Re: CVS commit: src/tests/kernel
On Sun, Nov 06, 2016 at 11:56:31PM +0700, Robert Elz wrote: > Date:Sun, 6 Nov 2016 16:24:16 + > From:"Kamil Rytarowski" <ka...@netbsd.org> > Message-ID: <20161106162416.95d77f...@cvs.netbsd.org> > > | assert_pid1 asserts that non-root user cannot attach to PID 1 as it is the > | /dev/init process. This tests is skipped if run as root. > > There's no need to skip it, just > > child=fork(); /* err if -1 */ > if (child == 0) { > (void)setuid(10); > if (ptrace(.) < 0) > _exit(errno); > else > _exit(0); > } > waitpid(child, , 0); > /* and check status */ > > If you're root, the setuid() works, and the child isn't root any more. > if you happened to be uid(10), the setuid() is a no-op, if you were some > other user the setuid() fails, but you don't care. Or use dedicated ATF properties : atf_tc_set_md_var(tc, "require.user", "unprivileged"); Documented in atf-test-case(4). -- Nicolas Joly Cluster & Computing Group Biology IT Center Institut Pasteur, Paris.
Re: CVS commit: src/sys/miscfs/procfs
On Sat, Jul 12, 2014 at 08:11:46PM +1000, matthew green wrote: Nicolas Joly writes: Module Name:src Committed By: njoly Date: Sat Jul 12 09:58:39 UTC 2014 Modified Files: src/sys/miscfs/procfs: procfs_linux.c Log Message: Use kproc2 to provide sensible informations for /proc/pid/stat. hmm, kernel code should really avoid using the userland compat layer. you're processing something from one known layout to another known layout.. userland needs it because it only knows one of them, which has a known forwards/backwards compat layout. I was trying to avoid duplicating all the needed computation ... /proc/pid/stat is another exact version (kernel) of the userland processes infos. Do we need another intermediate layer to fill a kernel only structure which can then be easily translated to kproc2 ? -- Nicolas Joly Biology IT Center Institut Pasteur, Paris.
Re: CVS commit: src/tests/fs/common
On Sat, Mar 16, 2013 at 05:25:00AM +, Julio M. Merino Vidal wrote: Module Name: src Committed By: jmmv Date: Sat Mar 16 05:25:00 UTC 2013 Modified Files: src/tests/fs/common: h_fsmacros.h Log Message: Mark the zfs tests as requiring root. This is wrong. The zfs tests already use rump so they should not require root. However, I've already spent much more time than I wanted trying to figure out why that's the case without much luck. If you can find why, just remove this hack. The zfs tests are failing when run as an unprivileged user in thread_create() ... because default limit for threads (160) is too low ! When run as root, the limit value (2500) seems to be high enough. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris.
Re: CVS commit: src/etc/rc.d
On Fri, Dec 14, 2012 at 06:42:26PM +, Alan Barrett wrote: Module Name: src Committed By: apb Date: Fri Dec 14 18:42:25 UTC 2012 Modified Files: src/etc/rc.d: random_seed Log Message: Avoid using programs from /usr/bin. This should fix PR 47326. - no need for dirname, because df -G can take a file name directly. - replace use of awk with a shell while read loop. - replace use of stat -s with ls -ldn. - no need for tail now that the use of stat has changed. While here, also add some shell quotes and improve the grammar in a comment. With this change, i do still see a bootstrap problem when ${random_file} doesn't exists ... njoly@lynche [~]# ls -l /var/db/entropy-file ls: /var/db/entropy-file: No such file or directory root@lynche [~]# /etc/rc.d/random_seed stop df: /var/db/entropy-file: No such file or directory Actually, if the file does not exists it will fail to create one. random_save() { oum=$(umask) umask 077 rm -Pf ${random_file} if ! fs_safe ${random_file}; then return 1 fi if rndctl -S ${random_file}; then echo Saved entropy to disk. fi } First, rm(1) is called before fs_safe() check which will always fail. Even with the rm call commented or better moved just before rndctl, the fs_safe call will fail in df(1) if file does not already exists. Chicken or Egg ... -- Nicolas Joly Biology IT Center Institut Pasteur, Paris.
Re: CVS commit: src/lib/libc/arch/alpha/gen
On Wed, Mar 21, 2012 at 10:42:58PM +0200, Alan Barrett wrote: On Wed, 21 Mar 2012, Havard Eidnes wrote: Modified Files: src/lib/libc/arch/alpha/gen: fpgetround.c fpsetround.c Log Message: Add some casts to get rid of bitwise op on signed value is non-portable warning from lint. I see no bitwise ops on signed values here. -return ((fpcrval.u64 58) 0x3); +return ((fp_rnd)(fpcrval.u64 58) 0x3); fpcrval.u64 is uint64_t. After the integer promotions, it's still uint64_t (unless that's smaller than int, which is not the case for any existing NetBSD port). After 58, it's still uint64_t. 0x3 is a signed int, but the usual arithmetic conversions should convert it to uint64_t. The commit message reference the wrong lint warning. /local/src/NetBSD/src/lib/libc/arch/alpha/gen/fpgetround.c(61): warning: conversion from 'unsigned long' to 'enum unnamed' may lose accuracy [132] /local/src/NetBSD/src/lib/libc/arch/alpha/gen/fpsetround.c(61): warning: conversion from 'unsigned long' to 'enum unnamed' may lose accuracy [132] -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/sbin/ifconfig
On Sat, Mar 17, 2012 at 06:53:31PM +0200, Jukka Ruohonen wrote: On Fri, Mar 16, 2012 at 10:25:08PM -0400, Christos Zoulas wrote: Module Name:src Committed By: christos Date: Sat Mar 17 02:25:08 UTC 2012 Modified Files: src/sbin/ifconfig: af_inetany.c Log Message: PR/43141: Tobias Nygren: Print an error on unknown interfaces. Now this is trival and it is fixed. But I guess there is a bug here in ATF. The test only checks that the exit code is not zero: atf_check -s not-exit:0 ifconfig nonexistent0 1.2.3.4/24 But why is stderr being involved? atf_check defaults to '-s exit:0 -o empty -e empty'. Adding '-e ignore' should do the trick. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/external/bsd/mdocml/dist
On Sun, Dec 18, 2011 at 04:58:02PM +0100, Joerg Sonnenberger wrote: On Sun, Dec 18, 2011 at 01:40:08PM +, Nicolas Joly wrote: Module Name:src Committed By: njoly Date: Sun Dec 18 13:40:08 UTC 2011 Modified Files: src/external/bsd/mdocml/dist: arch.in Log Message: Lowercase arch strings, and add missing ones. The former is wrong and should have at least been discussed on the mdocml mailing lists... Really ? Doesn't arch.in, like other .in files, intended for operating system customization ? -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: amd64 direct map causing crashes at boot (Re: CVS commit: src/sys/arch)
On Fri, Dec 09, 2011 at 02:41:16PM +0100, Manuel Bouyer wrote: On Thu, Dec 08, 2011 at 11:24:35PM +0100, Nicolas Joly wrote: On Thu, Dec 08, 2011 at 11:21:31PM +0100, Joerg Sonnenberger wrote: On Thu, Dec 08, 2011 at 11:10:41PM +0100, Nicolas Joly wrote: I do have two machines that fail to boot with this change. 1) A Dell Optiplex 740 that reboots early after loading the kernel. Includes 1 AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ CPU. 2) A Supermicro H8DG6-F motherboard (only 1 socket populated) later on boot mostly when sshd start (or fc-cache is updated). Includes 1 AMD Opteron(tm) Processor 6128A CPU Can you try to disable the use of 1GB pages? Already tested for the 6128A, still no luck. Can you print the content of mem_clusters[] and mem_cluster_count ? I wonder if we're missing some memory ... Here's the DEBUG_MEMLOAD output that Chuck already asked BIOS MEMORY MAP (12 ENTRIES): addr 0x0 size 0x93800 type 0x1 addr 0x93800 size 0xc800 type 0x2 addr 0xe6000 size 0x1a000 type 0x2 addr 0x10 size 0xb7d8 type 0x1 addr 0xb7e8e000 size 0x2000 type 0x9 addr 0xb7e9 size 0x24000 type 0x3 addr 0xb7eb4000 size 0x2c000 type 0x4 addr 0xb7ee size 0x2 type 0x2 addr 0xb7f0 size 0x10 type 0x2 addr 0xe000 size 0x1000 type 0x2 addr 0xffe0 size 0x20 type 0x2 addr 0x1 size 0x14800 type 0x1 loading first16q 0xf000-0x93000 (0xf-0x93) loading first4gq 0x133e000-0xb7e8 (0x133e-0xb7e80) loading default 0x1-0x24800 (0x10-0x248000) Does this fit your needs or do you need mem_clusters output too ? -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: amd64 direct map causing crashes at boot (Re: CVS commit: src/sys/arch)
On Fri, Dec 09, 2011 at 03:30:38PM +0100, Manuel Bouyer wrote: I'd like to see what mem_clusters[] looks like, especially mem_clusters[mem_cluster_cnt - 1] Added the printfs at the end of initx86_parse_memmap(), and got this: mem_cluster_cnt = 3 0 addr 0x0 size 0x93000 1 addr 0x10 size 0xb7d8 2 addr 0x1 size 0x14800 -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: amd64 direct map causing crashes at boot (Re: CVS commit: src/sys/arch)
On Thu, Dec 08, 2011 at 07:47:07AM -0800, Chuck Silvers wrote: I got another report yesterday morning that this is causing problems. from the info I've gotten from the other person so far, I'm at a loss to explain why this doesn't work for some people. I've added back some code to allow building a kernel with __HAVE_DIRECT_MAP disabled, just comment out that block in arch/amd64/include/types.h. That was me. I do have two machines that fail to boot with this change. 1) A Dell Optiplex 740 that reboots early after loading the kernel. Includes 1 AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ CPU. 2) A Supermicro H8DG6-F motherboard (only 1 socket populated) later on boot mostly when sshd start (or fc-cache is updated). Includes 1 AMD Opteron(tm) Processor 6128A CPU Since then i checked on a 3rd machine which has no problem so far. 3) A Sony VAIO VGNBZ12-VN laptop. Includes 1 Ïntel(R) Core(TM)2 Duo CPU P8600 @ 2.40GHz CPU. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: amd64 direct map causing crashes at boot (Re: CVS commit: src/sys/arch)
On Thu, Dec 08, 2011 at 11:21:31PM +0100, Joerg Sonnenberger wrote: On Thu, Dec 08, 2011 at 11:10:41PM +0100, Nicolas Joly wrote: I do have two machines that fail to boot with this change. 1) A Dell Optiplex 740 that reboots early after loading the kernel. Includes 1 AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ CPU. 2) A Supermicro H8DG6-F motherboard (only 1 socket populated) later on boot mostly when sshd start (or fc-cache is updated). Includes 1 AMD Opteron(tm) Processor 6128A CPU Can you try to disable the use of 1GB pages? Already tested for the 6128A, still no luck. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: amd64 direct map causing crashes at boot (Re: CVS commit: src/sys/arch)
On Thu, Dec 08, 2011 at 11:17:54PM +0100, Manuel Bouyer wrote: On Thu, Dec 08, 2011 at 11:10:41PM +0100, Nicolas Joly wrote: That was me. I do have two machines that fail to boot with this change. 1) A Dell Optiplex 740 that reboots early after loading the kernel. Includes 1 AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ CPU. 2) A Supermicro H8DG6-F motherboard (only 1 socket populated) later on boot mostly when sshd start (or fc-cache is updated). Includes 1 AMD Opteron(tm) Processor 6128A CPU Do they announce 1GB pages support ? Only the 6128A Opteron. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/sys/arch/alpha/alpha
On Tue, Jul 26, 2011 at 05:27:38PM +0200, Joerg Sonnenberger wrote: On Tue, Jul 26, 2011 at 02:59:04PM +, Nicolas Joly wrote: Module Name:src Committed By: njoly Date: Tue Jul 26 14:59:03 UTC 2011 Modified Files: src/sys/arch/alpha/alpha: autoconf.c Log Message: Make atoi func static, and constify Shouldn't this use strtol to implement atoi? We don't have strtol for kernel use. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/sys/arch/alpha/alpha
On Tue, Jul 26, 2011 at 06:45:31PM +0200, Joerg Sonnenberger wrote: On Tue, Jul 26, 2011 at 06:19:21PM +0200, Nicolas Joly wrote: On Tue, Jul 26, 2011 at 05:27:38PM +0200, Joerg Sonnenberger wrote: On Tue, Jul 26, 2011 at 02:59:04PM +, Nicolas Joly wrote: Module Name:src Committed By: njoly Date: Tue Jul 26 14:59:03 UTC 2011 Modified Files: src/sys/arch/alpha/alpha: autoconf.c Log Message: Make atoi func static, and constify Shouldn't this use strtol to implement atoi? We don't have strtol for kernel use. Then use strtoll. We have at least some conversion code in libkern. Looking how many atoi() copies we have in tree, does it make sense to add the following at the end of libkern.h LIBKERN_INLINE int atoi(const char *) __unused; #ifdef LIBKERN_BODY LIBKERN_INLINE int atoi(const char *s) { return (int)strtoll(s, NULL, 10); } #endif and rename the conflicting ones (dm, ptyfs and procfs). njoly@lanfeust [NetBSD/src] find sys -type f -name '*.[ch]'| xargs grep 'atoi(.*char' sys/arch/alpha/alpha/autoconf.c:static int atoi(const char *); sys/arch/alpha/alpha/autoconf.c:atoi(const char *s) sys/arch/amiga/dev/ite.c:inline static int atoi(const char *); sys/arch/amiga/dev/ite.c:atoi(const char *cp) sys/arch/atari/dev/ite.c:static inline int atoi(const char *); sys/arch/hp300/stand/common/devopen.c:atoi(char *cp) sys/arch/hp300/stand/common/samachdep.h:int atoi(char *); sys/arch/i386/stand/lib/bootmenu.c:atoi(const char *in) sys/arch/i386/stand/lib/bootmenu.h:int atoi(const char *); sys/arch/next68k/next68k/autoconf.c:static int atoi(const char *); sys/arch/next68k/next68k/autoconf.c:atoi(const char *s) sys/arch/next68k/stand/boot/devopen.c:int atoi(const char *); sys/arch/next68k/stand/boot/devopen.c:atoi(const char *cp) sys/arch/prep/stand/boot/devopen.c:int atoi(char *); sys/arch/prep/stand/boot/devopen.c:atoi(char *cp) sys/arch/rs6000/stand/boot/devopen.c:int atoi(char *); sys/arch/rs6000/stand/boot/devopen.c:atoi(char *cp) sys/arch/vax/boot/boot/devopen.c:intatoi(char *); sys/arch/x68k/dev/ite.c:inline static int atoi(const char *); sys/arch/x68k/dev/ite.c:atoi(const char *cp) sys/arch/zaurus/stand/zboot/bootmenu.c:atoi(const char *in) sys/arch/zaurus/stand/zboot/bootmenu.h:int atoi(const char *); sys/compat/ndis/subr_ntoskrnl.c:atoi(const char *str) sys/dev/dm/dm.h:uint64_t atoi(const char *); sys/dev/dm/dm_target_linear.c:atoi(const char *s) sys/dev/raidframe/rf_utils.c:rf_atoi(char *p) sys/dev/raidframe/rf_utils.h:int rf_atoi(char *p); sys/fs/ptyfs/ptyfs_vnops.c:static int atoi(const char *, size_t); sys/fs/ptyfs/ptyfs_vnops.c:atoi(const char *b, size_t len) sys/miscfs/procfs/procfs_vnops.c:static int atoi(const char *, size_t); sys/miscfs/procfs/procfs_vnops.c:atoi(const char *b, size_t len) -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/sys/arch/alpha/alpha
On Wed, Jul 27, 2011 at 02:33:24AM +0900, Izumi Tsutsui wrote: Looking how many atoi() copies we have in tree, does it make sense to add the following at the end of libkern.h Please don't use it for #ifdef _STANDALONE ones. libc's strtoll() is larger than dumb atoi()s. Fine. No need then, i'll only update the alpha one. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/distrib/sets/lists
On Mon, May 30, 2011 at 09:17:52AM +, Christos Zoulas wrote: In article 9fd1caa9-1c17-4198-a9cc-57f4de798...@gmail.com, Adam Hamsik haa...@gmail.com wrote: libdns was minor bumped after last bind update you should mark old entries obsolete and add new ones. Same should be done for all 64 archs. I think that this is not true for shared libraries; this happens automatically now. Perhaps not for .debug libraries? If that's the case it should be fixed. I tweaked postinstall, quite some time ago, to handle debug libraries too. AFAIK, it works. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/distrib/sets/lists
On Mon, May 30, 2011 at 12:52:57PM +0200, Adam Hamsik wrote: On May,Monday 30 2011, at 11:23 AM, Nicolas Joly wrote: On Mon, May 30, 2011 at 09:17:52AM +, Christos Zoulas wrote: In article 9fd1caa9-1c17-4198-a9cc-57f4de798...@gmail.com, Adam Hamsik haa...@gmail.com wrote: libdns was minor bumped after last bind update you should mark old entries obsolete and add new ones. Same should be done for all 64 archs. I think that this is not true for shared libraries; this happens automatically now. Perhaps not for .debug libraries? If that's the case it should be fixed. I tweaked postinstall, quite some time ago, to handle debug libraries too. AFAIK, it works. I was not aware of that, it might be a good idea to document this difference in set lists file which is different. e.g. in shl.mi now. Actually, only minor number changes are handled automatically for both shared/debug libraries. I do have, on my TODO list, to add a postinstall majorlib target (which will default to off) to ease builds when major is incremented. This is a real pain for MKDEBUG builds ... Just like kernel version bump versus module tree. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/distrib/sets/lists
On Mon, May 30, 2011 at 03:37:13PM +0200, Nicolas Joly wrote: On Mon, May 30, 2011 at 12:52:57PM +0200, Adam Hamsik wrote: On May,Monday 30 2011, at 11:23 AM, Nicolas Joly wrote: On Mon, May 30, 2011 at 09:17:52AM +, Christos Zoulas wrote: In article 9fd1caa9-1c17-4198-a9cc-57f4de798...@gmail.com, Adam Hamsik haa...@gmail.com wrote: libdns was minor bumped after last bind update you should mark old entries obsolete and add new ones. Same should be done for all 64 archs. I think that this is not true for shared libraries; this happens automatically now. Perhaps not for .debug libraries? If that's the case it should be fixed. I tweaked postinstall, quite some time ago, to handle debug libraries too. AFAIK, it works. I was not aware of that, it might be a good idea to document this difference in set lists file which is different. e.g. in shl.mi now. Actually, only minor number changes are handled automatically for both shared/debug libraries. I do have, on my TODO list, to add a postinstall majorlib target (which will default to off) to ease builds when major is incremented. This is a real pain for MKDEBUG builds ... Just like kernel version bump versus module tree. Well, not MKDEBUG, but MKUPDATE builds. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/distrib/sets/lists/comp
On Mon, May 09, 2011 at 03:17:54PM +0200, Adam Hamsik wrote: I'm getting this error with sources updated couple of minutes ago === 24 extra files in DESTDIR = Files in DESTDIR but missing from flist. File is obsolete or flist is out of date ? -- ./usr/include/perfuse.h ./usr/lib/i386/libperfuse.a ./usr/lib/i386/libperfuse.so ./usr/lib/i386/libperfuse.so.0 ./usr/lib/i386/libperfuse.so.0.0 ./usr/lib/i386/libperfuse_p.a ./usr/lib/i386/libperfuse_pic.a ./usr/lib/libperfuse.a ./usr/lib/libperfuse.so ./usr/lib/libperfuse.so.0 ./usr/lib/libperfuse.so.0.0 ./usr/lib/libperfuse_p.a ./usr/lib/libperfuse_pic.a ./usr/libdata/debug/usr/lib/i386/libperfuse.so.0.0.debug ./usr/libdata/debug/usr/lib/libperfuse.so.0.0.debug ./usr/libdata/debug/usr/sbin/perfused.debug ./usr/libdata/lint/llib-lperfuse.ln ./usr/sbin/perfused ./usr/share/man/cat3/libperfuse.0 ./usr/share/man/cat8/perfused.0 ./usr/share/man/html3/libperfuse.html ./usr/share/man/html8/perfused.html ./usr/share/man/man3/libperfuse.3 ./usr/share/man/man8/perfused.8 = end of 24 extra files === All these entries have the `perfuse' keyword, which means that they will be selected for MKPERFUSE=yes builds only. But MKPERFUSE variable is not defined anywhere ... Do we need to add another optional component here (Makefiles will then need some tweaks), or should the keywords need to be wiped out from lists ? Manu ? -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/tests/syscall
On Mon, Apr 04, 2011 at 08:32:14AM +0300, Jukka Ruohonen wrote: On Mon, Apr 04, 2011 at 01:59:50AM +, Christos Zoulas wrote: Check for getuid(2) == 0. Why not setuid() to fix it? That would indeed be the right thing to do. However, I am not quite sure how atf(7) actually works. It seems to create some temporary files such that simply changing the real and effective UID in the middle of a test gives tc-se:FATAL ERROR: Cannot create results file '/tmp/atf-run.21878a/tcr': Permission denied I am not sure whether this particular test deserves a more complex solution. Adding the following to the testcase header should do the trick: atf_tc_set_md_var(tc,require.user,unprivileged); In that case, if run as root, it will lower its privileges to the unprivileged user _atf before running the test. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/tests/syscall
On Mon, Apr 04, 2011 at 10:18:52AM +0300, Jukka Ruohonen wrote: On Mon, Apr 04, 2011 at 09:05:19AM +0200, Nicolas Joly wrote: Adding the following to the testcase header should do the trick: atf_tc_set_md_var(tc,require.user,unprivileged); In that case, if run as root, it will lower its privileges to the unprivileged user _atf before running the test. Thanks, that indeed does the job. If I may express my humble opinion, there seems to be currently way too much weird macro-magic and undocumented calls in atf(7)... Meta-data properties, like this one, are already described in atf-test-case(4) man page. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/tests
On Fri, Mar 25, 2011 at 10:00:32AM +, Jukka Ruohonen wrote: Module Name: src Committed By: jruoho Date: Fri Mar 25 10:00:32 UTC 2011 Modified Files: src/tests/include/sys: t_bitops.c src/tests/lib/libm: t_ceil.c t_floor.c Log Message: Bluntly skip the tests that fail under QEMU. XXX: system(3) is used for this? Skipping them is not the right solution IMO. I would expect them to fail if run under QEMU: if (system(cpuctl identify 0 | grep -q QEMU) == 0) atf_tc_expect_fail(PR misc/44767: floating point support broken in QEMU); To generate a diff of this commit: cvs rdiff -u -r1.4 -r1.5 src/tests/include/sys/t_bitops.c cvs rdiff -u -r1.1 -r1.2 src/tests/lib/libm/t_ceil.c cvs rdiff -u -r1.2 -r1.3 src/tests/lib/libm/t_floor.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/tests
On Fri, Mar 25, 2011 at 12:15:08PM +0200, Jukka Ruohonen wrote: On Fri, Mar 25, 2011 at 11:09:26AM +0100, Nicolas Joly wrote: Skipping them is not the right solution IMO. I have no strong opinion about this. But to me it seems that these bugs/tests are really testing an emulator instead of NetBSD itself. We do exercise with tests the tuple of hardware (or emulator) and OS. Tests should show existing problems on the tested infrastructure. And it could be a good way to know if/when QEMU fixed the reported problems, without following it too closely ... -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris.
Re: CVS commit: src/external/bsd/atf/dist/atf-sh
On Fri, Nov 26, 2010 at 12:04:36PM +, Antti Kantee wrote: Module Name: src Committed By: pooka Date: Fri Nov 26 12:04:36 UTC 2010 Modified Files: src/external/bsd/atf/dist/atf-sh: atf-check.cpp Log Message: Remove spammy (debug?) prints. took ages to figure out they were not coming from my application... XXX: the memcmp below looks suspicious XXX: Likewise for error messages. -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
Re: CVS commit: src/lib/libc/stdlib
On Thu, Sep 30, 2010 at 02:11:19PM +0100, Matthias Scheler wrote: On Thu, Sep 30, 2010 at 03:05:26PM +0200, Nicolas Joly wrote: One possibility could be to not free memory at all in setenv, but only with unsetenv. I'm not convinced about that. IMHO, it's a programmer error to call setenv more than once on the same variable without a corresponding unsetenv in between (just like malloc()/free() behaviour). Is there any standard which backs this statement? Not that i know. Otherwise, if i understand things correctly, it's more putenv that shoud work that way. Our putenv(3) is implemented via setenv(3). That may be part of the problem (at least for zsh 4.2). http://www.opengroup.org/onlinepubs/009695399/functions/putenv.html From the OpenGroup function description; this function does not copy the provided string, but use it directly instead. It's the caller responsability to clean it when not in use anymore. zsh 4.2 seems to follow this and try to deallocate the previous variable string when it has been replaced by a new one. But our putenv, which calls setenv, has already done the same ... -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
Re: CVS commit: src/lib/libc/stdlib
On Thu, Sep 30, 2010 at 12:41:34PM +, Matthias Scheler wrote: Module Name: src Committed By: tron Date: Thu Sep 30 12:41:33 UTC 2010 Modified Files: src/lib/libc/stdlib: setenv.c unsetenv.c Log Message: Be slightly more careful about freeing memory allocated for environment variables: only free memory if the current value points to the same memory area as the allocated block. This will prevent crashes if an application changes the order of the environment array. Unfortunately this is still not enough to stop zsh 4.2.* from crashing. zsh 4.3.* works fine before and after this change. Thanks Matthias, One possibility could be to not free memory at all in setenv, but only with unsetenv. IMHO, it's a programmer error to call setenv more than once on the same variable without a corresponding unsetenv in between (just like malloc()/free() behaviour). Otherwise, if i understand things correctly, it's more putenv that shoud work that way. -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
Re: CVS commit: src/bin/ps
On Wed, Jul 28, 2010 at 01:14:40PM +0200, Alan Barrett wrote: On Tue, 27 Jul 2010, Nicolas Joly wrote: v = ve-var; - doubleprintorsetwidth(v, getpcpu(k), 1, mode); + dbl = getpcpu(k); + doubleprintorsetwidth(v, dbl, (dbl = 100.0) ? 0 : 1, mode); } I think you need to test (dbl = 99.95); otherwise values in the range from 99.95 to 99.... will be printed with one decimal place, as 100.0. Right, will fix. -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
Re: CVS commit: src/usr.bin/mkdep
On Wed, May 26, 2010 at 02:07:34PM -0400, Christos Zoulas wrote: Module Name: src Committed By: christos Date: Wed May 26 18:07:34 UTC 2010 Modified Files: src/usr.bin/mkdep: mkdep.c Log Message: fix empty suffix case. Thanks again. But still not quite right, the .o extension remains while according to the man page it should be removed. nj...@lanfeust [~] mkdep -p -f main.dep main.c nj...@lanfeust [~] cat main.dep main.o: main.c nj...@lanfeust [~] mkdep -s '' -f main.dep main.c nj...@lanfeust [~] cat main.dep main.o: main.c -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
CVS commit: src/sys/conf
Module Name:src Committed By: njoly Date: Fri Mar 5 23:06:49 UTC 2010 Modified Files: src/sys/conf: Makefile.kern.inc Log Message: Use 'a b ...' contructs for running multiple commands on a single line in Makefiles, to properly detect/report failures. To generate a diff of this commit: cvs rdiff -u -r1.126 -r1.127 src/sys/conf/Makefile.kern.inc Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/conf
Module Name:src Committed By: njoly Date: Fri Mar 5 23:06:49 UTC 2010 Modified Files: src/sys/conf: Makefile.kern.inc Log Message: Use 'a b ...' contructs for running multiple commands on a single line in Makefiles, to properly detect/report failures. To generate a diff of this commit: cvs rdiff -u -r1.126 -r1.127 src/sys/conf/Makefile.kern.inc Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files. Modified files: Index: src/sys/conf/Makefile.kern.inc diff -u src/sys/conf/Makefile.kern.inc:1.126 src/sys/conf/Makefile.kern.inc:1.127 --- src/sys/conf/Makefile.kern.inc:1.126 Mon Mar 1 21:10:14 2010 +++ src/sys/conf/Makefile.kern.inc Fri Mar 5 23:06:49 2010 @@ -1,4 +1,4 @@ -# $NetBSD: Makefile.kern.inc,v 1.126 2010/03/01 21:10:14 darran Exp $ +# $NetBSD: Makefile.kern.inc,v 1.127 2010/03/05 23:06:49 njoly Exp $ # # This file contains common `MI' targets and definitions and it is included # at the bottom of each `MD' ${MACHINE}/conf/Makefile.${MACHINE}. @@ -123,7 +123,7 @@ # Add CTF sections for DTrace .if defined(CTFCONVERT) COMPILE_CTFCONVERT= ${_MKSHECHO}\ - ${CTFCONVERT} ${CTFFLAGS} ${.TARGET}; \ + ${CTFCONVERT} ${CTFFLAGS} ${.TARGET} \ ${CTFCONVERT} ${CTFFLAGS} ${.TARGET} .else COMPILE_CTFCONVERT= ${_MKSHECHO} @@ -131,19 +131,19 @@ # compile rules: rules are named ${TYPE}_${SUFFIX} where TYPE is NORMAL or # NOPROF and SUFFIX is the file suffix, capitalized (e.g. C for a .c file). -NORMAL_C?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET}; \ +NORMAL_C?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET} \ ${_MKSHECHO}\ - ${CC} ${CFLAGS} ${CPPFLAGS} ${PROF} -c $; \ - ${CC} ${CFLAGS} ${CPPFLAGS} ${PROF} -c $; \ + ${CC} ${CFLAGS} ${CPPFLAGS} ${PROF} -c $ \ + ${CC} ${CFLAGS} ${CPPFLAGS} ${PROF} -c $ \ ${COMPILE_CTFCONVERT} -NOPROF_C?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET}; \ +NOPROF_C?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET} \ ${_MKSHECHO}\ - ${CC} ${CFLAGS} ${CPPFLAGS} -c $; \ - ${CC} ${CFLAGS} ${CPPFLAGS} -c $; \ + ${CC} ${CFLAGS} ${CPPFLAGS} -c $ \ + ${CC} ${CFLAGS} ${CPPFLAGS} -c $ \ ${COMPILE_CTFCONVERT} -NORMAL_S?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET}; \ +NORMAL_S?= @${_MKSHMSG} compile ${.CURDIR:T}/${.TARGET} \ ${_MKSHECHO}\ - ${CC} ${AFLAGS} ${CPPFLAGS} -c $; \ + ${CC} ${AFLAGS} ${CPPFLAGS} -c $ \ ${CC} ${AFLAGS} ${CPPFLAGS} -c $ ##
CVS commit: src/sys/dev/i2c
Module Name:src Committed By: njoly Date: Mon Mar 1 17:35:21 UTC 2010 Modified Files: src/sys/dev/i2c: i2c.c Log Message: Do initialise i2c_attach_args new members, for indirect configuration too. To generate a diff of this commit: cvs rdiff -u -r1.24 -r1.25 src/sys/dev/i2c/i2c.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/dev/i2c
Module Name:src Committed By: njoly Date: Mon Mar 1 17:35:21 UTC 2010 Modified Files: src/sys/dev/i2c: i2c.c Log Message: Do initialise i2c_attach_args new members, for indirect configuration too. To generate a diff of this commit: cvs rdiff -u -r1.24 -r1.25 src/sys/dev/i2c/i2c.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files. Modified files: Index: src/sys/dev/i2c/i2c.c diff -u src/sys/dev/i2c/i2c.c:1.24 src/sys/dev/i2c/i2c.c:1.25 --- src/sys/dev/i2c/i2c.c:1.24 Sun Feb 28 11:34:42 2010 +++ src/sys/dev/i2c/i2c.c Mon Mar 1 17:35:21 2010 @@ -1,4 +1,4 @@ -/* $NetBSD: i2c.c,v 1.24 2010/02/28 11:34:42 martin Exp $ */ +/* $NetBSD: i2c.c,v 1.25 2010/03/01 17:35:21 njoly Exp $ */ /* * Copyright (c) 2003 Wasabi Systems, Inc. @@ -36,7 +36,7 @@ */ #include sys/cdefs.h -__KERNEL_RCSID(0, $NetBSD: i2c.c,v 1.24 2010/02/28 11:34:42 martin Exp $); +__KERNEL_RCSID(0, $NetBSD: i2c.c,v 1.25 2010/03/01 17:35:21 njoly Exp $); #include sys/param.h #include sys/systm.h @@ -108,6 +108,10 @@ ia.ia_size = cf-cf_loc[IICCF_SIZE]; ia.ia_type = sc-sc_type; + ia.ia_name = NULL; + ia.ia_ncompat = 0; + ia.ia_compat = NULL; + if (config_match(parent, cf, ia) 0) config_attach(parent, cf, ia, iic_print);
CVS commit: src/share/mk
Module Name:src Committed By: njoly Date: Mon Feb 22 12:33:22 UTC 2010 Modified Files: src/share/mk: bsd.README bsd.lib.mk Log Message: Change shared object file extension from .so to .pico, to avoid conflicts with shared libaries names libXX.so; as recently seen with MKUPDATE=yes builds for libelf and libnvpair. All now stalled regular .so files need to be manually removed from object directories. To generate a diff of this commit: cvs rdiff -u -r1.264 -r1.265 src/share/mk/bsd.README cvs rdiff -u -r1.303 -r1.304 src/share/mk/bsd.lib.mk Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src
Module Name:src Committed By: njoly Date: Mon Feb 22 12:35:32 UTC 2010 Modified Files: src: UPDATING Log Message: Note shared objects extension change, and stalled files. To generate a diff of this commit: cvs rdiff -u -r1.208 -r1.209 src/UPDATING Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src
Module Name:src Committed By: njoly Date: Mon Feb 22 12:35:32 UTC 2010 Modified Files: src: UPDATING Log Message: Note shared objects extension change, and stalled files. To generate a diff of this commit: cvs rdiff -u -r1.208 -r1.209 src/UPDATING Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files. Modified files: Index: src/UPDATING diff -u src/UPDATING:1.208 src/UPDATING:1.209 --- src/UPDATING:1.208 Thu Feb 4 22:25:21 2010 +++ src/UPDATING Mon Feb 22 12:35:32 2010 @@ -1,4 +1,4 @@ -$NetBSD: UPDATING,v 1.208 2010/02/04 22:25:21 roy Exp $ +$NetBSD: UPDATING,v 1.209 2010/02/22 12:35:32 njoly Exp $ This file (UPDATING) is intended to be a brief reference to recent changes that might cause problems in the build process, and a guide for @@ -16,6 +16,12 @@ Recent changes: ^^^ +20100222: + The shared objects file extension has been changed from .so to + .pico, in order to avoid conflicts with shared libraries names + libXX.so. All now stale regular .so files can be removed from + your object directories. + 20100204: The termcap database has been removed from the sources, but has not been marked obsolete so it is not removed
CVS commit: src/external/bsd/flex/dist
Module Name:src Committed By: njoly Date: Fri Feb 19 14:22:43 UTC 2010 Modified Files: src/external/bsd/flex/dist: initparse.c initscan.c Log Message: Regen initparse.c/initscan.c without #line directives, to avoid host tool depency problems when NETBSDSRCDIR != /usr/src. To generate a diff of this commit: cvs rdiff -u -r1.1 -r1.2 src/external/bsd/flex/dist/initparse.c cvs rdiff -u -r1.4 -r1.5 src/external/bsd/flex/dist/initscan.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
Re: CVS commit: src/external/cddl/osnet/lib/libnvpair
On Thu, Feb 18, 2010 at 07:39:05AM +1100, matthew green wrote: Well, I think it is worth it because it infrequently breaks random things. I am happy with .ro too for relocatable. let's go with .pico -- crunch already uses .ro. Fine with me. Unless someone beats me, i'll try to have a bsd.lib.mk patch to review/test soon ... -- Nicolas Joly Biological Software and Databanks. Institut Pasteur, Paris.
CVS commit: src/sys/dev/ic
Module Name:src Committed By: njoly Date: Tue Feb 9 23:04:15 UTC 2010 Modified Files: src/sys/dev/ic: nslm7x.c Log Message: Kill extra semi-column. To generate a diff of this commit: cvs rdiff -u -r1.52 -r1.53 src/sys/dev/ic/nslm7x.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/distrib/sets/lists/comp
Module Name:src Committed By: njoly Date: Fri Feb 5 12:46:33 UTC 2010 Modified Files: src/distrib/sets/lists/comp: mi Log Message: +infocmp.debug +tic.debug To generate a diff of this commit: cvs rdiff -u -r1.1383 -r1.1384 src/distrib/sets/lists/comp/mi Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/dev/pci
Module Name:src Committed By: njoly Date: Mon Feb 1 12:51:16 UTC 2010 Modified Files: src/sys/dev/pci: auich.c Log Message: Do not print ac97 link rate message unless verbose boot is requested. To generate a diff of this commit: cvs rdiff -u -r1.135 -r1.136 src/sys/dev/pci/auich.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/kern
Module Name:src Committed By: njoly Date: Mon Feb 1 12:58:04 UTC 2010 Modified Files: src/sys/kern: kern_ssp.c Log Message: Aprintify. To generate a diff of this commit: cvs rdiff -u -r1.3 -r1.4 src/sys/kern/kern_ssp.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/kern
Module Name:src Committed By: njoly Date: Mon Feb 1 16:14:58 UTC 2010 Modified Files: src/sys/kern: kern_ssp.c Log Message: Switch SSP init output to aprint_debug() instead of aprint_normal() under DIAGNOSTIC ifdefs. To generate a diff of this commit: cvs rdiff -u -r1.4 -r1.5 src/sys/kern/kern_ssp.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/share/man/man3
Module Name:src Committed By: njoly Date: Sun Jan 24 22:21:34 UTC 2010 Modified Files: src/share/man/man3: Makefile Log Message: Do not link timeradd.3 to itself. To generate a diff of this commit: cvs rdiff -u -r1.34 -r1.35 src/share/man/man3/Makefile Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/miscfs/kernfs
Module Name:src Committed By: njoly Date: Fri Jan 22 22:46:01 UTC 2010 Modified Files: src/sys/miscfs/kernfs: kernfs_vnops.c Log Message: Remove unneeded strlen() call in KFShostname case. To generate a diff of this commit: cvs rdiff -u -r1.139 -r1.140 src/sys/miscfs/kernfs/kernfs_vnops.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/sys/arch/amd64/amd64
Module Name:src Committed By: njoly Date: Mon Jan 18 21:55:40 UTC 2010 Modified Files: src/sys/arch/amd64/amd64: procfs_machdep.c Log Message: Do not check more than 32 bits against ci_feature_flags, to avoid printing bogus data on /proc/cpuinfo flags line. To generate a diff of this commit: cvs rdiff -u -r1.12 -r1.13 src/sys/arch/amd64/amd64/procfs_machdep.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.