From 6f9c0342f33f52a2bda98d3709fbefd678bd46d9 Mon Sep 17 00:00:00 2001 From: Thomas Schwinge Date: Sun, 13 May 2012 06:06:11 +0200 Subject: IRC. --- open_issues/bpf.mdwn | 23 ++++ open_issues/glibc.mdwn | 83 +++++++++++++- open_issues/glibc/getlogin_r.mdwn | 45 ++++++++ open_issues/glibc/octave.mdwn | 35 ++++++ open_issues/gnumach_memory_management.mdwn | 17 +++ open_issues/gnumach_page_cache_policy.mdwn | 35 ++++++ open_issues/gnumach_vm_map_red-black_trees.mdwn | 142 ++++++++++++++++++++++++ open_issues/libpthread.mdwn | 53 +++------ open_issues/libpthread_CLOCK_MONOTONIC.mdwn | 58 ++++++++++ open_issues/libpthread_dlopen.mdwn | 49 +++++++- open_issues/o_exec.mdwn | 54 +++++++++ open_issues/packaging_libpthread.mdwn | 114 ++++++++++++++++--- open_issues/system_call_mechanism.mdwn | 43 ++++++- open_issues/user-space_device_drivers.mdwn | 10 +- 14 files changed, 696 insertions(+), 65 deletions(-) create mode 100644 open_issues/glibc/getlogin_r.mdwn create mode 100644 open_issues/glibc/octave.mdwn create mode 100644 open_issues/gnumach_page_cache_policy.mdwn create mode 100644 open_issues/gnumach_vm_map_red-black_trees.mdwn create mode 100644 open_issues/libpthread_CLOCK_MONOTONIC.mdwn create mode 100644 open_issues/o_exec.mdwn (limited to 'open_issues') diff --git a/open_issues/bpf.mdwn b/open_issues/bpf.mdwn index 2a8c897a..e24d761b 100644 --- a/open_issues/bpf.mdwn +++ b/open_issues/bpf.mdwn @@ -562,3 +562,26 @@ This is a collection of resources concerning *Berkeley Packet Filter*s. DDE accordingly antrik: i agree antrik: eth-multiplexer would be the right place + + +## IRC, freenode, #hurd, 2012-04-24 + + braunr: Is BPF fully supported by now? Can it be used for + isc-dhcp? + gnu_srs: bpf isn't supported at all + gnu_srs: instead of emulating it, i added a hurd-specific module + in libpcap + if isc-dhcp can use libpcap, then fine + (otherwise we could create a hurd-specific patch for dhcp that + uses the in-kernel bpf filter implementation) + gnu_srs: can't it use a raw socket ? + it can + it's just that the shape of the patch to do so wasn't exactly how + they needed it + so they have to rework it a bit + and that takes time + ok + antrik: for now, we prefer encapsulating the system specific code + in libpcap, and let users of that library benefit from it + instead of implementing the low level bpf interface, which + nonetheless has some system-specific variants .. diff --git a/open_issues/glibc.mdwn b/open_issues/glibc.mdwn index b15c880a..26fa6012 100644 --- a/open_issues/glibc.mdwn +++ b/open_issues/glibc.mdwn @@ -202,11 +202,59 @@ Last reviewed up to the [[Git mirror's d40c5d54cb551acba4ef1617464760c5b3d41a14 `CLOCK_REALTIME_ALARM`, `O_PATH`, `PTRACE_*` (for example, cbff0d9689c4d68578b6a4f0a17807232506ea27), `RLIMIT_RTTIME`, `SEEK_DATA` (`unistd.h`), `SEEK_HOLE` (`unistd.h`) - `clock_adjtime`, `fallocate`, `fallocate64`, `getcontext` (and - `setcontext`), `name_to_handle_at`, `open_by_handle_at`, - `process_vm_readv`, `process_vm_writev`, `sendmmsg`, + `clock_adjtime`, `fallocate`, `fallocate64`, `name_to_handle_at`, + `open_by_handle_at`, `process_vm_readv`, `process_vm_writev`, `sendmmsg`, `setns`, `sync_file_range` + * `getcontext`/`setcontext` + + Needed for [[gccgo]]. + + IRC, freenode, #hurd, 2012-04-19: + + How much work/knowledge is needed to implement + getcontext/setcontext? + Any already implemented alternatives available? + x86 registers knowledge, as well as unix signal masks + there's the linux implementation that can be taken as an + exxample, but the signal part has to be rewritten + Well, it's a pity they are not implemented. That's the + remaining hurdle to get gccgo working :-( + uh :/ + Everything builds, but the testsuite fails due to these + missing functions. + Regarding getcontext/setcontext they seem to be written + in assembly for linux but the code is not very long. + How much effort would it be to write something similar + for Hurd? Anybody fluent in asm? + And registers and signals. + gnu_srs: Signals is the key thing -- everything else we + can probably just copy. I have never/not yet looked at it, + though. + For kfreebsd it is written in C: kern_context.c, 3/4 in + one file: getcontext, setcontext, swapcontext, not makecontext. + Dunno how much assembly calls used though. + Hi, any preferences about implementing get/setcontext in + C or Asm? + gnu_srs: I think these will have to be implemented in + assembly. Based on the Linux x86 variants. + + IRC, freenode, #hurd, 2012-04-20: + + youpi: Your understanding of that is better than mine + -- the *context stuff can't be very useful at the moment, because + when the user changes uc_stack.ss_sp (which the glibc tests are + doing), we're losing access to the _hurd_threadvars. Correct? + At least the getcontext test works, the other get a + SIGILL. + others + _hurd_threadvars issue is just guessing. + tschwinge: yes, threadvars are on the stack + threadvars is not much code, it should just work, but care + has to be taken on the libpthread/libthread side, which does some + initialization + OK, that at least matches my understanding. + * `syncfs` We should be easily able to implement that one. @@ -264,6 +312,35 @@ Last reviewed up to the [[Git mirror's d40c5d54cb551acba4ef1617464760c5b3d41a14 * `timespec_get` (74033a2507841cf077e31221de2481ff30b43d51) + * `waitflags.h` (`WEXITED`, `WNOWAIT`, `WSTOPPED`, `WCONTINUED`) + + IRC, freenode, #hurd, 2012-04-20: + + in glibc, we use the generic waitflags.h which, unlike + linux's version, does not define WEXITED, WNOWAIT, WSTOPPED, + WCONTINUED + should the generic bits/waitflags.h define them anyway, + since they are posix? + well, we'd have to implement them anyway + but otherwise, I'd say yes + sure, but since glibc headers should expose at least + everything declared by posix, i thought they should be defined + anyway + that might bring bugs + some applications might be #ifdefing them + and break when they are defined but not working + i guess they would define them to 0, andd having them to + non-zero values shouldn't break them (since those values don't do + anything, so they would act as if they were 0.. or not?) + no, I mean they would do something else, not define them to + 0 + like posix/tst-waitid.c, you mean? + yes + + For specific packages: + + * [[octave]] + * Create `t/cleanup_kernel-features.h`. * Add tests from Linux kernel commit messages for `t/dup3` et al. diff --git a/open_issues/glibc/getlogin_r.mdwn b/open_issues/glibc/getlogin_r.mdwn new file mode 100644 index 00000000..9980ea1f --- /dev/null +++ b/open_issues/glibc/getlogin_r.mdwn @@ -0,0 +1,45 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!meta title="getlogin_r"]] + +[[!tag open_issue_glibc]] + + +# IRC, freenode, #hurd, 2012-04-23 + + * pinotree spots how our getlogin_r() implementation uses a static + buffer... + oO + braunr: yeah, the _r means "reentrantbutnotreally" xD + pinotree: that's amazing .. + pinotree: without having checked the actual implementation, that + doesn't sound like a problem... + caching doesn't break reentrancy. the problem with the plain + getlogin() is that a static buffer is *returned to the user* + antrik: http://paste.debian.net/164727/ + ah, caching + i don't think this is caching at all + pinotree: OK, not actually caching... but same effect as far as I + can tell + pinotree: or is it the fixed size of the temporary variable you + are concerned about? + antrik: my concern was about the "static" of the buffer used for + the get_login rpc + pinotree: I'm not sure that's a problem. can the login actually + ever change for a running process? + dunno + but if so, it would be pointless to do the rpc every time + instead of just once + pinotree: true + * pinotree would just make it non-static and be safe + pinotree: well, one might argue that allocating a KiB of stack + space is not very friendly, especially in a low-level library... + not sure what the general policy is about such things in libc diff --git a/open_issues/glibc/octave.mdwn b/open_issues/glibc/octave.mdwn new file mode 100644 index 00000000..b12b7558 --- /dev/null +++ b/open_issues/glibc/octave.mdwn @@ -0,0 +1,35 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!tag open_issue_glibc]] + + +# IRC, OFTC, #debian-hurd, 2012-04-23 + + diffing the octave i386 vs hurd-i386 build logs gives + interesting surprises + checking whether this system has an arbitrary file name length + limit... no | checking whether this system has an arbitrary + file name length limit... yes + ? + not only that + checking whether getcwd handles long file names properly... yes + | checking whether getcwd handles long file names properly... no, but it + is partly worki+ + ? + -checking whether fdopendir works... yes + +checking whether fdopendir works... no + (- is i386, + is hurd-i386) + -checking whether getlogin_r works with small buffers... yes + +checking whether getlogin_r works with small buffers... no + -checking for working mkstemp... yes + +checking for working mkstemp... no + +checking for working nanosleep... no (mishandles large + arguments) diff --git a/open_issues/gnumach_memory_management.mdwn b/open_issues/gnumach_memory_management.mdwn index f8e27e62..9feb30c8 100644 --- a/open_issues/gnumach_memory_management.mdwn +++ b/open_issues/gnumach_memory_management.mdwn @@ -2116,3 +2116,20 @@ There is a [[!FF_project 266]][[!tag bounty]] on this task. remember) use "physical" instead of "real memory" braunr: thank you. + + +# IRC, freenode, #hurd, 2012-04-22 + + youpi: tschwinge: when the slab code was added, a few new files + made into gnumach that come from my git repo and are used in other + projects as well + they're licensed under BSD upstream and GPL in gnumach, and though + it initially didn't disturb me, now it does + i think i should fix this by leaving the original copyright and + adding the GPL on top + sure, submit a patch + hm i have direct commit acces if im right + then fix it :) + do you want to review ? + I don't think there is any need to + ok diff --git a/open_issues/gnumach_page_cache_policy.mdwn b/open_issues/gnumach_page_cache_policy.mdwn new file mode 100644 index 00000000..75fcdd88 --- /dev/null +++ b/open_issues/gnumach_page_cache_policy.mdwn @@ -0,0 +1,35 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!tag open_issue_gnumach]] + + +# IRC, freenode, #hurd, 2012-04-26 + + another not-too-long improvement would be changing the page cache + policy + to drop the 4000 objects limit, you mean ? + yes + do you still have my patch attempt ? + no + let me grab that + oh i won't start it right away you know + i'll ask for it when i do + k + (otherwise i fell i'll just loose it again eh) + :) + but i imagine it's not too hard to achieve + yes + i also imagine to set a large threshold of free pages to avoid + deadlocks + which will still be better than the current situation where we + have either lots of free pages because tha max limit is reached, or lots + of pressure and system freezes :/ + yes diff --git a/open_issues/gnumach_vm_map_red-black_trees.mdwn b/open_issues/gnumach_vm_map_red-black_trees.mdwn new file mode 100644 index 00000000..d77bea32 --- /dev/null +++ b/open_issues/gnumach_vm_map_red-black_trees.mdwn @@ -0,0 +1,142 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!tag open_issue_gnumach]] + + +# IRC, freenode, #hurd, 2012-04-23 + + btw, i'm running a gnumach version using red-black trees for vm + map entries + braunr: sounds fashionable ;-) + braunr: with some perf improvement? + looks promising for our ext2fs instances showing several thousands + of map entries + youpi: i'm not using it for lookups yet + but the tree is there, maintained, used for both regular and copy + maps, and it doesn't crash + good :) + antrik: isn't it ? :) + youpi: and the diff stat is like 50/15 + braunr: what's the goal of using the fashionable trees? + antrik: speeding up lookups in address spaces + braunr: so the idea is that if we have a heavily fragmented + address space, the performance penalty is smaller? + yes + OK + I take it you gave up on attempts to actually decrease + fragmentation?... + it's not as good as reducing fragmentation, which requires + implementing a powerful merge, but it's still better + yes + it's too messy for my brain :/ + I see + oh + it will add some overhead though + I guess log(n) ? + but if there is a significant performance gain, it'll be worth it + yes + i was more thinking about the memory overhead + right now it's a linear list? + I don't think we care nowadays :) + antrik: yes + ouch + antrik: yes ... :> + the original authors expected vm maps to have like 30 entries + so they used a list for the maps, and a hash table for the + object/offset to physical page lookups + there is a small lookup cache though, which is a nice optimization + my code now uses it first, and falls back to the RB tree if the + hint didn't help + braunr: well, don't forget to check whether it actually *is* still + an optimisation, when using fashionable trees ;-) + antrik: i checked that already :) + i did the same in x15 + I see + both bsd and linux uses a similar technique + use* + (well, bsd actually use what is done in mach :) + (or perhaps the other way around... ;-) ) + i don't think so, as the bsd vm is really the mach vm + but we don't care much + oh, right... that part actually went full circle + youpi: i have a patch ready for test on machines with significant + amounts of map entries (e.g. the buildds ..) + youpi: i won't have time for tests tonight, are you interested ? + (i've been running it for 15 minutes without any issue for now) + I'd say post to the list + ok + braunr: your patch uses the rb tree for lookups, right? + braunr: the buildd using rbtree seems swift + but maybe it's just a psychologic effect :) + the chroot ext2fs already has 1392 lines in vminfo + an rbtree can't hurt there :) + braunr: it really seems faster + the reboot might have helped too + benchmarks shall say + for now, I'll just let ironforge use it + youpi: it's always fast after a reboot ;-) + sure + but still + I mean + *obviously* the reboot has helped + but it might not be all + at least it feels so + and obviously only benchmarks can say + the major benefit AIUI is rather that the slowdown happening over + time will be less noticable + +[[performance/degradation]]. + + "over time" is actually quite fast + ext2 fills up very quickly when you build a package + it went up to 1700 lines very quickly + and stabilized around there + yeah, it can be very fast under heavy load + that's why I say reboot seems not all + it's already not so fresh + with 1700 lines in vminfo + well, I don't know how much of the slowdown I'm experiencing + (after doing stuff under memory pressure) is actually related to VM map + fragmentation... + might be all, might be none, might be something in between + then try his patch + guess I should play a bit with vminfo... + well, currently I actually experience pretty little slowdown, as + for certain reasons (only indirectly related to the Hurd) I'm not running + mutt on this machine, so I don't really have memory pressure... + + +## IRC, freenode, #hurd, 2012-04-24 + + youpi: yes, it uses bst lookups + youpi: FYI, the last time i checked, one ext2fs instance had 4k+ + map entries, and another around 7.5k + (on ironforge) + + +## IRC, freenode, #hurd, 2012-04-24 + + braunr: $ sudo vminfo 624 | wc -l + 22957 + there's no way it can not help :) + youpi: 23k, that's really huge + + +## IRC, freenode, #hurd, 2012-04-26 + + youpi: any new numbers wrt the rbtree patch ? + well, buildd times are not really accurate :) + but what I can tell is that it managed to build qtwebkit fine + ok + so the patch is probably safe :) + i'll commit it soon then + I'd say feel free to, yes + thanks diff --git a/open_issues/libpthread.mdwn b/open_issues/libpthread.mdwn index 614f1271..c5054b7f 100644 --- a/open_issues/libpthread.mdwn +++ b/open_issues/libpthread.mdwn @@ -1,4 +1,5 @@ -[[!meta copyright="Copyright © 2010, 2011 Free Software Foundation, Inc."]] +[[!meta copyright="Copyright © 2010, 2011, 2012 Free Software Foundation, +Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -25,43 +26,19 @@ There is a [[!FF_project 275]][[!tag bounty]] on this task. [[!inline pages=community/gsoc/project_ideas/pthreads feeds=no]] +## IRC, freenode, #hurd, 2012-04-26 -# pthread/stubs issue w/ dlopen'ed libraries + youpi: just to be sure: even if libpthread is compiled inside + glibc (with proper symbols forwarding etc), it doesn't change that you + cannot use both cthreads and pthreads in the same app, right? -IRC, freenode, #hurd, 2010-01-24 +[[Packaging_libpthread]]. - youpi: hm, thought about the pthread/stubs issue w/ dlopen'ed - libraries - currently looks like libstdc++ on hurd links to pthread-stubs, - we're the only one with such configuration - i was looking at the gcc 4.4 patch hurd-pthread.diff, could it - be it does not set THREADLIBS in the configure.ac switch case? - that's expected - on linux the libc provides hooks itself, on hurd-i386 it's - pthread-stubs - why not explicitly link to pthread though? - because there is no strict need to, for applications that don't - need libpthread - the dlopen case is a tricky case that pthread-stubs had not thought - about - hm - what if the pthread stubs would be moved in our glibc? - that's what we should do yes - (ideally) - but for this we need to build libpthread along glibc, to get it - really working - and that's the tricky part (Makefile & such) which hasn't been done - yet - why both (stubs + actual libpthread)? - because you need the stubs to be able to call the actual libpthread - as soon libpthread gets dlopened for instance - +as - i see - (remember that nptl does this if you want to see how) - (it's the libc files in nptl/) - (and forward.c) - also if libpthreads gets integrated with glibc don't we need to - switch the hurd from cthreads then? Which has been the blocker all this - time AFAIR? - we don't _need_ to - ok + it's the same libpthread + symbol forwarding does not magically resolve that libpthread lacks + some libthread features :) + i know, i was referring about the clash between actively using + both + there'll still be the issue that only one will be initialized + and one that provides libc thread safety functions, etc. + that's what i wanted to knew, thanks :) diff --git a/open_issues/libpthread_CLOCK_MONOTONIC.mdwn b/open_issues/libpthread_CLOCK_MONOTONIC.mdwn new file mode 100644 index 00000000..f9195540 --- /dev/null +++ b/open_issues/libpthread_CLOCK_MONOTONIC.mdwn @@ -0,0 +1,58 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!meta title="libpthread: CLOCK_MONOTONIC"]] + +[[!tag open_issue_glibc open_issue_libpthread]] + +[[!message-id "201204220058.37328.toscano.pino@tiscali.it"]] + + +# IRC, freenode, #hurd- 2012-04-22 + + youpi: what i thought would be creating a + glib/hurd/hurdtime.{c,h}, adding _hurd_gettimeofday and + _hurd_clock_{gettime,settime,getres} to it and making the current .c in + sysdeps call those + yep + that's unfortunate, but that's what nptl actually does + this way we could add inside hurdtime.c the mapped time stuff + too + most probably a noobish question, but why does rt link to + pthread? + no idea :) + possibly due to aio implementation + ./sysdeps/pthread/aio_cancel.c + probably due to that + (and others) + + +## IRC, freenode, #hurd- 2012-04-23 + + pinotree: about librt vs libpthread, don't worry too much for now + libpthread can lib against the already-installed librt + it does work + youpi: wouldn't that cause a dependency loop? + pinotree: what dependency loop? + youpi: librt wouldd link to pthread, no? + and ? + I don't think it's an issue that .so link with each other + it isn't? + well, try + * pinotree never did that + but I don't think it will pose problem + well, perhaps initialization order + anyway, I now have packages built, I'll test that + pinotree: ok, it seems it doesn't work indeed :) + early in initialization + pinotree: the initialization bug might actually not be due to librt + at all + pinotree: yes, things work even with -lrt + wow diff --git a/open_issues/libpthread_dlopen.mdwn b/open_issues/libpthread_dlopen.mdwn index fb665c67..05a07ef2 100644 --- a/open_issues/libpthread_dlopen.mdwn +++ b/open_issues/libpthread_dlopen.mdwn @@ -1,4 +1,4 @@ -[[!meta copyright="Copyright © 2011 Free Software Foundation, Inc."]] +[[!meta copyright="Copyright © 2011, 2012 Free Software Foundation, Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -10,7 +10,52 @@ License|/fdl]]."]]"""]] [[!tag open_issue_libpthread]] -IRC, OFTC, #debian-hurd, 2011-07-21. +[[!toc]] + + +# IRC, freenode, #hurd, 2010-01-24 + + youpi: hm, thought about the pthread/stubs issue w/ dlopen'ed + libraries + currently looks like libstdc++ on hurd links to pthread-stubs, + we're the only one with such configuration + i was looking at the gcc 4.4 patch hurd-pthread.diff, could it + be it does not set THREADLIBS in the configure.ac switch case? + that's expected + on linux the libc provides hooks itself, on hurd-i386 it's + pthread-stubs + why not explicitly link to pthread though? + because there is no strict need to, for applications that don't + need libpthread + the dlopen case is a tricky case that pthread-stubs had not thought + about + hm + what if the pthread stubs would be moved in our glibc? + that's what we should do yes + (ideally) + but for this we need to build libpthread along glibc, to get it + really working + +[[Packaging_libpthread]]. + + and that's the tricky part (Makefile & such) which hasn't been done + yet + why both (stubs + actual libpthread)? + because you need the stubs to be able to call the actual libpthread + as soon libpthread gets dlopened for instance + +as + i see + (remember that nptl does this if you want to see how) + (it's the libc files in nptl/) + (and forward.c) + also if libpthreads gets integrated with glibc don't we need to + switch the hurd from cthreads then? Which has been the blocker all this + time AFAIR? + we don't _need_ to + ok + + +# IRC, OFTC, #debian-hurd, 2011-07-21 there's one known issue with pthreads you can't dlopen() it diff --git a/open_issues/o_exec.mdwn b/open_issues/o_exec.mdwn new file mode 100644 index 00000000..3f77a0f2 --- /dev/null +++ b/open_issues/o_exec.mdwn @@ -0,0 +1,54 @@ +[[!meta copyright="Copyright © 2012 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!meta title="O_EXEC"]] + +[[!tag open_issue_glibc open_issue_hurd]] + + +# IRC, freenode, #hurd, 2012-04-24 + + interesting, glibc on every OS except hurd (so including linux + too) does not define O_EXEC + can somebody please help me understand a POSIX behaviour? + it's about fexecve: + http://pubs.opengroup.org/onlinepubs/9699919799/functions/fexecve.html + basically, it seems to me (reading the "errors" and "application + usage" sections) that O_EXEC for open() the fd is not mandatory, and if + not used fexecve will check for file permission at call time? + because currently libdiskfs and libnetfs require the fd to be + open with O_EXEC + "Since execute permission is checked by fexecve(), the file + description fd need not have been opened with the O_EXEC flag" + this one makes it clear checking for O_EXEC is wrong + it looks like O_EXEC is only needed when you want to have files + for which only the execution permission is set + but not the read one + (i don't understand the "and write" part though) + "exec will fail if the mode of the file associated with fd does + not grant execute permission to the calling process at the time fexecve() + is called." + this one strengthens the impression you have, that fexecve indeed + checks file permissions at the time it's called + pinotree: hope it helps + so it implies the following: + O_RDONLY → exec works if the file is readable + exec works if the file is readable and/or executable (although + without read permissions you can't check it) + (well, fexecve) + O_EXEC → exec requires that the permission of the file at + fexecve() time have +x + i'd say ye so far + yes + so we need to fix lib{disk,net}fs then + seems so + enlighting, merci braunr + de rien + :) diff --git a/open_issues/packaging_libpthread.mdwn b/open_issues/packaging_libpthread.mdwn index fa3d4312..f8ef4ae7 100644 --- a/open_issues/packaging_libpthread.mdwn +++ b/open_issues/packaging_libpthread.mdwn @@ -1,4 +1,5 @@ -[[!meta copyright="Copyright © 2010, 2011 Free Software Foundation, Inc."]] +[[!meta copyright="Copyright © 2010, 2011, 2012 Free Software Foundation, +Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -10,9 +11,13 @@ License|/fdl]]."]]"""]] [[!tag open_issue_libpthread open_issue_glibc]] -IRC, #hurd, 2010-07-31 +[[!toc]] - My idea was to have a separate libpthread package. What do you think about that? + +# IRC, freenode, #hurd, 2010-07-31 + + My idea was to have a separate libpthread package. What do you + think about that? in the long term, that can't work with glibc because of the thread stub stuff @@ -21,30 +26,103 @@ IRC, #hurd, 2010-07-31 it's not really possible to keep synchronized because you have to decide which package you unpack first (when upgrading) - Hmm, how is that different if two shared libraries are in one package vs. two packages? It isn't atomic either way? Aren't sonames / versioned library packages solving that? + Hmm, how is that different if two shared libraries are in one + package vs. two packages? It isn't atomic either way? Aren't sonames / + versioned library packages solving that? ... for incompatible forward changes? that'd be a mess to maintain - Drepper doesn't have this constraint and thus adds members of private fields at will - OK, but how is it different then if the libpthread is in the Hurd package? + Drepper doesn't have this constraint and thus adds members of + private fields at will + OK, but how is it different then if the libpthread is in the + Hurd package? I'm not saying it's better to have libpthread in the Hurd package OK. - I'm saying it's useless to package it separately when Drepper makes everything to have us put it along glibc + I'm saying it's useless to package it separately when Drepper makes + everything to have us put it along glibc Then, to goal is to have it in glibc? OK. :-) - OK, I can accommodate to that. Isn't not that we'd want to switch libpthread to something else so quickly. - So our official goal is to have libpthread in glibc, at least for Debian purposese? + OK, I can accommodate to that. Isn't not that we'd want to + switch libpthread to something else so quickly. + So our official goal is to have libpthread in glibc, at least + for Debian purposese? for any port purpose Ack. - provided you're using glibc, you're deemed to ship libpthread with it + provided you're using glibc, you're deemed to ship libpthread with + it because of the strong relations Drepper puts between them - (just to remind: we already have bugs just because our current libpthread isn't bound enough to glibc: dlopen()ing a library depending on libpthread doesn't work, for instance) - yeah, pthread-stubs is linked to almost everywhere -lpthread isn't used + (just to remind: we already have bugs just because our current + libpthread isn't bound enough to glibc: dlopen()ing a library depending + on libpthread doesn't work, for instance) + yeah, pthread-stubs is linked to almost everywhere -lpthread + isn't used (would be nice to not have those issues anymore...) - So -- what do we need to put it into glibc? We can make libpthread a Git submodule (or move the code; but it's shared also for Neal's viengoos, so perhaps the submodule is better?), plus some glibc make foo, plus some other adaptions (stubs, etc.) - Does that sound about right, or am I missing something fundamental? + So -- what do we need to put it into glibc? We can make + libpthread a Git submodule (or move the code; but it's shared also for + Neal's viengoos, so perhaps the submodule is better?), plus some glibc + make foo, plus some other adaptions (stubs, etc.) + Does that sound about right, or am I missing something + fundamental? I actually don't know what a git submodule permits :) looks like a good thing for this, yes - Unfortunately I can't allocate much time at the moment to work on this. :-/ - well, as long as I know where we're going, I can know how to package stuff in Debian - That sounds like a plan to me. libpthread -> glibc as submodule. - (note: actually, the interface between glibc and the libpthread is the responsibility of the libpthread: it gives a couple of .c files to be shipped in libc.so) + Unfortunately I can't allocate much time at the moment to work + on this. :-/ + well, as long as I know where we're going, I can know how to + package stuff in Debian + That sounds like a plan to me. libpthread -> glibc as + submodule. + (note: actually, the interface between glibc and the libpthread is + the responsibility of the libpthread: it gives a couple of .c files to be + shipped in libc.so) + + +# IRC, freenode, #hurd, 2012-04-21 + + had you tried to build libpthread as a glibc addon? + youpi: No, I only know about libpthread in Hurd build system, + and libpthread stand-alone (with the Auto* stuff that I added), but not + yet as a glibc add-on. + k + I'm trying it atm + Oh, OK. + that should fix the no-add-needed issue in gcc/binutils, as well as + the pthread_threads assertion errors in threaded plugins + (once I add forward.c, but that part should not be hard) + that means also less use of pthread-stubs^ + ? + tschwinge: do you remember whether sysdeps/mach/bits/spin* are used + by anybody? + they are half-finished (no __PTHREAD_SPIN_LOCK_INITIALIZER), and + come in the way when building in glibc + also, any reason for using ia32 and not i386? glibc uses the latter + pinotree: rid of pthread-stubs yes + \o/ + youpi: You mean sysdeps/mach/i386/machine-lock.h? No idea + about that one, sorry. + I'm talking about libpthread + not glibc + Oh. + sysdeps/ia32/bits/spin-lock.h:# define + __PTHREAD_SPIN_LOCK_INITIALIZER ((__pthread_spinlock_t) 0) + Anyway, no idea about that either. + that one is meant to be used with the spin-lock.h just below + +-inline + also, I guess signal/ was for the l4 port? + youpi: I guess so. + tschwinge: I have an issue with sysdeps pt files: + sysdeps/hurd/pt-getspecific.c is not looked for by libc ; symlinking into + sysdeps/mach/hurd/pt-getspecific.c works + we don't have a non-mach sysdeps directory? + youpi: if you add sysdeps/mach/hurd/Implies containing only + "hurd", does sysdeps/hurd work? + ah, right + youpi: did it work? (and, it was needed in sysdeps/mach/hurd, or + in libpthread/sysdeps/mach/hurd?) + pinotree: it worked, it was for libpthread + good: I got libpthread built and forward working + + +## IRC, freenode, #hurd, 2012-04-23 + + phew + confirmed that moving libpthread to glibc fixes the gcc/binutils + no-add-needed issue diff --git a/open_issues/system_call_mechanism.mdwn b/open_issues/system_call_mechanism.mdwn index 5598148c..68418097 100644 --- a/open_issues/system_call_mechanism.mdwn +++ b/open_issues/system_call_mechanism.mdwn @@ -1,4 +1,4 @@ -[[!meta copyright="Copyright © 2011 Free Software Foundation, Inc."]] +[[!meta copyright="Copyright © 2011, 2012 Free Software Foundation, Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -10,8 +10,47 @@ License|/fdl]]."]]"""]] [[!tag open_issue_gnumach]] -IRC, freenode, #hurd, 2011-05-07 +[[!toc]] + + +# IRC, freenode, #hurd, 2011-05-07 very simple examples: system calls use old call gates, which are the slowest path to kernel space modern processors have dedicated instructions now + + +# IRC, freenode, #hurd, 2012-04-22 + + rah: basically, system calls are slower on mach because they use + call gates instead of newer sysenter/sysexit + braunr: sysenter/exit is a x86_64 thing + rah: apart from that, the code can't get much simpler, and *I* + know, for i have studied it, and wrote a compatible version in a clone + attempt + braunr: on a x86_64 port we'd probably use sysenter/exit + youpi: no there are 32-bits instructions, i don't remember if + they're called sysenter, it's in my thesis though so i'm sure of it :) + braunr: ah, the other part + is linux-x86 using them? + youpi: yes, glibc uses them + and does it really change much nowadays? + what is the actual difference between int 80 and sysenter? + less checking + checking what? + the idt? + ring levels for example + well, checking a ring is fast :) + depending on the original and requested levels, there are lookups + in tables + sysenter always assume 3 to 0 and 0 to 3 for sysexit + ah, also it assumes things about segments + so that indeed makes context things simpler + right + but mach doesn't uses int 0x80 + it uses an lcall + which is a bit slower from what I could read some time ago + (not sure if it's still relevant) + actually in 64bit mode I had to catch lcall from the invalid + instruction trap + perhaps it got dropped in 64bit mode diff --git a/open_issues/user-space_device_drivers.mdwn b/open_issues/user-space_device_drivers.mdwn index 70c3c6dc..25168fce 100644 --- a/open_issues/user-space_device_drivers.mdwn +++ b/open_issues/user-space_device_drivers.mdwn @@ -1,4 +1,5 @@ -[[!meta copyright="Copyright © 2009, 2011 Free Software Foundation, Inc."]] +[[!meta copyright="Copyright © 2009, 2011, 2012 Free Software Foundation, +Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -64,7 +65,7 @@ Also see [[device drivers and IO systems]]. ## System Boot -IRC, freenode, #hurd, 2011-07-27 +### IRC, freenode, #hurd, 2011-07-27 < braunr> btw, was there any formulation of the modifications required to have disk drivers in userspace ? @@ -88,6 +89,11 @@ IRC, freenode, #hurd, 2011-07-27 < Tekk_> mhm < braunr> s/disk/storage/ +### IRC, freenode, #hurd, 2012-04-25 + + btw, remember the initrd thing? + I just came across task.c in libstore/ :) + # Plan -- cgit v1.2.3