OSDN Git Service

uclinux-h8/uClibc.git
19 years agoDo not grovel about in the ELF header for the page size when
Eric Andersen [Wed, 28 Jul 2004 14:52:51 +0000 (14:52 -0000)]
Do not grovel about in the ELF header for the page size when
running on uClinux, which at runtime uses the FLAT file format.

19 years agoBe certain to include sys/sysmacros.h when major() and minor()
Eric Andersen [Tue, 27 Jul 2004 23:14:44 +0000 (23:14 -0000)]
Be certain to include sys/sysmacros.h when major() and minor()
are being used

19 years agoRemove all reference to __GLIBC_HAVE_LONG_LONG
Eric Andersen [Tue, 27 Jul 2004 22:48:36 +0000 (22:48 -0000)]
Remove all reference to __GLIBC_HAVE_LONG_LONG

19 years agoSupport linux 2.6.x wide device major/minor numbers
Eric Andersen [Mon, 26 Jul 2004 12:36:24 +0000 (12:36 -0000)]
Support linux 2.6.x wide device major/minor numbers

20 years agoFix ftruncate64 and truncate64 for powerpc and mips
Eric Andersen [Wed, 21 Jul 2004 10:00:04 +0000 (10:00 -0000)]
Fix ftruncate64 and truncate64 for powerpc and mips
 -Erik

20 years agoUpdate todo list
Eric Andersen [Sat, 17 Jul 2004 22:39:13 +0000 (22:39 -0000)]
Update todo list

20 years agoUpdate sigaction for arm
Eric Andersen [Sat, 17 Jul 2004 09:36:08 +0000 (09:36 -0000)]
Update sigaction for arm

20 years agoAdd sigaction for mips
Eric Andersen [Sat, 17 Jul 2004 00:39:14 +0000 (00:39 -0000)]
Add sigaction for mips

20 years agoFollow glibc's lead and add a gcc-3.4.1 required symbol to crt0.S.
Manuel Novoa III [Fri, 16 Jul 2004 20:23:19 +0000 (20:23 -0000)]
Follow glibc's lead and add a gcc-3.4.1 required symbol to crt0.S.

20 years agoHide the arm libfloat support for now. No longer needed for gcc 3.4.1 and
Manuel Novoa III [Fri, 16 Jul 2004 20:22:22 +0000 (20:22 -0000)]
Hide the arm libfloat support for now.  No longer needed for gcc 3.4.1 and
won't be needed for 3.3.4 either when I get some time to clean up that
toolchain which also suffers from the g++ include dir search order bug.

20 years agoRevert this stuff for now
Eric Andersen [Fri, 16 Jul 2004 13:23:21 +0000 (13:23 -0000)]
Revert this stuff for now

20 years agoPete Popov writes:
Eric Andersen [Fri, 16 Jul 2004 11:09:34 +0000 (11:09 -0000)]
Pete Popov writes:

Hi Erik,

I'm not sure why the NIOS support is not in uClibc -- perhaps the patch
was rejected or never submitted? In any case, I'm playing with some NIOS
stuff and created this patch against 0.9.26. The work was done by
Microtronix. I'm not sure who else contributed to it. It would be great
to have the NIOS support available in uClibc so developers don't have to
go searching for these bits.

Pete

20 years agoCleanup some cruft
Eric Andersen [Thu, 15 Jul 2004 10:14:10 +0000 (10:14 -0000)]
Cleanup some cruft

20 years agoupdate packaging a bit
Eric Andersen [Thu, 15 Jul 2004 09:21:07 +0000 (09:21 -0000)]
update packaging a bit

20 years agoCleanup a few of the more egregiously broken sysconf values to
Eric Andersen [Thu, 15 Jul 2004 09:09:24 +0000 (09:09 -0000)]
Cleanup a few of the more egregiously broken sysconf values to
actually match reality.  In particular, _SC_GETPW_R_SIZE_MAX and
_SC_GETGR_R_SIZE_MAX were causing us problems with programs such
as libglib, since they were always returning -1, which is a bit
smaller than the actual passwd and group max buffer sizes.
 -Erik

20 years ago_SC_PAGESIZE is standard. Some ancient legacy unix variants used _SC_PAGE_SIZE
Eric Andersen [Thu, 15 Jul 2004 08:31:31 +0000 (08:31 -0000)]
_SC_PAGESIZE is standard.  Some ancient legacy unix variants used _SC_PAGE_SIZE
instead, but we are not such a system and should not propagate such things.

20 years ago_SC_PAGESIZE is standard. Some ancient legacy unix variants used _SC_PAGE_SIZE
Eric Andersen [Thu, 15 Jul 2004 08:22:48 +0000 (08:22 -0000)]
_SC_PAGESIZE is standard.  Some ancient legacy unix variants used _SC_PAGE_SIZE
instead, but we are not such a system and should not propagate such things.

20 years agoUse MAP_PRIVATE whenever __ARCH_HAS_MMU__ is set.
Eric Andersen [Thu, 15 Jul 2004 08:06:25 +0000 (08:06 -0000)]
Use MAP_PRIVATE whenever __ARCH_HAS_MMU__ is set.
Only use MAP_SHARED when mmu-less.

20 years agoAtsushi Nemoto writes:
Eric Andersen [Thu, 15 Jul 2004 07:52:22 +0000 (07:52 -0000)]
Atsushi Nemoto writes:

This is a patch for uClibc/libc/sysdeps/linux/mips/sys/procfs.h pulled
from glibc CVS.

This fixes gdb with threaded application (with glibc environment).
Although debugging threaded programs seems not working on uClibc/mips
yet, this patch will be required when it becomes working.

This patch can be applied to any version of uClibc (including current
CVS).

20 years agoh8300 updates from Yoshinori Sato
Eric Andersen [Thu, 15 Jul 2004 07:34:00 +0000 (07:34 -0000)]
h8300 updates from Yoshinori Sato

20 years agoPatch from Yoshinori Sato to add H8/300 relocation types.
Eric Andersen [Thu, 15 Jul 2004 07:05:07 +0000 (07:05 -0000)]
Patch from Yoshinori Sato to add H8/300 relocation types.

20 years agoPeter Kjellerstedt at axis.com writes:
Eric Andersen [Thu, 15 Jul 2004 05:54:13 +0000 (05:54 -0000)]
Peter Kjellerstedt at axis.com writes:

Hello,

the attached patch should bring extra/config in line
with the Linux 2.6.7 sources.

The following are the commit messages for the respective
files from the Linux bk-repository:

checklist.c:
* fix menuconfig choice item help display

confdata.c:
* config: choice fix
* kconfig: don't rename target dir when saving config

expr.c, expr.h:
* config: disable debug prints

mconf.c:
* fix menuconfig choice item help display

menu.c:
* Kconfig: use select statements

symbol.c:
* config: choice fix
* Avoid bogus warning about recursive dependencies
* c99 struct initialiser conversions

textbox.c:
* janitor: don't init statics to 0

util.c:
* fix lxdialog behaviour

//Peter

20 years agoSigh. I got things working this morning, then checked stuff in from the wrong
Eric Andersen [Wed, 14 Jul 2004 23:07:46 +0000 (23:07 -0000)]
Sigh.  I got things working this morning, then checked stuff in from the wrong
tree.  Bad boy, No doughnut.
 -Erik

20 years agoBased on a patch from Alexandre Oliva, make sure _dl_malloc returns a nicely
Eric Andersen [Wed, 14 Jul 2004 12:27:03 +0000 (12:27 -0000)]
Based on a patch from Alexandre Oliva, make sure _dl_malloc returns a nicely
aligned pointer that may be aligned up to page_size.  Also add _dl_free,

20 years agoUpdate FAQ a bit
Eric Andersen [Mon, 5 Jul 2004 18:45:50 +0000 (18:45 -0000)]
Update FAQ a bit

20 years agoJoakim Tjernlund writes:
Eric Andersen [Sun, 27 Jun 2004 01:16:10 +0000 (01:16 -0000)]
Joakim Tjernlund writes:

Hi yet again :)

in dl-startup.c when performing boot strap relocation the following test
exists to make sure that only "_dl_" symbols are relocated:
/* We only do a partial dynamic linking right now.  The user
 is not supposed to define any symbols that start with a
 '_dl', so we can do this with confidence. */
 if (!symname || !_dl_symbol(symname)) {
        continue;
 }

However on PPC(and the other archs as well I suspect) all symbols are
"_dl_" symbols so the test is never true. The test can be removed and the
whole loop simplified(smaller). This also makes it possible to
simplify elfinterp.c

This remove the scanning of ldso.so relocs, making relocation faster.

I have tested this on PPC and it works well.
Do you think this optimization will work for the other arches as well?
I can't see why not.

     Jocke

* Tested on x86, arm, mipsel, and powerpc by Erik and works nicely
 -Erik

20 years agoDisable support for R_ARM_PC24, just like R_PPC_REL24 these should
Eric Andersen [Thu, 24 Jun 2004 07:52:08 +0000 (07:52 -0000)]
Disable support for R_ARM_PC24, just like R_PPC_REL24 these should
only occur when shared libs are not compiled with -fPIC

20 years agoLet dl-string.h include dl-sysdep.h, since it has proper include guards
Eric Andersen [Sun, 20 Jun 2004 11:10:49 +0000 (11:10 -0000)]
Let dl-string.h include dl-sysdep.h, since it has proper include guards

20 years agoWhine about R_PPC_REL24 relocations, and tell people to compile
Eric Andersen [Sun, 20 Jun 2004 05:39:04 +0000 (05:39 -0000)]
Whine about R_PPC_REL24 relocations, and tell people to compile
shared libraries with -fPIC

20 years agoJoakim Tjernlund writes:
Eric Andersen [Sat, 19 Jun 2004 21:38:41 +0000 (21:38 -0000)]
Joakim Tjernlund writes:

Make old weak symbol handling work, if someone wants to
enable it.

20 years agoJoakim Tjernlund writes:
Eric Andersen [Sat, 19 Jun 2004 21:33:04 +0000 (21:33 -0000)]
Joakim Tjernlund writes:

Since MIPS don't have COPY relocs it seems apropiate to define
DL_NO_COPY_RELOCS. This will optimize dl_find_hash somewhat.

20 years agoJoakim Tjernlund writes:
Eric Andersen [Sat, 19 Jun 2004 21:32:41 +0000 (21:32 -0000)]
Joakim Tjernlund writes:

Glibc does not use COPY class for SHN_COMMOM in MIPS. Make uClibc ldso
the same.

20 years agoBradley D. LaRonde writes:
Eric Andersen [Sat, 19 Jun 2004 19:54:40 +0000 (19:54 -0000)]
Bradley D. LaRonde writes:

dl-string.h references do_rem, but do_rem is a #define in <arch>/dl-sysdep.h
which is not included by dl-sysdep.h.  This causes a problem in libdl:

    In file included from ../../ldso/include/ldso.h:27, from libdl.c:33:
    ../../ldso/include/dl-string.h: In function `_dl_simple_ltoa':
    ../../ldso/include/dl-string.h:216: warning: implicit declaration of
function `do_rem'

Include dl-sysdep.h in dl-string.h before using do_rem.

20 years agonote new mailing list posting policy
Eric Andersen [Sat, 19 Jun 2004 06:28:24 +0000 (06:28 -0000)]
note new mailing list posting policy

20 years agoStephen Hemminger from osdl dot org writes:
Eric Andersen [Sat, 12 Jun 2004 08:41:18 +0000 (08:41 -0000)]
Stephen Hemminger from osdl dot org writes:

Rather than copy more code back into the bridge-utilities, how about applying this
change to uClibc?  I hate when packages get cluttered for workarounds for other
incompatibilities.

The problem is that SIOCGIFCONF only lists interfaces that have IP addresses, so it
doesn't find the other interfaces that are being used for bridging. It could be fixed
in the kernel to return all interfaces, but then something else might break; and still
it mean a kernel update for the 2.4 users.

The whole use of ifindex in the bridge API is a bad idea. But we probably have to live
with it for compatibility.

Patch against uClibc 0.9.26

20 years agoJakub Bogusz from pld-linux dot org writes:
Eric Andersen [Sat, 12 Jun 2004 08:38:39 +0000 (08:38 -0000)]
Jakub Bogusz from pld-linux dot org writes:

Hello,

I managed to get ldso (and thus shared linking to uClibc) to work on
sparc (actually sparc64 kernel with 32-bit userspace), at least on
simple "hello world" program (more complex ones not tested).

Some notes on attached patch (against 0.9.26, would require some work
to apply on current CVS - but I tested 0.9.26, not CVS):

- ELF magic cannot be examined by _dl_strncmp so early, probably because of
  string constant, like on ppc/mips/sh
  (note that early SEND_STDERR still crashes when trying to do _dl_strlen
   - I suppose that string constants require relocation; but adding
   load_addr didn't help, just ELF header was displayed instead of crash)
- mmap() is syscall6 like on ppc/mips/sh, not old i386 mmap()
- for generic sparc (i.e. not sparcv8/sparcv9) gcc produces .udiv/.urem
  calls for unsigned integer / and % operators - so these operations
  must be avoided. I copied do_rem definition from arm header.
  But / and % are used also in _dl_simple_ltoa() and
  _dl_simple_ltoahex(); in ltoahex gcc optimizes it to shifts (but
  I think it's safer to use shifts explicitly, not rely on
  optimization...).
  I changed % in ltoa to do_rem, but as there was no do_div definition,
  I changed all "%d" specifiers to "%x" to avoid crashes (this changes
  wouldn't be needed if _dl_simple_ltoa() were fixed to not use
  division on sparc).
- "#define SOLARIS_COMPATIBLE" in ld_sysdep.h broke ldso on Linux
  because of redefining _dl_linux_resolve only in some places (it was
  still referenced in INIT_GOT before redefinition). So
  _dl_linux_resolve redefinition should be moved before INIT_GOT
  definition or removed.
- sparc64 kernel requires mmap() addresses to be aligned to 8192, not
  4096, otherwise mmap() call failed
- reloc_entry must be shifted by 10, not 12 (I found similar operation
  in glibc sources)

Aside of sparc-specific fixes:
- I moved some _dl_dprintf()s inside if(_dl_debug_*) conditions (to avoid
  debugging messages when LD_DEBUG is not defined)
- it seems that there was possible off-by-one in ltoa and ltoahex?
  they are called with char[22] as 1st argument, and then '\0' is stored
  in local[22] (_before_ p decrementation)... or am I missing something?
  If not, fix is included in patch.

20 years agoPhilip Craig at snapgear dot com writes:
Eric Andersen [Sat, 12 Jun 2004 06:12:21 +0000 (06:12 -0000)]
Philip Craig at snapgear dot com writes:

Strlen was counting the last 3 bytes incorrectly for big endian arm.

20 years agoPer comments from Bernhard Rosenkraenzer, adjust gcc 3.3.x generated asm
Eric Andersen [Sat, 12 Jun 2004 05:52:22 +0000 (05:52 -0000)]
Per comments from Bernhard Rosenkraenzer, adjust gcc 3.3.x generated asm
with s/i686.get_pc_thunk.bx/get_pc_thunk_bx/g to make gcc 3.4 happy.

20 years agoOops. Missed a spot.
Eric Andersen [Sat, 12 Jun 2004 05:43:09 +0000 (05:43 -0000)]
Oops.  Missed a spot.

20 years agoscrap the goofy time + getpid() based "randomness" and use
Eric Andersen [Sun, 6 Jun 2004 08:08:25 +0000 (08:08 -0000)]
scrap the goofy time + getpid() based "randomness" and use
the real thing.

20 years agoRemoved unused file.
Tobias Anderberg [Mon, 24 May 2004 10:03:59 +0000 (10:03 -0000)]
Removed unused file.

20 years agoCope when the __NR_get_kernel_syms is undefined (i.e. 2.6.x arm-linux)
Eric Andersen [Sat, 15 May 2004 08:21:58 +0000 (08:21 -0000)]
Cope when the __NR_get_kernel_syms is undefined (i.e. 2.6.x arm-linux)

20 years agoFix changed link into the mailing list archive.
Manuel Novoa III [Fri, 14 May 2004 19:49:40 +0000 (19:49 -0000)]
Fix changed link into the mailing list archive.

20 years agoJoakim Tjernlund writes:
Eric Andersen [Fri, 14 May 2004 11:12:25 +0000 (11:12 -0000)]
Joakim Tjernlund writes:

Hi Erik

It seems to me that __pthread_once and __pthread_initialize_minimal could be made
WEAKs with no stub. The code in rpc_thread.c and __uClibc_main.c appears to expect this.
Also, __pthread_return_0 __pthread_return_1 and __pthread_return_void can be static, not to
pollute the name space.

  Jocke

20 years agoAlexandre Oliva writes:
Eric Andersen [Fri, 14 May 2004 10:59:09 +0000 (10:59 -0000)]
Alexandre Oliva writes:

Here are a number of minor changes to FR-V-specific bits of the uClibc
port:

- I've adjusted the definition of _dl_mmap to cope with the fact that
  there will be a definition for an mmap2-only system.

- We don't have COPY relocs, so optimize the copy reloc-related code
  away.

- Change the page size to 16KiB, to match the ABI spec, and not a
  stale value I'd copied from a linker config file.  Oops.

- Fix error handling in clone and vfork; parts of the changes by David
  Howells <dhowells@redhat.com>

- Rearrange includes in crtreloc.c.

- Change the API of __syscall_error, to reduce code size.

- Improve __syscall_return in terms of code size, so as to enable
  tail-calling of __syscall_error, at least within libc.  Ideally,
  __syscall_error should be hidden within libc.so, but this didn't
  work because of libdl.  I haven't looked into why, and figured I'd
  leave it visible for now.

- Rename enumerators and macros in sys/ucontext.h to reduce namespace
  pollution.

20 years ago__data_start needs to be added to all crt0.S files that don't currently
Eric Andersen [Fri, 14 May 2004 10:51:16 +0000 (10:51 -0000)]
__data_start needs to be added to all crt0.S files that don't currently
have it.  It is used by the boehm gc, amoung other things.

20 years agoKill off the now obsolete __libc_vfork symbol
Eric Andersen [Fri, 14 May 2004 10:39:05 +0000 (10:39 -0000)]
Kill off the now obsolete __libc_vfork symbol

20 years agoAlexandre Oliva writes:
Eric Andersen [Fri, 14 May 2004 10:29:45 +0000 (10:29 -0000)]
Alexandre Oliva writes:

This patch introduces optimized versions of memcpy and memset for
frv.

20 years agoPatch from Alexandre Oliva:
Eric Andersen [Fri, 14 May 2004 10:26:05 +0000 (10:26 -0000)]
Patch from Alexandre Oliva:

On Mar 20, 2004, Erik Andersen <andersen@codepoet.org> wrote:

> If you supply a FR-V specific link.h header into
> libc/sysdeps/linux/frv/ then you can do the same sortof
> thing that mips does i.e. with sgidefs.h in the headers
> target in libc/sysdeps/linux/mips/Makefile

Thanks, this patch implements your suggestion.

20 years agoAlexandre Oliva writes:
Eric Andersen [Fri, 14 May 2004 10:22:10 +0000 (10:22 -0000)]
Alexandre Oliva writes:

A program that requests __pthread_sig_debug to be blocked will
self-deadlock when it requests a thread to be created, because the
debugger (rda or gdb) will never get the signal, so it won't wake up
the pthread manager as expected.

This patch fixes it.

20 years agoupdate libdl so it also groks the new _dl_errno
Eric Andersen [Thu, 13 May 2004 09:23:34 +0000 (09:23 -0000)]
update libdl so it also groks the new _dl_errno

20 years agoAdd define for SHM_HUGETLB
Eric Andersen [Thu, 13 May 2004 05:53:37 +0000 (05:53 -0000)]
Add define for SHM_HUGETLB

20 years agoBradley D. LaRonde writes:
Eric Andersen [Thu, 13 May 2004 00:19:47 +0000 (00:19 -0000)]
Bradley D. LaRonde writes:

Oops, this adds the page alignment offset to the mprotected size.

Regards,
Brad

20 years agoPatch from Bradley D. LaRonde:
Eric Andersen [Wed, 12 May 2004 23:11:30 +0000 (23:11 -0000)]
Patch from Bradley D. LaRonde:
ld.so mprotect rwx .dyanmic segment for mips DT_DEBUG
Allow writing debug_addr into the .dynamic segment.
Even though the program header is marked RWE, the kernel
gives it to us rx.

20 years agoAdd a local '_dl_errno' to be used by syscalls in ldso, allowing
Eric Andersen [Wed, 12 May 2004 22:54:59 +0000 (22:54 -0000)]
Add a local '_dl_errno' to be used by syscalls in ldso, allowing
useful syscall failure diagnostics.

20 years agoAs noticed by the sharp eyed Joakim Tjernlund, I stupidly
Eric Andersen [Tue, 11 May 2004 19:13:26 +0000 (19:13 -0000)]
As noticed by the sharp eyed Joakim Tjernlund, I stupidly
hard coded 4096 instead of PAGE_SIZE.  Because I'm an idiot.

20 years agoSetup a _dl_pagesize value for use by staticly linked apps
Eric Andersen [Tue, 11 May 2004 11:35:10 +0000 (11:35 -0000)]
Setup a _dl_pagesize value for use by staticly linked apps

20 years agomake certain that getpagesize() returns correct the value for mips
Eric Andersen [Tue, 11 May 2004 11:14:58 +0000 (11:14 -0000)]
make certain that getpagesize() returns correct the value for mips
by extracting the value from the ELF header.

20 years agoPurge trailing white space
Eric Andersen [Tue, 11 May 2004 10:45:25 +0000 (10:45 -0000)]
Purge trailing white space

20 years agoBradley D. LaRonde, brad at laronde dot org writes:
Eric Andersen [Tue, 11 May 2004 07:21:47 +0000 (07:21 -0000)]
Bradley D. LaRonde, brad at laronde dot org writes:

If I change that ELF_RTYPE_CLASS_COPY to ELF_RTYPE_CLASS_PLT to tell
_dl_find_hash to ignore stubs when resolving undefined functions without
stubs, the dlopen tests all pass.  dlopen gets a pointer to the libc.so
malloc instead of a pointer to the libpthread malloc stub.  Yay!  :-)

20 years agoSave r0 before modifying it.
Tobias Anderberg [Mon, 10 May 2004 12:59:17 +0000 (12:59 -0000)]
Save r0 before modifying it.

20 years agoObligatory forgotten file...
Manuel Novoa III [Sat, 8 May 2004 05:20:24 +0000 (05:20 -0000)]
Obligatory forgotten file...

20 years agoChange clock() to allow wrapping.
Manuel Novoa III [Sat, 8 May 2004 05:12:14 +0000 (05:12 -0000)]
Change clock() to allow wrapping.
Add timegm() function.
Make lookup_tzname() static (as it should have been).
Have strftime() get timezone information from the passed struct
  for the %z and %Z conversions when using struct tm extensions.

20 years agoPeter Mazinger asked for this several months ago.
Manuel Novoa III [Fri, 7 May 2004 10:53:57 +0000 (10:53 -0000)]
Peter Mazinger asked for this several months ago.

20 years agoPatch from Yoshinori Sato <ysato@users.sourceforge.jp>.
Manuel Novoa III [Fri, 7 May 2004 10:30:59 +0000 (10:30 -0000)]
Patch from Yoshinori Sato <ysato@users.sourceforge.jp>.

20 years ago"Fix" this so that at least things will link when threading is disabled.
Manuel Novoa III [Fri, 7 May 2004 10:22:52 +0000 (10:22 -0000)]
"Fix" this so that at least things will link when threading is disabled.
Also fix the sizeof() issue since the change to a dynamicly allocated buf.
Note!  This is still broken wrt threading, but so is the glibc version.
I'm just commiting this for new until I can test my rewrite.

20 years agoFix a couple of issues in configurations I hadn't tested.
Manuel Novoa III [Fri, 7 May 2004 10:19:14 +0000 (10:19 -0000)]
Fix a couple of issues in configurations I hadn't tested.

20 years agoAdd missing include.
Manuel Novoa III [Fri, 7 May 2004 10:00:36 +0000 (10:00 -0000)]
Add missing include.

20 years agoUCLIBC_COMPLETELY_PIC does nothing, so kill it
Eric Andersen [Fri, 7 May 2004 06:52:18 +0000 (06:52 -0000)]
UCLIBC_COMPLETELY_PIC does nothing, so kill it

20 years agoChange the global 'buf' to 'servbuf' since some functions in here use
Eric Andersen [Fri, 7 May 2004 05:26:31 +0000 (05:26 -0000)]
Change the global 'buf' to 'servbuf' since some functions in here use
a local named 'buf' and we want to avoid shadowing that.

20 years agoAlexandre Oliva writes:
Eric Andersen [Thu, 6 May 2004 09:16:11 +0000 (09:16 -0000)]
Alexandre Oliva writes:

The vfork() wrapper defined in libpthread, that's used to run
pthread_atfork()-registered handlers, is not only a very bad idea,
it's broken and useless.  Here's the rationale:

[---------snip----------]

Since the implementation as it stands is broken (linking a program
that vfork()s and exec()s on the child and wait()s on the parent works
unless you happen to link with libpthread), and I can't think of
any workable solution, I suggest that we simply remove the vfork()
overrider in the non-MMU case.  Yes, we might lose some small amount
of functionality here, but it's not like people running uClinux expect
anything resembling actual fork() to work.

20 years agoFix screwed up formatting
Eric Andersen [Wed, 5 May 2004 06:56:30 +0000 (06:56 -0000)]
Fix screwed up formatting

20 years agoJeroen Dobbelaere writes:
Eric Andersen [Wed, 5 May 2004 06:56:03 +0000 (06:56 -0000)]
Jeroen Dobbelaere writes:

Because variables are linked to fixed registers, there is a problem in :
     (*__errno_location())=(-_r0);

As __errno_location() uses r0 to return the address of the errno location,
the negated address will be assigned instead of the error code.

Attached patch will resolve this.

20 years agoMinor .deb updates
Eric Andersen [Fri, 30 Apr 2004 22:59:38 +0000 (22:59 -0000)]
Minor .deb updates

20 years agoThese headers should not be included in uClibc -- they should be
Eric Andersen [Fri, 30 Apr 2004 22:20:01 +0000 (22:20 -0000)]
These headers should not be included in uClibc -- they should be
part of the ppp instead.

20 years agoIt turns out that asm/reg.h is missing from the 2.6.x kernel headers,
Eric Andersen [Fri, 23 Apr 2004 21:39:34 +0000 (21:39 -0000)]
It turns out that asm/reg.h is missing from the 2.6.x kernel headers,
which prevents libthread_db (used by gdb) from compiling.  Include a
copy within include/sys/user.h for mips, per what was done in glibc.

20 years agoblah
Eric Andersen [Tue, 20 Apr 2004 18:51:24 +0000 (18:51 -0000)]
blah

20 years agoFix grammar and spelling
Eric Andersen [Tue, 20 Apr 2004 17:15:53 +0000 (17:15 -0000)]
Fix grammar and spelling

20 years agoMention uwoody
Eric Andersen [Tue, 20 Apr 2004 17:09:40 +0000 (17:09 -0000)]
Mention uwoody

20 years agoCope with gcc 3.4's more aggressive persuit of attribute unused
Eric Andersen [Tue, 20 Apr 2004 06:33:11 +0000 (06:33 -0000)]
Cope with gcc 3.4's more aggressive persuit of attribute unused

20 years agoWe don't need these...
Eric Andersen [Tue, 20 Apr 2004 06:27:59 +0000 (06:27 -0000)]
We don't need these...

20 years agoFix typo.
Manuel Novoa III [Wed, 14 Apr 2004 16:21:27 +0000 (16:21 -0000)]
Fix typo.

20 years agoFix a bug noticed by Oleg I. Vdovikin, oleg at cs dot msu dot su.
Eric Andersen [Tue, 6 Apr 2004 12:19:22 +0000 (12:19 -0000)]
Fix a bug noticed by Oleg I. Vdovikin, oleg at cs dot msu dot su.
This code in uClibc would attempt to set the default facility if
none was specified.  Except none being specified is 0, which is
also LOG_KERN, therefore klogd could never get kernel messages
logged as LOG_KERN.

20 years agoglibc does not provide pppio.h, and our doing so conflicts
Eric Andersen [Wed, 24 Mar 2004 22:21:07 +0000 (22:21 -0000)]
glibc does not provide pppio.h, and our doing so conflicts
with the ppp package

20 years agoPer patch from Stefan Holst, do not try to compile these
Eric Andersen [Fri, 19 Mar 2004 20:53:54 +0000 (20:53 -0000)]
Per patch from Stefan Holst, do not try to compile these
syscalls if they are not supported.

20 years agoPatch from Stefan Holst to fix argv[0]
Eric Andersen [Fri, 19 Mar 2004 20:38:10 +0000 (20:38 -0000)]
Patch from Stefan Holst to fix argv[0]

20 years agoBased on a patch from Alexandre Oliva, allow uClibc to compile
Eric Andersen [Fri, 19 Mar 2004 12:03:30 +0000 (12:03 -0000)]
Based on a patch from Alexandre Oliva, allow uClibc to compile
with __NR_mmap is not available (i.e. only __NR_mmap2)

20 years agoAlexandre Oliva writes:
Eric Andersen [Thu, 18 Mar 2004 11:43:18 +0000 (11:43 -0000)]
Alexandre Oliva writes:

This patch arranges for the .so files in say /usr/lib to be soft links
to ../../lib, instead of to /some/arbitrary/pathname/lib.  This
enables seamless relocation of a toolchain containing the development
and run time trees in a sys-root.

20 years agoAlexandre Oliva writes:
Eric Andersen [Thu, 18 Mar 2004 11:41:12 +0000 (11:41 -0000)]
Alexandre Oliva writes:

I added this function after I posted the last version of the FR-V
patch.  Add syscall.c.

20 years agoSupply '__bzero' with 'bzero' as just a weak alias (since it is not SuSv3)
Eric Andersen [Thu, 18 Mar 2004 11:38:02 +0000 (11:38 -0000)]
Supply '__bzero' with 'bzero' as just a weak alias (since it is not SuSv3)

20 years agoBased on a patch from Alexandre Oliva, remove all reference to 'bzero' (which
Eric Andersen [Thu, 18 Mar 2004 11:28:51 +0000 (11:28 -0000)]
Based on a patch from Alexandre Oliva, remove all reference to 'bzero' (which
is not a SuSv3 symbol).  Rather than using __bzero internally per Alexandre's
original patch, use memset instead.

20 years agoLook in /usr/X11R6/lib as well
Eric Andersen [Thu, 18 Mar 2004 11:17:14 +0000 (11:17 -0000)]
Look in /usr/X11R6/lib as well

20 years agoReduce memory used by static buffers and allocate that memory dynamicly
Eric Andersen [Thu, 18 Mar 2004 11:12:34 +0000 (11:12 -0000)]
Reduce memory used by static buffers and allocate that memory dynamicly
instead.  Based on an initial patch from Tobias Anderberg, but reworked.  I
asked Tobias to look into doing something more like what is done in busybox,
but that proved to be a pain.

One possible concern is that these buffers will probably show up as
memory leaks i.e. with valgrind.  Perhaps we should add in an atexit
call to free this memory right after we allocate it?

20 years agoJoakim Tjernlund writes:
Eric Andersen [Thu, 18 Mar 2004 08:00:40 +0000 (08:00 -0000)]
Joakim Tjernlund writes:

Hi

I just noticed that gcc has an "b"(Address base register) operand that
will match all "r" registers but r0. It is a better fix then adding
r0 to the clobber list.

What do you think?

20 years agoCope gracefully with missing module syscalls
Eric Andersen [Fri, 12 Mar 2004 00:21:20 +0000 (00:21 -0000)]
Cope gracefully with missing module syscalls

20 years agoRewrite the err/warn functions as they were broken (__noreturn__ funcs
Manuel Novoa III [Thu, 11 Mar 2004 10:10:53 +0000 (10:10 -0000)]
Rewrite the err/warn functions as they were broken (__noreturn__ funcs
were returning).  Anyway, also make them threadsafe and smaller.  The
error.c file still needs work.

20 years agoBug fix: gethostbyname2_r would fail if /etc/host was missing.
Manuel Novoa III [Wed, 10 Mar 2004 20:43:23 +0000 (20:43 -0000)]
Bug fix: gethostbyname2_r would fail if /etc/host was missing.
Bug fix: gethostbyname_r checked errno without first setting it to a
         known value.

20 years agoMore detail about malloc-simple
Eric Andersen [Tue, 9 Mar 2004 22:33:56 +0000 (22:33 -0000)]
More detail about malloc-simple

20 years agoUpdate todo list indicating the items remaining before 0.9.27
Eric Andersen [Tue, 9 Mar 2004 21:57:12 +0000 (21:57 -0000)]
Update todo list indicating the items remaining before 0.9.27

20 years agoFix comment
Eric Andersen [Mon, 8 Mar 2004 20:57:21 +0000 (20:57 -0000)]
Fix comment