| 00:21.29 | *** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 01:07.24 | CIA-10 | 03sjhill 07uClibc-nptl * r13246 10/libc/stdlib/arc4random.c: Copy from trunk. |
| 01:08.42 | CIA-10 | 03sjhill 07uClibc-nptl * r13247 10/libc/sysdeps/linux/x86_64/bits/sem.h: Copy from trunk. |
| 01:44.11 | *** join/#uclibc flatronf701B (n=flatronf@202.75.186.154) |
| 01:49.10 | CIA-10 | 03vapier * r13248 10uClibc/libc/sysdeps/linux/x86_64/bits/ (environments.h sigcontext.h stat.h): import headers from glibc |
| 01:49.54 | CIA-10 | 03vapier * r13249 10uClibc/libc/sysdeps/linux/common/llseek.c: we dont want to truncate the offset when dealing with the syscall3 version as this will screw up some 64bit hosts |
| 02:05.55 | CIA-10 | 03vapier * r13250 10uClibc/test/misc/seek.c: test for seek funcs |
| 02:22.14 | khem | hi I need to use thumb aware return for ARM asm functions in libc/string/arm |
| 02:23.01 | khem | but I think the build calls assembler directly to assemble assembly routines so we can not tak advantage of gcc defines to write conditional code |
| 02:23.20 | khem | all functions currently use mov pc, lr |
| 02:23.46 | khem | We need to have bx lr in case of thumb variants |
| 02:24.12 | CIA-10 | 03vapier * r13251 10uClibc/Makefile.in: only chown on include subdir in DEVEL_PREFIX |
| 02:24.16 | khem | it could be controlled via gcc built in defines |
| 02:24.42 | khem | but then we need to use gcc driver to assemble |
| 02:24.47 | khem | these .S files |
| 02:29.10 | SpanKY | ? |
| 02:29.34 | SpanKY | .S files are preprocessed |
| 02:29.37 | SpanKY | .s are not |
| 02:29.57 | khem | SpanKY: oh so they are using GCC to preprocess |
| 02:30.18 | SpanKY | i'm pretty sure |
| 02:30.20 | SpanKY | lemme check |
| 02:30.41 | SpanKY | they must |
| 02:30.47 | SpanKY | look at like x86_64/memcpy.S |
| 02:30.55 | SpanKY | there's so much preprosser in that heh |
| 02:31.03 | khem | actually I was thinking of using gcc built-in defines which are decided based on commandline options |
| 02:31.06 | SpanKY | the AS display is misleading |
| 02:31.18 | khem | I believed in that :) |
| 02:31.26 | SpanKY | run `make V=1` to see the actual commands |
| 02:31.38 | khem | thanks for pointer |
| 02:31.46 | khem | I will do that |
| 02:31.52 | khem | now |
| 02:34.37 | khem | you are right so its calling gcc |
| 02:36.31 | SpanKY | if it didnt, pretty much every .S file we have in there would break heh |
| 02:42.26 | khem | SpanKY: I posted a patch assuming the same |
| 02:43.30 | khem | SpanKY: Is there also a arch system include file like sysdeps.h like glibc |
| 02:45.01 | khem | I see some defines in sysdeps/linux/arm/bits/ |
| 02:45.08 | SpanKY | what are you looking for ? |
| 02:45.12 | SpanKY | (like what macro) |
| 02:45.20 | khem | I want to add a macro |
| 02:45.43 | SpanKY | what kind of macro :) |
| 02:45.56 | khem | rather a define which is something like this |
| 02:47.06 | khem | #if (!defined (__ARM_ARCH_2__) && !defined (__ARM_ARCH_3__) \ |
| 02:47.06 | khem | <PROTECTED> |
| 02:47.06 | khem | # define __USE_BX__ |
| 02:47.06 | khem | #endif |
| 02:47.37 | khem | and then I will use this define to write conditional return |
| 02:48.21 | SpanKY | we dont really have an include like that |
| 02:48.47 | SpanKY | or rather, we dont have a header file that would normally accomodate such a thing |
| 02:49.04 | SpanKY | usually those sort of things we make a config option |
| 02:49.07 | khem | will it be acceptable to add one |
| 02:49.15 | khem | I see |
| 02:50.24 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-79-42.d-ip.magma.ca) |
| 02:50.34 | khem | actually this info can be obtained from CONFIG_CPU as well but how to do it |
| 02:51.26 | khem | e.g. CONFIG_ARM926T when set we should use BX to return |
| 02:51.59 | khem | is there a way to get this info during compilation |
| 02:52.45 | khem | or may be I can modify CPU_FLAGS to add a define in Rules.mak |
| 02:53.04 | SpanKY | you could use kconfig |
| 02:53.15 | SpanKY | make it an arm-specific option |
| 02:53.29 | SpanKY | extra/Configs/Config.arm |
| 02:53.37 | khem | OK |
| 02:53.55 | SpanKY | config USE_BX |
| 02:54.03 | SpanKY | depends on !CPU_... |
| 02:54.52 | khem | how will I use it in .S file then (sorry for stupid questions) |
| 02:57.35 | khem | SpanKY: will it generate autoconfig.h file like kernel will all defines |
| 02:59.23 | SpanKY | yes, and whenever you include <features.h>, you'll get access to all config options |
| 02:59.44 | khem | SpanKY: aha thanks a bunch |
| 03:02.15 | SpanKY | np |
| 03:06.52 | SpanKY | khem: know much about crt* files ? |
| 03:07.37 | CIA-10 | 03vapier * r13252 10uClibc/ldso/ldso/dl-elf.c: fix a (very unlikely) page leak |
| 03:07.51 | CIA-10 | 03landley * r13253 10busybox/util-linux/mdev.c: The major:minor read from dev ends with \n, need to trim that. |
| 03:16.27 | Newsome | yay! |
| 03:20.05 | Newsome | hmm, different fix, but whatever |
| 03:27.22 | *** join/#uclibc lsilva (n=leandro@20150019155.user.veloxzone.com.br) |
| 04:59.13 | *** join/#uclibc bryan (n=Bryan@cpe-70-123-129-170.austin.res.rr.com) |
| 05:52.46 | khem | SpanKY: sorry I was away for a bit |
| 06:13.53 | CIA-10 | 03landley * r13254 10busybox/util-linux/mdev.c: (log message trimmed) |
| 06:13.53 | CIA-10 | Frank Sorenson added hotplug support to mdev. (I tweaked it a bit. Need |
| 06:13.53 | CIA-10 | to come up with a test suite for all the stuff that requires root access. |
| 06:54.47 | *** join/#uclibc blindvt_ (n=bf@M851P003.adsl.highway.telekom.at) |
| 07:26.45 | *** join/#uclibc blindvt__ (n=bf@M926P015.adsl.highway.telekom.at) |
| 07:39.14 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
| 08:01.52 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:01.59 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.12 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.23 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 10:34.19 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
| 10:42.14 | CIA-10 | 03psm * r13255 10uClibc/include/fcntl.h: Add fcntl64 prototype |
| 10:50.15 | *** join/#uclibc blindvt` (n=blindvt`@chello213047219050.surfer.at) |
| 11:23.56 | CIA-10 | 03vapier * r13256 10uClibc/test/dlopen/Makefile: use LDFLAGS when linking .so files as pointed out by Peter S. Mazinger |
| 11:38.04 | *** join/#uclibc wigyori (i=wigyori@fehercapa.kektintahal.pirospolip.hu) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc SamChi (n=SamChi@37.Red-83-45-225.dynamicIP.rima-tde.net) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc Kaloz (i=kaloz@195.56.146.238) [NETSPLIT VICTIM] |
| 00:21.29 | *** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 01:07.24 | CIA-10 | 03sjhill 07uClibc-nptl * r13246 10/libc/stdlib/arc4random.c: Copy from trunk. |
| 01:08.42 | CIA-10 | 03sjhill 07uClibc-nptl * r13247 10/libc/sysdeps/linux/x86_64/bits/sem.h: Copy from trunk. |
| 01:44.11 | *** join/#uclibc flatronf701B (n=flatronf@202.75.186.154) |
| 01:49.10 | CIA-10 | 03vapier * r13248 10uClibc/libc/sysdeps/linux/x86_64/bits/ (environments.h sigcontext.h stat.h): import headers from glibc |
| 01:49.54 | CIA-10 | 03vapier * r13249 10uClibc/libc/sysdeps/linux/common/llseek.c: we dont want to truncate the offset when dealing with the syscall3 version as this will screw up some 64bit hosts |
| 02:05.55 | CIA-10 | 03vapier * r13250 10uClibc/test/misc/seek.c: test for seek funcs |
| 02:22.14 | khem | hi I need to use thumb aware return for ARM asm functions in libc/string/arm |
| 02:23.01 | khem | but I think the build calls assembler directly to assemble assembly routines so we can not tak advantage of gcc defines to write conditional code |
| 02:23.20 | khem | all functions currently use mov pc, lr |
| 02:23.46 | khem | We need to have bx lr in case of thumb variants |
| 02:24.12 | CIA-10 | 03vapier * r13251 10uClibc/Makefile.in: only chown on include subdir in DEVEL_PREFIX |
| 02:24.16 | khem | it could be controlled via gcc built in defines |
| 02:24.42 | khem | but then we need to use gcc driver to assemble |
| 02:24.47 | khem | these .S files |
| 02:29.10 | SpanKY | ? |
| 02:29.34 | SpanKY | .S files are preprocessed |
| 02:29.37 | SpanKY | .s are not |
| 02:29.57 | khem | SpanKY: oh so they are using GCC to preprocess |
| 02:30.18 | SpanKY | i'm pretty sure |
| 02:30.20 | SpanKY | lemme check |
| 02:30.41 | SpanKY | they must |
| 02:30.47 | SpanKY | look at like x86_64/memcpy.S |
| 02:30.55 | SpanKY | there's so much preprosser in that heh |
| 02:31.03 | khem | actually I was thinking of using gcc built-in defines which are decided based on commandline options |
| 02:31.06 | SpanKY | the AS display is misleading |
| 02:31.18 | khem | I believed in that :) |
| 02:31.26 | SpanKY | run `make V=1` to see the actual commands |
| 02:31.38 | khem | thanks for pointer |
| 02:31.46 | khem | I will do that |
| 02:31.52 | khem | now |
| 02:34.37 | khem | you are right so its calling gcc |
| 02:36.31 | SpanKY | if it didnt, pretty much every .S file we have in there would break heh |
| 02:42.26 | khem | SpanKY: I posted a patch assuming the same |
| 02:43.30 | khem | SpanKY: Is there also a arch system include file like sysdeps.h like glibc |
| 02:45.01 | khem | I see some defines in sysdeps/linux/arm/bits/ |
| 02:45.08 | SpanKY | what are you looking for ? |
| 02:45.12 | SpanKY | (like what macro) |
| 02:45.20 | khem | I want to add a macro |
| 02:45.43 | SpanKY | what kind of macro :) |
| 02:45.56 | khem | rather a define which is something like this |
| 02:47.06 | khem | #if (!defined (__ARM_ARCH_2__) && !defined (__ARM_ARCH_3__) \ |
| 02:47.06 | khem | <PROTECTED> |
| 02:47.06 | khem | # define __USE_BX__ |
| 02:47.06 | khem | #endif |
| 02:47.37 | khem | and then I will use this define to write conditional return |
| 02:48.21 | SpanKY | we dont really have an include like that |
| 02:48.47 | SpanKY | or rather, we dont have a header file that would normally accomodate such a thing |
| 02:49.04 | SpanKY | usually those sort of things we make a config option |
| 02:49.07 | khem | will it be acceptable to add one |
| 02:49.15 | khem | I see |
| 02:50.24 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-79-42.d-ip.magma.ca) |
| 02:50.34 | khem | actually this info can be obtained from CONFIG_CPU as well but how to do it |
| 02:51.26 | khem | e.g. CONFIG_ARM926T when set we should use BX to return |
| 02:51.59 | khem | is there a way to get this info during compilation |
| 02:52.45 | khem | or may be I can modify CPU_FLAGS to add a define in Rules.mak |
| 02:53.04 | SpanKY | you could use kconfig |
| 02:53.15 | SpanKY | make it an arm-specific option |
| 02:53.29 | SpanKY | extra/Configs/Config.arm |
| 02:53.37 | khem | OK |
| 02:53.55 | SpanKY | config USE_BX |
| 02:54.03 | SpanKY | depends on !CPU_... |
| 02:54.52 | khem | how will I use it in .S file then (sorry for stupid questions) |
| 02:57.35 | khem | SpanKY: will it generate autoconfig.h file like kernel will all defines |
| 02:59.23 | SpanKY | yes, and whenever you include <features.h>, you'll get access to all config options |
| 02:59.44 | khem | SpanKY: aha thanks a bunch |
| 03:02.15 | SpanKY | np |
| 03:06.52 | SpanKY | khem: know much about crt* files ? |
| 03:07.37 | CIA-10 | 03vapier * r13252 10uClibc/ldso/ldso/dl-elf.c: fix a (very unlikely) page leak |
| 03:07.51 | CIA-10 | 03landley * r13253 10busybox/util-linux/mdev.c: The major:minor read from dev ends with \n, need to trim that. |
| 03:16.27 | Newsome | yay! |
| 03:20.05 | Newsome | hmm, different fix, but whatever |
| 03:27.22 | *** join/#uclibc lsilva (n=leandro@20150019155.user.veloxzone.com.br) |
| 04:59.13 | *** join/#uclibc bryan (n=Bryan@cpe-70-123-129-170.austin.res.rr.com) |
| 05:52.46 | khem | SpanKY: sorry I was away for a bit |
| 06:13.53 | CIA-10 | 03landley * r13254 10busybox/util-linux/mdev.c: (log message trimmed) |
| 06:13.53 | CIA-10 | Frank Sorenson added hotplug support to mdev. (I tweaked it a bit. Need |
| 06:13.53 | CIA-10 | to come up with a test suite for all the stuff that requires root access. |
| 06:54.47 | *** join/#uclibc blindvt_ (n=bf@M851P003.adsl.highway.telekom.at) |
| 07:26.45 | *** join/#uclibc blindvt__ (n=bf@M926P015.adsl.highway.telekom.at) |
| 07:39.14 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
| 08:01.52 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:01.59 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.12 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.23 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 10:34.19 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
| 10:42.14 | CIA-10 | 03psm * r13255 10uClibc/include/fcntl.h: Add fcntl64 prototype |
| 10:50.15 | *** join/#uclibc blindvt` (n=blindvt`@chello213047219050.surfer.at) |
| 11:23.56 | CIA-10 | 03vapier * r13256 10uClibc/test/dlopen/Makefile: use LDFLAGS when linking .so files as pointed out by Peter S. Mazinger |
| 11:38.04 | *** join/#uclibc wigyori (i=wigyori@fehercapa.kektintahal.pirospolip.hu) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc SamChi (n=SamChi@37.Red-83-45-225.dynamicIP.rima-tde.net) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc Kaloz (i=kaloz@195.56.146.238) [NETSPLIT VICTIM] |
| 12:08.48 | CIA-10 | 03aldot * r13257 10busybox/networking/libiproute/iproute.c: - mark argument who of print_route as unused. |
| 12:18.37 | *** join/#uclibc prpplague (n=billybob@72.22.152.142) |
| 13:13.45 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-79-42.d-ip.magma.ca) |
| 13:15.52 | CIA-10 | 03aldot * r13258 10busybox/networking/libiproute/iptunnel.c: (log message trimmed) |
| 13:15.53 | CIA-10 | - remove warning about redeclaration of __constant_htons |
| 13:15.53 | CIA-10 | - remove warning about format '%ld' expects type 'long int *', but argument #num |
| 14:51.31 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 14:56.16 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 14:59.07 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:04.14 | *** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 15:17.58 | vapier | psm: alive ? |
| 15:19.16 | psm | yeah |
| 15:19.32 | vapier | trying to rename ld-uClibc |
| 15:19.37 | vapier | but it's being bitchy |
| 15:19.47 | *** part/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 15:20.28 | psm | for x86_64? |
| 15:21.11 | vapier | yeah, i want to call it ld64-... |
| 15:21.18 | vapier | for future multilib fun |
| 15:21.42 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:21.42 | vapier | and so i can debug i686/x86_64 on my host system easily :) |
| 15:22.09 | vapier | http://rafb.net/paste/results/4DnRuE21.html |
| 15:22.16 | vapier | here's what i got so far but doesnt seem to be working right |
| 15:23.54 | psm | maybe somewhere hardcoded? in test it was iirc |
| 15:24.14 | vapier | it's failing `make` in topdir |
| 15:24.48 | psm | you need some rule to build maybe in ldso? |
| 15:25.15 | vapier | i thought all of that stuff had moved to using $(UCLIBC_LDSO) |
| 15:25.42 | psm | maybe not the compiler/link rules |
| 15:26.29 | vapier | k, i'll peep there |
| 15:26.48 | vapier | i'm also hoping that is the reason for my python bug in amd64/uclibc |
| 15:26.48 | psm | CFLAGS-ld-uClibc.so fails for sure |
| 15:28.38 | vapier | hrm it's working now |
| 15:28.39 | vapier | lakdsjfalskf |
| 15:28.44 | vapier | Portage 2.1_pre3-r1 (uclibc/amd64, gcc-3.4.5, unavailable, 2.6.14.3-grsec x86_64) |
| 15:29.01 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:30.02 | *** join/#uclibc blindvt (n=bf@M959P016.adsl.highway.telekom.at) |
| 15:30.40 | vapier | we should change ld-uClibc-so-y to ldso-y |
| 15:32.21 | vapier | psm: btw, your gcc-4.0.x ssp stuff, where did you get it ? |
| 15:32.22 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:35.36 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:36.01 | psm | its mainly redhat, there is a bug about it on bugs.g.o, you need other patches as well before applying that one |
| 15:36.32 | psm | you should use it only w/ uClibc svn, glibc is not ok |
| 15:37.12 | vapier | could you post the updated stuff to that bug ? |
| 15:37.27 | vapier | we want to move 4.0.x into ~x86 and we'll have to mask it on hardened otherwise :/ |
| 15:38.14 | CIA-10 | 03aldot * r13259 10busybox/libbb/obscure.c: (log message trimmed) |
| 15:38.15 | CIA-10 | - shrink simple obscure stuff a tiny bit: |
| 15:38.15 | CIA-10 | <PROTECTED> |
| 15:38.20 | psm | I explained here in this channel, that moving to arch will be a failure, if you do not get glibc correctly before that |
| 15:38.40 | psm | sorry: on gentoo-embedded |
| 15:38.44 | vapier | right, we need to update the symbols |
| 15:39.51 | vapier | you going to be around this evening psm ? |
| 15:40.19 | vapier | say in 8 hours ? |
| 15:40.32 | vapier | we can get the glibc crap all ironed out then in #gentoo-embedded |
| 15:40.58 | psm | that is 00:40 here, not sure |
| 15:41.20 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:41.44 | psm | I do not help w/ glibc-ssp until the fedora patch is not updated to 2005[11-12]xx |
| 15:43.04 | psm | but I said that 2 month ago already, nothing happened ;-( |
| 15:43.20 | vapier | generally i try to avoid hardened related stuff |
| 15:43.27 | blindvt` | any idea when somebody will go ahead an tag that branches/busybox-1_1 thingy? |
| 15:43.38 | vapier | blindvt: landley is the only one likely to do that |
| 15:44.04 | psm | fedora patch is not hardened, it is more than a year old |
| 15:44.30 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:45.04 | blindvt` | i hope landley does remember which rev he wanted to tag, imho i waited long enough to checkin stuff (i forgot to look if there was already a branch when i checked a few bits in eralier today.. ) |
| 15:45.35 | blindvt` | earlier even |
| 15:46.01 | vapier | psm: generally i try to avoid that stuff as well because i've never touched it before |
| 15:46.09 | vapier | psm: if i had it my way, i'd never touch glibc :) |
| 15:46.20 | vapier | but it seems that no one else is touching it so it needs to be updated |
| 15:46.26 | vapier | which is what i'd like to go through with you tonight |
| 15:47.17 | psm | I'll try to be up, havent slept much lately |
| 15:47.35 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:48.26 | vapier | well we can do it another time |
| 15:53.28 | vapier | bbiab |
| 16:31.50 | *** join/#uclibc ulf_k (n=ulf_kypk@p54BDB125.dip0.t-ipconnect.de) |
| 16:40.59 | *** join/#uclibc Newsome (n=sorenson@obelix.cs.byu.edu) |
| 16:41.16 | *** join/#uclibc rfs (n=rsiemsen@209-87-231-114.storm.ca) |
| 17:02.09 | *** join/#uclibc JockeHome (n=JockeHom@84-217-91-123.tn.glocalnet.net) |
| 00:21.29 | *** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 01:07.24 | CIA-10 | 03sjhill 07uClibc-nptl * r13246 10/libc/stdlib/arc4random.c: Copy from trunk. |
| 01:08.42 | CIA-10 | 03sjhill 07uClibc-nptl * r13247 10/libc/sysdeps/linux/x86_64/bits/sem.h: Copy from trunk. |
| 01:44.11 | *** join/#uclibc flatronf701B (n=flatronf@202.75.186.154) |
| 01:49.10 | CIA-10 | 03vapier * r13248 10uClibc/libc/sysdeps/linux/x86_64/bits/ (environments.h sigcontext.h stat.h): import headers from glibc |
| 01:49.54 | CIA-10 | 03vapier * r13249 10uClibc/libc/sysdeps/linux/common/llseek.c: we dont want to truncate the offset when dealing with the syscall3 version as this will screw up some 64bit hosts |
| 02:05.55 | CIA-10 | 03vapier * r13250 10uClibc/test/misc/seek.c: test for seek funcs |
| 02:22.14 | khem | hi I need to use thumb aware return for ARM asm functions in libc/string/arm |
| 02:23.01 | khem | but I think the build calls assembler directly to assemble assembly routines so we can not tak advantage of gcc defines to write conditional code |
| 02:23.20 | khem | all functions currently use mov pc, lr |
| 02:23.46 | khem | We need to have bx lr in case of thumb variants |
| 02:24.12 | CIA-10 | 03vapier * r13251 10uClibc/Makefile.in: only chown on include subdir in DEVEL_PREFIX |
| 02:24.16 | khem | it could be controlled via gcc built in defines |
| 02:24.42 | khem | but then we need to use gcc driver to assemble |
| 02:24.47 | khem | these .S files |
| 02:29.10 | SpanKY | ? |
| 02:29.34 | SpanKY | .S files are preprocessed |
| 02:29.37 | SpanKY | .s are not |
| 02:29.57 | khem | SpanKY: oh so they are using GCC to preprocess |
| 02:30.18 | SpanKY | i'm pretty sure |
| 02:30.20 | SpanKY | lemme check |
| 02:30.41 | SpanKY | they must |
| 02:30.47 | SpanKY | look at like x86_64/memcpy.S |
| 02:30.55 | SpanKY | there's so much preprosser in that heh |
| 02:31.03 | khem | actually I was thinking of using gcc built-in defines which are decided based on commandline options |
| 02:31.06 | SpanKY | the AS display is misleading |
| 02:31.18 | khem | I believed in that :) |
| 02:31.26 | SpanKY | run `make V=1` to see the actual commands |
| 02:31.38 | khem | thanks for pointer |
| 02:31.46 | khem | I will do that |
| 02:31.52 | khem | now |
| 02:34.37 | khem | you are right so its calling gcc |
| 02:36.31 | SpanKY | if it didnt, pretty much every .S file we have in there would break heh |
| 02:42.26 | khem | SpanKY: I posted a patch assuming the same |
| 02:43.30 | khem | SpanKY: Is there also a arch system include file like sysdeps.h like glibc |
| 02:45.01 | khem | I see some defines in sysdeps/linux/arm/bits/ |
| 02:45.08 | SpanKY | what are you looking for ? |
| 02:45.12 | SpanKY | (like what macro) |
| 02:45.20 | khem | I want to add a macro |
| 02:45.43 | SpanKY | what kind of macro :) |
| 02:45.56 | khem | rather a define which is something like this |
| 02:47.06 | khem | #if (!defined (__ARM_ARCH_2__) && !defined (__ARM_ARCH_3__) \ |
| 02:47.06 | khem | <PROTECTED> |
| 02:47.06 | khem | # define __USE_BX__ |
| 02:47.06 | khem | #endif |
| 02:47.37 | khem | and then I will use this define to write conditional return |
| 02:48.21 | SpanKY | we dont really have an include like that |
| 02:48.47 | SpanKY | or rather, we dont have a header file that would normally accomodate such a thing |
| 02:49.04 | SpanKY | usually those sort of things we make a config option |
| 02:49.07 | khem | will it be acceptable to add one |
| 02:49.15 | khem | I see |
| 02:50.24 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-79-42.d-ip.magma.ca) |
| 02:50.34 | khem | actually this info can be obtained from CONFIG_CPU as well but how to do it |
| 02:51.26 | khem | e.g. CONFIG_ARM926T when set we should use BX to return |
| 02:51.59 | khem | is there a way to get this info during compilation |
| 02:52.45 | khem | or may be I can modify CPU_FLAGS to add a define in Rules.mak |
| 02:53.04 | SpanKY | you could use kconfig |
| 02:53.15 | SpanKY | make it an arm-specific option |
| 02:53.29 | SpanKY | extra/Configs/Config.arm |
| 02:53.37 | khem | OK |
| 02:53.55 | SpanKY | config USE_BX |
| 02:54.03 | SpanKY | depends on !CPU_... |
| 02:54.52 | khem | how will I use it in .S file then (sorry for stupid questions) |
| 02:57.35 | khem | SpanKY: will it generate autoconfig.h file like kernel will all defines |
| 02:59.23 | SpanKY | yes, and whenever you include <features.h>, you'll get access to all config options |
| 02:59.44 | khem | SpanKY: aha thanks a bunch |
| 03:02.15 | SpanKY | np |
| 03:06.52 | SpanKY | khem: know much about crt* files ? |
| 03:07.37 | CIA-10 | 03vapier * r13252 10uClibc/ldso/ldso/dl-elf.c: fix a (very unlikely) page leak |
| 03:07.51 | CIA-10 | 03landley * r13253 10busybox/util-linux/mdev.c: The major:minor read from dev ends with \n, need to trim that. |
| 03:16.27 | Newsome | yay! |
| 03:20.05 | Newsome | hmm, different fix, but whatever |
| 03:27.22 | *** join/#uclibc lsilva (n=leandro@20150019155.user.veloxzone.com.br) |
| 04:59.13 | *** join/#uclibc bryan (n=Bryan@cpe-70-123-129-170.austin.res.rr.com) |
| 05:52.46 | khem | SpanKY: sorry I was away for a bit |
| 06:13.53 | CIA-10 | 03landley * r13254 10busybox/util-linux/mdev.c: (log message trimmed) |
| 06:13.53 | CIA-10 | Frank Sorenson added hotplug support to mdev. (I tweaked it a bit. Need |
| 06:13.53 | CIA-10 | to come up with a test suite for all the stuff that requires root access. |
| 06:54.47 | *** join/#uclibc blindvt_ (n=bf@M851P003.adsl.highway.telekom.at) |
| 07:26.45 | *** join/#uclibc blindvt__ (n=bf@M926P015.adsl.highway.telekom.at) |
| 07:39.14 | *** join/#uclibc psaksa (n=pate@ip212-226-134-137.adsl.kpnqwest.fi) |
| 08:01.52 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:01.59 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.12 | *** join/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 08:02.23 | *** part/#uclibc marcusbrutus (n=marcusbr@ppp108-88.static.internode.on.net) |
| 10:34.19 | *** join/#uclibc psm (n=mps@host-6.mikroweb.hu) |
| 10:42.14 | CIA-10 | 03psm * r13255 10uClibc/include/fcntl.h: Add fcntl64 prototype |
| 10:50.15 | *** join/#uclibc blindvt` (n=blindvt`@chello213047219050.surfer.at) |
| 11:23.56 | CIA-10 | 03vapier * r13256 10uClibc/test/dlopen/Makefile: use LDFLAGS when linking .so files as pointed out by Peter S. Mazinger |
| 11:38.04 | *** join/#uclibc wigyori (i=wigyori@fehercapa.kektintahal.pirospolip.hu) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc SamChi (n=SamChi@37.Red-83-45-225.dynamicIP.rima-tde.net) [NETSPLIT VICTIM] |
| 11:38.04 | *** join/#uclibc Kaloz (i=kaloz@195.56.146.238) [NETSPLIT VICTIM] |
| 12:08.48 | CIA-10 | 03aldot * r13257 10busybox/networking/libiproute/iproute.c: - mark argument who of print_route as unused. |
| 12:18.37 | *** join/#uclibc prpplague (n=billybob@72.22.152.142) |
| 13:13.45 | *** join/#uclibc thraxisp (n=thraxisp@ottawa-hs-209-217-79-42.d-ip.magma.ca) |
| 13:15.52 | CIA-10 | 03aldot * r13258 10busybox/networking/libiproute/iptunnel.c: (log message trimmed) |
| 13:15.53 | CIA-10 | - remove warning about redeclaration of __constant_htons |
| 13:15.53 | CIA-10 | - remove warning about format '%ld' expects type 'long int *', but argument #num |
| 14:51.31 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 14:56.16 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 14:59.07 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:04.14 | *** join/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 15:17.58 | vapier | psm: alive ? |
| 15:19.16 | psm | yeah |
| 15:19.32 | vapier | trying to rename ld-uClibc |
| 15:19.37 | vapier | but it's being bitchy |
| 15:19.47 | *** part/#uclibc Newsome (n=sorenson@216-190-206-130.customer.csolutions.net) |
| 15:20.28 | psm | for x86_64? |
| 15:21.11 | vapier | yeah, i want to call it ld64-... |
| 15:21.18 | vapier | for future multilib fun |
| 15:21.42 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:21.42 | vapier | and so i can debug i686/x86_64 on my host system easily :) |
| 15:22.09 | vapier | http://rafb.net/paste/results/4DnRuE21.html |
| 15:22.16 | vapier | here's what i got so far but doesnt seem to be working right |
| 15:23.54 | psm | maybe somewhere hardcoded? in test it was iirc |
| 15:24.14 | vapier | it's failing `make` in topdir |
| 15:24.48 | psm | you need some rule to build maybe in ldso? |
| 15:25.15 | vapier | i thought all of that stuff had moved to using $(UCLIBC_LDSO) |
| 15:25.42 | psm | maybe not the compiler/link rules |
| 15:26.29 | vapier | k, i'll peep there |
| 15:26.48 | vapier | i'm also hoping that is the reason for my python bug in amd64/uclibc |
| 15:26.48 | psm | CFLAGS-ld-uClibc.so fails for sure |
| 15:28.38 | vapier | hrm it's working now |
| 15:28.39 | vapier | lakdsjfalskf |
| 15:28.44 | vapier | Portage 2.1_pre3-r1 (uclibc/amd64, gcc-3.4.5, unavailable, 2.6.14.3-grsec x86_64) |
| 15:29.01 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:30.02 | *** join/#uclibc blindvt (n=bf@M959P016.adsl.highway.telekom.at) |
| 15:30.40 | vapier | we should change ld-uClibc-so-y to ldso-y |
| 15:32.21 | vapier | psm: btw, your gcc-4.0.x ssp stuff, where did you get it ? |
| 15:32.22 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:35.36 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:36.01 | psm | its mainly redhat, there is a bug about it on bugs.g.o, you need other patches as well before applying that one |
| 15:36.32 | psm | you should use it only w/ uClibc svn, glibc is not ok |
| 15:37.12 | vapier | could you post the updated stuff to that bug ? |
| 15:37.27 | vapier | we want to move 4.0.x into ~x86 and we'll have to mask it on hardened otherwise :/ |
| 15:38.14 | CIA-10 | 03aldot * r13259 10busybox/libbb/obscure.c: (log message trimmed) |
| 15:38.15 | CIA-10 | - shrink simple obscure stuff a tiny bit: |
| 15:38.15 | CIA-10 | <PROTECTED> |
| 15:38.20 | psm | I explained here in this channel, that moving to arch will be a failure, if you do not get glibc correctly before that |
| 15:38.40 | psm | sorry: on gentoo-embedded |
| 15:38.44 | vapier | right, we need to update the symbols |
| 15:39.51 | vapier | you going to be around this evening psm ? |
| 15:40.19 | vapier | say in 8 hours ? |
| 15:40.32 | vapier | we can get the glibc crap all ironed out then in #gentoo-embedded |
| 15:40.58 | psm | that is 00:40 here, not sure |
| 15:41.20 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:41.44 | psm | I do not help w/ glibc-ssp until the fedora patch is not updated to 2005[11-12]xx |
| 15:43.04 | psm | but I said that 2 month ago already, nothing happened ;-( |
| 15:43.20 | vapier | generally i try to avoid hardened related stuff |
| 15:43.27 | blindvt` | any idea when somebody will go ahead an tag that branches/busybox-1_1 thingy? |
| 15:43.38 | vapier | blindvt: landley is the only one likely to do that |
| 15:44.04 | psm | fedora patch is not hardened, it is more than a year old |
| 15:44.30 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:45.04 | blindvt` | i hope landley does remember which rev he wanted to tag, imho i waited long enough to checkin stuff (i forgot to look if there was already a branch when i checked a few bits in eralier today.. ) |
| 15:45.35 | blindvt` | earlier even |
| 15:46.01 | vapier | psm: generally i try to avoid that stuff as well because i've never touched it before |
| 15:46.09 | vapier | psm: if i had it my way, i'd never touch glibc :) |
| 15:46.20 | vapier | but it seems that no one else is touching it so it needs to be updated |
| 15:46.26 | vapier | which is what i'd like to go through with you tonight |
| 15:47.17 | psm | I'll try to be up, havent slept much lately |
| 15:47.35 | *** join/#uclibc ashes (n=ashes@modemcable085.56-130-66.mc.videotron.ca) |
| 15:48.26 | vapier | well we can do it another time |
| 15:53.28 | vapier | bbiab |
| 16:31.50 | *** join/#uclibc ulf_k (n=ulf_kypk@p54BDB125.dip0.t-ipconnect.de) |
| 16:40.59 | *** join/#uclibc Newsome (n=sorenson@obelix.cs.byu.edu) |
| 16:41.16 | *** join/#uclibc rfs (n=rsiemsen@209-87-231-114.storm.ca) |
| 17:02.09 | *** join/#uclibc JockeHome (n=JockeHom@84-217-91-123.tn.glocalnet.net) |
| 18:41.19 | *** join/#uclibc thraxisp (n=thraxisp@ottgate.precidia.com) |
| 18:42.45 | *** join/#uclibc woglinde (n=woglinde@e178105179.adsl.alicedsl.de) |
| 19:05.01 | *** join/#uclibc nitinkg (n=ngupta@gateway-1237.mvista.com) [NETSPLIT VICTIM] |
| 19:08.25 | *** join/#uclibc tty56 (n=johannes@2001:6f8:1331:3:3d82:7326:f3ee:cc7a) |
| 20:29.52 | khem | has some one seen unresolved symbol "__ctype_b" |
| 20:48.54 | psm | have you enabled XLOCALE? |
| 20:49.15 | khem | psm: yes I think |
| 20:49.20 | khem | let me check |
| 20:49.29 | psm | dont do it |
| 20:49.47 | psm | or you 0.9.28 |
| 20:49.57 | khem | UCLIBC_HAS_XLOCALE=y |
| 20:50.01 | khem | oh I see |
| 20:50.13 | khem | I am using 0.9.28 |
| 20:50.43 | khem | infact a snapshot from 17/11/2005 |
| 20:51.36 | khem | psm: I will disable it and try a new build |
| 20:52.59 | psm | either vanilla or current, nothing between, because it changed all the time so noone will help w/ a 2 month old snapshot |
| 20:53.31 | khem | psm: OK I will update my snapshot too |
| 20:54.44 | psm | vapier: *stat64 redefining, I have to modify that code, how can I test your failure (__stat64 wont exist anymore, so I do not know what to undef) ? |
| 21:05.47 | khem | psm: thanks the issue is gone if I disable xlocale |
| 21:21.35 | vapier | psm: just start using your hidden_proto() and stuff and i'll let you know when it fails :) |
| 21:23.28 | psm | ok, I remove your redefines and comments (but not these days there is much to do before I commit) |
| 21:24.17 | vapier | if it fails i'll show you where/why, but not commit stuff |
| 21:24.24 | vapier | no point in clobbering each other constantly |
| 21:24.42 | vapier | need to figure out why sem stuff fails on x86_64/uclibc ... |
| 21:32.29 | blindvt | btw.. i remember to have seen an empty __linkage_warning when linking an app (busybox, iirc) with uClibc. Is this known and expected? |
| 22:51.18 | SpanKY | blindvt: yes, ive seen that |
| 22:51.28 | SpanKY | i think it's the fault of the linker, but i havent checked |
| 22:57.21 | *** join/#uclibc ulf_k (n=ulf_kypk@p54BDB125.dip0.t-ipconnect.de) |
| 23:29.36 | *** join/#uclibc blindvt_ (n=bf@M1019P026.adsl.highway.telekom.at) |
| 23:32.29 | *** join/#uclibc ulf_k_ (n=ulf_kypk@p54BDB593.dip0.t-ipconnect.de) |
| 23:53.33 | *** join/#uclibc ulf_k_ (n=ulf_kypk@p54BDB593.dip0.t-ipconnect.de) |