Re: CVS commit: src/include
On Wed, Feb 16, 2011 at 02:29:36PM -0500, Christos Zoulas wrote: Module Name: src Committed By: christos Date: Wed Feb 16 19:29:35 UTC 2011 Modified Files: src/include: rmt.h Log Message: handle ssp To generate a diff of this commit: cvs rdiff -u -r1.6 -r1.7 src/include/rmt.h Can we please stop adding even more cludges for this? Joerg
Re: CVS commit: [netbsd-5] src/sys/arch/sparc/sparc
On Wed, Feb 16, 2011 at 09:33:25PM +, Manuel Bouyer wrote: Module Name: src Committed By: bouyer Date: Wed Feb 16 21:33:25 UTC 2011 Modified Files: src/sys/arch/sparc/sparc [netbsd-5]: cpu.c cpuvar.h pmap.c timer_sun4m.c Log Message: Apply patch, requested my mrg in ticket 1553: sys/arch/sparc/sparc/cpu.c: patch sys/arch/sparc/sparc/cpuvar.h: patch sys/arch/sparc/sparc/pmap.c:patch sys/arch/sparc/sparc/timer_sun4m.c: patch - print the curpcb in ddb mach cpu output as well. - bump the size of cpus[] by one, so we have a NULL pointer at the end, from tsutsui - for MP kernels, copy the loop to find the bootcpu in mainbus_attach() into getcacheinfo_obp() so we can get cache properties on the bootcpu before calling main() - in getcpuinfo(), move the call of getmid() before the call to getcacheinfo() so that the above change to getcacheinfo_obp() can work - move the struct cpu_info setup to the end of the initial kernel page setup and don't access this space until after we have switched to the kernel pagetables - revive most of the old CPUINFO_VA alignment/congruency code from the old alloc_cpuinfo_global_va() function, and ensure that all cpuinfo structures are sanely aligned. this makes hypersparc work again - introduce a new way to free the wasted pages back to UVM, as we can't simply uvm_unmap() them this early in bootstrap - make sure to initialise the cpuinfo sz in all cases. noted by martin. - add per-cpu event counters for lev10 and lev14 interrupts. - make CPU_INFO_FOREACH() set the iterator count to '0' in the !MP case. - add some disabled MP code to poke other cpus on level 14 interrupts. - add a diagnostic to ensure that cpus[0] == cpu0's cpu_info-ci_self - if a cpu doesn't have any mappings allocated, don't copy them. this occurs if a cpu isn't attached (such as a MP kernel with only cpu0 listed in the config file..) - fix the previous to compile !MULTIPROCESSOR. I forgot: from tsutsui: - Make sure to initialize cpus[0] which will also be used as cpuinfo for sun4 and sun4c, not only for sun4m. Okay'ed by mrg@. I fixed the CHANGES-5.2 entry -- Manuel Bouyer bou...@antioche.eu.org NetBSD: 26 ans d'experience feront toujours la difference --
Re: CVS commit: src/include
In article 20110216211601.ga24...@britannica.bec.de, Joerg Sonnenberger jo...@britannica.bec.de wrote: On Wed, Feb 16, 2011 at 02:29:36PM -0500, Christos Zoulas wrote: Module Name: src Committed By:christos Date:Wed Feb 16 19:29:35 UTC 2011 Modified Files: src/include: rmt.h Log Message: handle ssp To generate a diff of this commit: cvs rdiff -u -r1.6 -r1.7 src/include/rmt.h Can we please stop adding even more cludges for this? I wish, please show a better way. christos
Re: CVS commit: src/usr.bin/grep
Joerg Sonnenberger jo...@netbsd.org writes: Module Name: src Committed By: joerg Date: Wed Feb 16 01:31:34 UTC 2011 Modified Files: src/usr.bin/grep: Makefile file.c grep.1 grep.c grep.h queue.c util.c Added Files: src/usr.bin/grep: fastgrep.c I just did `more fastgrep.c' and found following piece of code. The usage of wflag is an obvious regression from OpenBSD code. | if (fg-len = 14 | strncmp(pat + (fg-bol ? 1 : 0), [[::]], 7) == 0 | strncmp(pat + (fg-bol ? 1 : 0) + fg-len - 7, [[::]], 7) == 0) { | fg-len -= 14; | /* Word boundary is handled separately in util.c */ | wflag = true; | } | | /* |* Copy pattern minus '^' and '$' characters as well as word |* match character classes at the beginning and ending of the |* string respectively. |*/ | fg-pattern = grep_malloc(fg-len + 1); | strlcpy((char *)fg-pattern, pat + (bol ? 1 : 0) + wflag, | fg-len + 1); enami.
Re: CVS commit: src/usr.bin/grep
On Thu, Feb 17, 2011 at 12:06:30PM +0900, enami tsugutomo wrote: Joerg Sonnenberger jo...@netbsd.org writes: Module Name:src Committed By: joerg Date: Wed Feb 16 01:31:34 UTC 2011 Modified Files: src/usr.bin/grep: Makefile file.c grep.1 grep.c grep.h queue.c util.c Added Files: src/usr.bin/grep: fastgrep.c I just did `more fastgrep.c' and found following piece of code. The usage of wflag is an obvious regression from OpenBSD code. In which sense? Joerg
Re: CVS commit: src/usr.bin/grep
In which sense? The meaning of `... + wflag' was to skip 7 chars. enami.