# 20130908: libssh becomes private
+OLD_LIBS+=usr/lib/libssh.so.5
Author: des
Date: Sun Sep 8 09:46:22 2013
New Revision: 255385
URL: http://svnweb.freebsd.org/changeset/base/255385
Log:
Add a stock libmap32.conf for amd64. The first two lines have no effect
except to document the hardcoded standard library search path for 32-bit
binaries. The third lin
Author: des
Date: Sun Sep 8 09:40:23 2013
New Revision: 255384
URL: http://svnweb.freebsd.org/changeset/base/255384
Log:
Create a private library directory (LIBPRIVATEDIR) for libraries which
we don't want to expose but which can't or shouldn't be static.
To mark a library as private, de
On Sun, Sep 08, 2013 at 12:00:33PM +0400, Slawa Olhovchenkov wrote:
S> On Sun, Sep 08, 2013 at 11:22:08AM +0400, Gleb Smirnoff wrote:
S>
S> > On Sun, Sep 08, 2013 at 01:47:28AM +0400, Slawa Olhovchenkov wrote:
S> > S> As we can see by activity in list planed: ipf 5.1, native iSCSI,
S> > S> switch
On Sun, Sep 08, 2013 at 11:22:08AM +0400, Gleb Smirnoff wrote:
> On Sun, Sep 08, 2013 at 01:47:28AM +0400, Slawa Olhovchenkov wrote:
> S> As we can see by activity in list planed: ipf 5.1, native iSCSI,
> S> switch off GCC (with broken by clang kernel part), patch ld.so and many
> others.
> S> Ch
On Sun, Sep 08, 2013 at 11:19:12AM +0400, Gleb Smirnoff wrote:
> On Sun, Sep 08, 2013 at 01:29:57AM +0400, Slawa Olhovchenkov wrote:
> S> I think now -HEAD can't be more risky by one or two patches.
>
> Yes, it can.
>
> Do you run HEAD actually on regular basis when you make such
> an assertion?
On Sun, Sep 08, 2013 at 01:47:28AM +0400, Slawa Olhovchenkov wrote:
S> As we can see by activity in list planed: ipf 5.1, native iSCSI,
S> switch off GCC (with broken by clang kernel part), patch ld.so and many
others.
S> Changed in loader and take out modules from kernel at this time is little-o.
On 07.09.2013 17:30, Mikolaj Golub wrote:
Hi,
On Sat, Aug 24, 2013 at 11:17:25AM +, Andre Oppermann wrote:
Author: andre
Date: Sat Aug 24 11:17:25 2013
New Revision: 254773
URL: http://svnweb.freebsd.org/changeset/base/254773
Log:
Resolve the confusion between the head_list and the hook
On Sun, Sep 08, 2013 at 01:29:57AM +0400, Slawa Olhovchenkov wrote:
S> I think now -HEAD can't be more risky by one or two patches.
Yes, it can.
Do you run HEAD actually on regular basis when you make such
an assertion?
--
Totus tuus, Glebius.
___
svn
45 matches
Mail list logo
