From KernelTools
commit 159d53df088f2a7bf8e99e6c1ee1d3fe56ad6a74
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Thu Mar 22 03:36:02 2007 -0400
[CLEANUP] Cleanup tree and add a HACKING doc.
We've been meaning to cleanup the tree for a while. The git repo should be n
and clean now, and if you follow the HACKING doc, easily build-populated too
Signed-off-by: Jon Masters <jcm@redhat.com>
commit 7277a538f4508e95f8ae90f1c990157344efce60
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Tue Mar 20 15:52:30 2007 -0400
[PATCH] Handle multiple kernels in modinfo.
As part of some other work that needed information about various installed
modules, modinfo was modified to support a '-k' kernel parameter.
Signed-off-by: Keiichiro Tokunaga <ktokunag@redhat.com>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 02d23829e6ad80d1aba1107545c301b293d500d7
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 21:12:13 2007 -0500
process: get in the habit of updating ChangeLog too.
On each update, I should regenerate the configure/configure.ac/Makefile.in
files and update the ChangeLog too just as I bump the revision.
Signed-off-by: Jon Masters <jcm@redhat.com>
commit 2eebb5a5d4f8781ea7bcbd168e587e5d3b6edfe9
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 20:48:30 2007 -0500
build-system: re-add Makefile.in to default distribution
Luiz removed Makefile.in via .gitignore. I understand why this was removed,
but again, people expect to do the configure dance out of the box. I guess
for now, I will rebuild configure and Makefile.in on each update using the
autotools on my devel box - anyone doing something weird can run autotools
for themselves. The advantage here is less "it's broken" email :-)
Signed-off-by: Jon Masters <jcm@redhat.com>
commit 45ee16ab753fb6055eb5692ae04902c0d8c4b9c0
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 20:40:25 2007 -0500
build-system: re-add a default configure script for ease of building.
Luiz took the configure script out via .gitignore. I understand why, but I
think we should have a default script built with a quasi-recent autoconf
in the distribution /because/ people expect to be able to run it. If
they actually need to update configure, they probably know how.
I guess I should automate having configure be rebuilt.
Signed-off-by: Jon Masters <jcm@redhat.com>
commit 8c96fa391eb4257eaaba527f9b0fd5031733b1bb
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 15:20:29 2007 -0500
build-system: update configure.ac's m-i-t version number.
Signed-off-by: Jon Masters <jcm@redhat.com>
commit 899710e9b873950e3c901aef2d32d5466273d345
Merge: 4619ec5... b51060a...
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 15:17:37 2007 -0500
Merge branch 'depmod-misc-fixes' of http://distro2.conectiva.com.br/~lcapitu
commit 4619ec5af5636f413187f9ccff59173a8880d63e
Merge: a5644bb... 057f2d8...
Author: Jon Masters <jcm@perihelion.int.jonmasters.org>
Date: Wed Feb 21 15:17:24 2007 -0500
Merge branch 'misc-fixes' of http://distro2.conectiva.com.br/~lcapitulino/gi
commit a5644bb415e4bf53d6e13739b67c783a3ae4d119
Author: Tobias Stoeckmann <tobias@bugol.de>
Date: Wed Feb 21 15:09:59 2007 -0500
[PATCH] Correct strdup error handling and insmod.static handling
There is one more problem about memory allocation in insmod: strdup("") coul
fail. If this happens, realloc() accesses later on strlen(options). A call l
strlen(NULL) segfaults on my system (glibc 2.4).
Another thing worth to mention is that strstr() on argv[0] raises another
little problem - or perhaps intented behaviour:
/my/path/insmod.static/../../../sbin/insmod
(or anything else that contains insmod.static)
This call results in using old insmod. Fix for this is to check only the str
after last occurence of "/".
Signed-off-by: Jon Masters <jcm@redhat.com>
Signed-off-by: Tobias Stoeckmann <tobias@bugol.de>
commit b51060a0c71317be3f5e948efcbae0cca128879c
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Feb 10 12:51:33 2007 -0200
depmod: Allocate do_module()'s newpath in the stack
No reason to allocate in the heap.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit eb70ebcaabe3a5f6d7e676253b9894b793000a87
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 19:41:23 2007 -0200
depmod: Add checks to strdup()'s return value
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 8de7f234a9265c6cbec8a9a244580a354164b7d5
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 18:11:25 2007 -0200
depmod: do_module() leak fix
When a higher priority module is found, we have to free the memory
allocated by the module we're about to remove from the list.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 921c49d670e70affaa1371de4301b8ab306b58e0
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 17:10:47 2007 -0200
depmod: read_config_file() should abort on error
read_config_file() silently skips fopen()'s errors when opening
config files.
It's ok to ignore ENOENT, since config files are optional. But any
other error should be reported as usual.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit aeb9d5b558c1e2b521a6148f89eeed61acbbbb31
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 16:17:14 2007 -0200
depmod.8: Add -C text
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit a2bb6381f25bcf24a704b68e48d890b8047b64bc
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 15:41:55 2007 -0200
depmod.8: Remove duplicated 'basedir' from -b option
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 750bd14ebf77bfdfd84dcef541c9bf87607460a0
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Fri Feb 9 15:38:11 2007 -0200
depmod.8: List options in alphabetical order
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 057f2d825dc873527d553e73e1fa4a579b9433bb
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:50 2006 -0200
depmod: load_system_map() minor optimization.
The load_system_map() function calls strlen() to get the length of the
"__ksymtab_" string twice as follows:
* For each line of the System.map file
* For each symbol which starts with "__ksymtab_"
It's smarter to call it once, before the loop is entered. That's what
this patch does.
Addittionally a string duplication is also fixed.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 6c7e2fb62b120f55d8bee65ba673aeeeb023c1a0
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:50 2006 -0200
depmod: Fixes load_system_map() NULL pointer.
depmod's load_system_map() function assumes that System.map file lines
aways ends with newlines ('\n').
However, in some Linux distributions it may not be true, which will lead
to a Segmentation Fault.
To fix the problem we have to check 'ptr' right after we look for the
newline character using strchr().
This is bug #2, report and suggested fix by shenlinf@cn.ibm.com.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 0375dce5857ce8c2b3044f5d05252f65062f2980
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:50 2006 -0200
repo: More files to .gitignore
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 8ff372ddd54d253b4dd62ff418eeccf856762c41
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:50 2006 -0200
build-system: updates configure.ac's m-i-t version number.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit a30f24cbd26315f0708f7a15c9efed89c88db6f0
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:49 2006 -0200
build-system: renames configure.in file.
The .in extension is deprecated, the .ac one is preferred now.
More information can be found in the autoconf manual.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit 28e17a33b5fba6fc19ad139810b196c4e9d35950
Author: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Date: Sat Dec 23 12:04:49 2006 -0200
build-system: Removes autotools generated files.
These files should not be added to the repository because they're generated
automatically by the autotools.
More information can be found in the autoconf and automake manuals.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
commit b0c1e293dc962368117875c3684cee6362a79238
Author: Ladislav Michl <ladis@linux-mips.org>
Date: Mon Jan 15 17:23:39 2007 +0100
[PATCH] [Fwd: module-init-tools-3.3-pre1 - config.sub doesn't recognize uCli
On Mon, Jan 08, 2007 at 11:01:34PM +1100, Rusty Russell wrote:
> Hi Ladislav,
>
> Jon Masters is the new maintainer, fwding to him...
Hi Rusty,
thanks for information. For new maintainer's convenience a patch updating
config.guess and confing.sub to the lastest and greatest version as
provided by FSF can be found at the end of this email.
Thanks,
ladis
> -------- Forwarded Message --------
> From: Ladislav Michl <ladis@linux-mips.org>
> To: Rusty Russell <rusty@rustcorp.com.au>
> Subject: module-init-tools-3.3-pre1 - config.sub doesn't recognize
> uClibc based system
> Date: Mon, 8 Jan 2007 11:17:10 +0100
>
> Hi,
>
> would you mind to update config.sub to some more recent version? The one
> with timestamp 2005-07-08 handles my gcc-3.4.3-uclibc-0.9.27 toolchain
> correctly. Current one fails with:
> checking target system type... Invalid configuration `arm-linux-uclibc': s
> configure: error: /bin/sh ./config.sub arm-linux-uclibc failed
>
> Thanks,
> ladis
Update config.guess and confing.sub to the lastest version as
provided by FSF
Signed-off-by Ladislav Michl <ladis@linux-mips.org>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit f22bfcb02ec4651726ce8ada0355e70fb0667121
Author: Sergey Vlasov <vsu@altlinux.ru>
Date: Wed Jan 3 18:13:06 2007 +0300
modprobe: Fix format string bug in do_command()
gcc produces the following warning on x86_64:
modprobe.c:786: warning: field precision should have type 'int', but argumen
This bug is currently harmless on x86_64, because parameters are
passed in registers, but could cause misbehavior on other 64-bit
platforms.
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
commit 758caafc788fcff9808df0aa12523af38ee11753
Author: Sergey Vlasov <vsu@altlinux.ru>
Date: Wed Jan 3 18:01:40 2007 +0300
modprobe: Do not pass cmdline_opts to subordinate modules
Options from the modprobe command line should be passed only to the
install/remove command for the module specified on the command line,
and not to modules which the requested module depends on.
Fixes http://bugzilla.kerneltools.org/show_bug.cgi?id=1
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
commit 0373793164656126f9934334b60fcfc5554506c6
Author: Sergey Vlasov <vsu@altlinux.ru>
Date: Tue Jan 2 20:44:00 2007 +0300
Add .gitignore file to unclutter "git status" output
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
commit 15150274c3393a18b6b170e3268a61389a214ac7
Author: Sergey Vlasov <vsu@altlinux.ru>
Date: Tue Jan 2 20:35:59 2007 +0300
Remove more garbage from the repository
config.status is generated by ./configure and should not be stored
in the repository; autom4te.cache/* and .deps/* also should not be
there.
Signed-off-by: Sergey Vlasov <vsu@altlinux.ru>
commit a7d7fa164af4e20c1727933d975509e5a4e6cb59
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:25:36 2006 -0500
I previously applied a patch to add --dump-modversions to modprobe but this
not covered by the usage instructions on that command.
commit 079fb7b6a2c5f47f08211ef42e5a77416fa1cd43
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:22:01 2006 -0500
-------- Forwarded Message --------
From: Steven J. Hill <sjhill@realitydiluted.com>
To: rusty@rustcorp.com.au
Subject: Remove usage of legacy 'index' function.
Date: Mon, 04 Dec 2006 20:04:00 -0600
Rusty,
Attached is a patch to remove usage of the legacy 'index' function. Thanks.
-Steve
commit a005f8388c5fec4e84b69cc88dfcf0f5f2ff1376
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:17:36 2006 -0500
Subject: modinfo: Memory leak fixes.
Only happens when more than one module is passed in the command-line.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit daa5d47c1ddb06a8f2382ad0038adf7ec394746c
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:16:35 2006 -0500
Subject: rmmod: Changes exit() return value.
In the current implementation if rmmod fails, it will return the last
error it got.
This is not very good because it's not consistent and if the error is
negative, we'll have not standard return error codes from rmmod (eg, 255)
This patch changes that behaivor by always returning 1 if at least one
of the modules to remove failed.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 51ad6022167afba83e6da9e6671c0b1e906aef79
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:15:52 2006 -0500
Subject: zlibsupport: Adds missing fstat() return value check.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit aa030a80e4e48f03fdd7b29b2e30a344027028e6
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:15:01 2006 -0500
Subject: zlibsupport: Adds missing realloc() check.
Without this check we may leak memory.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit b052df7ecccb06bd16f824316185b47b6e1fb38d
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:14:01 2006 -0500
Subject: depmod: adds missing fclose() in load_system_map().
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 8301378badce21e71c4fcca98a05f8a56febc697
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:12:59 2006 -0500
Subject: depmod: adds missing options to manpage.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 6df350fc9913b5b2fe58c69f416ae0060700020f
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:11:58 2006 -0500
Subject: depmod: print_usage() minor fixes.
Adds missing options, and kernel modules now are .ko.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 5244693471120f98230f4c3ad6e8f200ce86deb9
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:10:47 2006 -0500
Subject: insmod: Fixes errno override.
If the read() call in grab_file() fails, errno can be overriden by the
call to close().
It's better to stay with read()'s error because it probably has the
explanation we're interested in.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 2832a37703706d362a77969a2b13c5e4fea57d26
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:05:17 2006 -0500
Subject: insmod: Adds missing realloc() check.
Without this check we'll get a NULL pointer dereference if realloc() fails.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit 556e85135b4751e9de1bda99249a2c6b63de3c81
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:03:52 2006 -0500
Subject: insmod: Fixes grab_file() NULL pointer.
If realloc() fails, it will return NULL and we'll pass 'NULL + *size' to
read().
This patch fixes it by adding the proper check and a new variable to store
the realloc() return value, to avoid leaking memory.
Signed-off-by: Luiz Fernando N. Capitulino <lcapitulino@mandriva.com.br>
Signed-off-by: Jon Masters <jcm@jonmasters.org>
commit d33fa4603a2e76c912fcbe54e967b3cdebf4a32b
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:02:19 2006 -0500
Luiz Fernando N. Capitulino mentioned that some dud files were left after my
last release. Sorry about that; the files have now been cleaned out.
commit 9ba2d09c2261c04fa72dcf614497b9f622fada6c
Author: Jonathan Masters <jcm@jcm.boston.redhat.com>
Date: Fri Dec 8 05:01:05 2006 -0500
Begin tracking post-3.3-pre3 development with git.
Thanks to Luiz Fernando N. Capitulino for his mail, suggesting that it might
really good to begin tracking development with a proper SCM. This is especia
important now that we're trying to actively encourage greater involvement in
development once more. I'm looking for more patches and git trees to merge!
