]> Zhao Yanbai Git Server - minix.git/log
minix.git
17 years agoUse temporary binary directory as bootstrap
Ben Gras [Thu, 8 Feb 2007 16:18:48 +0000 (16:18 +0000)]
Use temporary binary directory as bootstrap

17 years agoExtend cdprobe probe list to c1, and reorder minors to do most-likely
Ben Gras [Thu, 8 Feb 2007 15:56:58 +0000 (15:56 +0000)]
Extend cdprobe probe list to c1, and reorder minors to do most-likely
first.

Make at_wini include instance number in error messages.

17 years agosupply instance to 2nd at_wini instance.
Ben Gras [Thu, 8 Feb 2007 14:23:03 +0000 (14:23 +0000)]
supply instance to 2nd at_wini instance.
requires a little cooperation from at_wini.

17 years agoStart a 2nd copy of at_wini, for /dev/c1*. This requires a slightly
Ben Gras [Thu, 8 Feb 2007 14:04:59 +0000 (14:04 +0000)]
Start a 2nd copy of at_wini, for /dev/c1*. This requires a slightly
larger rs.

17 years agoMake /dev/c1* device nodes on disk and on the boot ramdisk.
Ben Gras [Thu, 8 Feb 2007 13:51:35 +0000 (13:51 +0000)]
Make /dev/c1* device nodes on disk and on the boot ramdisk.

 . include c1* nodes in std in MAKEDEV
 . this requires a slightly larger shell
 . this requires a larger blocksize on the boot ramdisk (to fit
   /dev/ in direct blocks for mkfs with a proto file)
 . also more inodes and kB's on the boot ramdisk

17 years agoSanity check in clock - process is supposed to be runnable when it's
Ben Gras [Thu, 8 Feb 2007 12:59:29 +0000 (12:59 +0000)]
Sanity check in clock - process is supposed to be runnable when it's
interrupted.

17 years agoRetired DEV_{READ,WRITE,GATHER,SCATTER,IOCTL} (safe versions *_S are to
Ben Gras [Wed, 7 Feb 2007 16:22:19 +0000 (16:22 +0000)]
Retired DEV_{READ,WRITE,GATHER,SCATTER,IOCTL} (safe versions *_S are to
be used and drivers should never receieve these 'unsafe' variants
any more).

17 years agoMostly bugfixes of bugs triggered by the test set.
Ben Gras [Thu, 1 Feb 2007 17:50:02 +0000 (17:50 +0000)]
Mostly bugfixes of bugs triggered by the test set.

bugfixes:
 SYSTEM:
 . removed
        rc->p_priv->s_flags = 0;
   for the priv struct shared by all user processes in get_priv(). this
   should only be done once. doing a SYS_PRIV_USER in sys_privctl()
   caused the flags of all user processes to be reset, so they were no
   longer PREEMPTIBLE. this happened when RS executed a policy script.
   (this broke test1 in the test set)

 VFS/MFS:
 . chown can change the mode of a file, and chmod arguments are only
   part of the full file mode so the full filemode is slightly magic.
   changed these calls so that the final modes are returned to VFS, so
   that the vnode can be kept up-to-date.
   (this broke test11 in the test set)

 MFS:
 . lookup() checked for sizeof(string) instead of sizeof(user_path),
   truncating long path names
   (caught by test 23)
 . truncate functions neglected to update ctime
   (this broke test16)

 VFS:
 . corner case of an empty filename lookup caused fields of a request
   not to be filled in in the lookup functions, not making it clear
   that the lookup had failed, causing messages to garbage processes,
   causing strange failures.
   (caught by test 30)
 . trust v_size in vnode when doing reads or writes on non-special
   files, truncating i/o where necessary; this is necessary for pipes,
   as MFS can't tell when a pipe has been truncated without it being
   told explicitly each time.
   when the last reader/writer on a pipe closes, tell FS about
   the new size using truncate_vn().
   (this broke test 25, among others)
 . permission check for chdir() had disappeared; added a
   forbidden() call
   (caught by test 23)

new code, shouldn't change anything:
 . introduced RTS_SET, RTS_UNSET, and RTS_ISSET macro's, and their
   LOCK variants. These macros set and clear the p_rts_flags field,
   causing a lot of duplicated logic like

       old_flags = rp->p_rts_flags;            /* save value of the flags */
       rp->p_rts_flags &= ~NO_PRIV;
       if (old_flags != 0 && rp->p_rts_flags == 0) lock_enqueue(rp);

   to change into the simpler

       RTS_LOCK_UNSET(rp, NO_PRIV);

   so the macros take care of calling dequeue() and enqueue() (or lock_*()),
   as the case may be). This makes the code a bit more readable and a
   bit less fragile.
 . removed return code from do_clocktick in CLOCK as it currently
   never replies
 . removed some debug code from VFS
 . fixed grant debug message in device.c

preemptive checks, tests, changes:
 . added return code checks of receive() to SYSTEM and CLOCK
 . O_TRUNC should never arrive at MFS (added sanity check and removed
   O_TRUNC code)
 . user_path declared with PATH_MAX+1 to let it be null-terminated
 . checks in MFS to see if strings passed by VFS are null-terminated

 IS:
 . static irq name table thrown out

17 years agoAlso exclude /home
Ben Gras [Thu, 1 Feb 2007 17:10:03 +0000 (17:10 +0000)]
Also exclude /home

17 years agosimpler /usr/bin/vi check
Ben Gras [Thu, 1 Feb 2007 17:09:49 +0000 (17:09 +0000)]
simpler /usr/bin/vi check

17 years agoOnly install /usr/bin/vi as a hardlink to elvis if there's nothing already
Ben Gras [Wed, 31 Jan 2007 13:00:39 +0000 (13:00 +0000)]
Only install /usr/bin/vi as a hardlink to elvis if there's nothing already
there (so e.g. nvi isn't overwritten).

17 years agoput mfs in /sbin, where mount expects it.
Ben Gras [Wed, 24 Jan 2007 15:50:04 +0000 (15:50 +0000)]
put mfs in /sbin, where mount expects it.

17 years ago. no .vimrc in standard user's dirs
Ben Gras [Mon, 22 Jan 2007 17:08:36 +0000 (17:08 +0000)]
. no .vimrc in standard user's dirs
. install a vi hardlink to elvis

17 years agoDon't reply to the caller on RS_DOWN until process is actually dead -
Ben Gras [Mon, 22 Jan 2007 16:44:03 +0000 (16:44 +0000)]
Don't reply to the caller on RS_DOWN until process is actually dead -
otherwise (e.g.) mounts right after an unmount of the same device don't
work (duplicate label).

17 years agoNo /usr/src.* shenanigans any more.
Ben Gras [Mon, 22 Jan 2007 15:47:40 +0000 (15:47 +0000)]
No /usr/src.* shenanigans any more.

17 years agoInstall sources in /usr/bigsrc, not /usr/src/commands.
Ben Gras [Mon, 22 Jan 2007 15:42:40 +0000 (15:42 +0000)]
Install sources in /usr/bigsrc, not /usr/src/commands.

17 years agoramdisk can be set only once.
Ben Gras [Mon, 22 Jan 2007 15:37:26 +0000 (15:37 +0000)]
ramdisk can be set only once.

17 years agovfs/mount/rs/service changes:
Ben Gras [Mon, 22 Jan 2007 15:25:41 +0000 (15:25 +0000)]
vfs/mount/rs/service changes:
 . changed umount() and mount() to call 'service', so that it can include
   a custom label, so that umount() works again (RS slot gets freed now).
   merged umount() and mount() into one file to encode keep this label
   knowledge in one file.
 . removed obsolete RS_PID field and RS_RESCUE rescue command
 . added label to RS_START struct
 . vfs no longer does kill of fs process on unmount (which was failing
   due to RS_PID request not working)
 . don't assume that if error wasn't one of three errors, that no error
   occured in vfs/request.c
mfs changes:
 . added checks to copy statements to truncate copies at buffer sizes
   (left in debug code for now)
 . added checks for null-terminatedness, if less than NAME_MAX was copied
 . added checks for copy function success
is changes:
 . dump rs label
drivers.conf changes:
 . added acl for mfs so that mfs can be started with 'service start',
   so that a custom label can be provided

17 years agoAdded register clobber warning to read_tsc.
Ben Gras [Mon, 22 Jan 2007 15:15:30 +0000 (15:15 +0000)]
Added register clobber warning to read_tsc.

17 years agoMany more IDs.
Philip Homburg [Fri, 19 Jan 2007 17:27:51 +0000 (17:27 +0000)]
Many more IDs.

17 years agoPCI IDs for Accton SMC2-1211TX
Philip Homburg [Thu, 18 Jan 2007 11:40:13 +0000 (11:40 +0000)]
PCI IDs for Accton SMC2-1211TX

17 years agoremove debug (%d) in fsck
Ben Gras [Tue, 16 Jan 2007 16:52:51 +0000 (16:52 +0000)]
remove debug (%d) in fsck

17 years ago. added checks for buffer sizes in sys_datacopy() functions in mfs,
Ben Gras [Tue, 16 Jan 2007 14:50:10 +0000 (14:50 +0000)]
. added checks for buffer sizes in sys_datacopy() functions in mfs,
  print debug message if copy is truncated
. increased buffer in lookup() to be PATH_MAX instead of NAME_MAX
. sanity check in fetch_name() in vfs to see if name fits, and
  is null-terminated
. first check i < NAME_MAX, then string[i] in search_dir, as we're
  not supposed to look at string[NAME_MAX]

17 years ago. added super-user check for mount
Ben Gras [Tue, 16 Jan 2007 13:57:35 +0000 (13:57 +0000)]
. added super-user check for mount
. corrected device match for unmount (otherwise unmount would
  proceed with bogus mount slot, often sending messages to 0 (PM))
. added some sanity checking to fs process number
. made fs_sendrec PRIVATE to request.c

17 years agoFsck needs more memory for bigger filesystems
Ben Gras [Tue, 16 Jan 2007 12:59:13 +0000 (12:59 +0000)]
Fsck needs more memory for bigger filesystems

17 years agoDon't panic if opening a block device fails.
Ben Gras [Fri, 12 Jan 2007 17:16:51 +0000 (17:16 +0000)]
Don't panic if opening a block device fails.

17 years agoTruncate s_max_size at LONG_MAX, now that off_t is signed.
Ben Gras [Fri, 12 Jan 2007 16:53:12 +0000 (16:53 +0000)]
Truncate s_max_size at LONG_MAX, now that off_t is signed.

17 years ago. removed readclock command and cmos driver.
Ben Gras [Fri, 12 Jan 2007 16:35:04 +0000 (16:35 +0000)]
. removed readclock command and cmos driver.
. replaced by a readclock 'driver' that runs once, a re-imported version
  of the minix 2.0.4 readclock command.
. this has also restored cmos writing.
. readclock wrapper script calls service command to run /bin/readclock.drv
  once.

17 years agoUse bitwise not instead of logical not on PIE flag when disabling periodic
Ben Gras [Fri, 12 Jan 2007 16:33:41 +0000 (16:33 +0000)]
Use bitwise not instead of logical not on PIE flag when disabling periodic
interrupts to avoid clobbering register B. This seems to have fixed the
corrupting-CMOS bug when enabling profiling.

17 years agouse servers/inet/mq.[ch] to queue messages using mq_queue() in
Ben Gras [Fri, 12 Jan 2007 13:33:12 +0000 (13:33 +0000)]
use servers/inet/mq.[ch] to queue messages using mq_queue() in
libdriver.  at_wini now queues messages it can't handle it receives when
waiting for an interrupt. this way it can do receive(ANY) and timeouts
should be working again (were broken for VFS, as with the advent of VFS,
at_wini could get requests from a filesystem while it was waiting for an
interrupt - as a hack, the receive() was changed to receive(HARDWARE)).

Added mq.c to libdriver, and made libdriver an actual library that
drivers link with -L../libdriver -ldriver. (So adding files, if
necessary, is easier next time.)

17 years agoIn some configurations of non-working networking, packman will hang a
Ben Gras [Thu, 11 Jan 2007 14:35:29 +0000 (14:35 +0000)]
In some configurations of non-working networking, packman will hang a
very long time or forever retrieving the packages list. It's tricky to
test connectivity automatically, so ask every time.  Reported by
Stanislav Kapustin <kapustin_stanislav@hotmail.com>.

17 years agoZlib is now an optional package.
Ben Gras [Tue, 9 Jan 2007 17:08:31 +0000 (17:08 +0000)]
Zlib is now an optional package.

17 years agoTake arch+version specific packages List file.
Ben Gras [Mon, 8 Jan 2007 15:10:27 +0000 (15:10 +0000)]
Take arch+version specific packages List file.

17 years agoctty command without parameters to disable serial console.
Philip Homburg [Mon, 8 Jan 2007 14:30:11 +0000 (14:30 +0000)]
ctty command without parameters to disable serial console.

17 years agoVFS cleanup (mostly open).
Philip Homburg [Fri, 5 Jan 2007 16:36:55 +0000 (16:36 +0000)]
VFS cleanup (mostly open).

17 years agoDon't let rs232 code send REVIVEs.
Ben Gras [Thu, 4 Jan 2007 12:06:04 +0000 (12:06 +0000)]
Don't let rs232 code send REVIVEs.

17 years agoUse svn instead of cvs
Ben Gras [Fri, 22 Dec 2006 16:43:30 +0000 (16:43 +0000)]
Use svn instead of cvs

17 years agointerface.h is unused (remainder from Ingmar's work)
Ben Gras [Fri, 22 Dec 2006 15:50:17 +0000 (15:50 +0000)]
interface.h is unused (remainder from Ingmar's work)

17 years agoarch/i386/include/protect.h is unnecessary; add make.conf to etc/ and
Ben Gras [Fri, 22 Dec 2006 15:48:27 +0000 (15:48 +0000)]
arch/i386/include/protect.h is unnecessary; add make.conf to etc/ and
note about it in docs/UPDATING.

17 years agoSplit of architecture-dependent and -independent functions for i386,
Ben Gras [Fri, 22 Dec 2006 15:22:27 +0000 (15:22 +0000)]
Split of architecture-dependent and -independent functions for i386,
mainly in the kernel and headers. This split based on work by
Ingmar Alting <iaalting@cs.vu.nl> done for his Minix PowerPC architecture
port.

 . kernel does not program the interrupt controller directly, do any
   other architecture-dependent operations, or contain assembly any more,
   but uses architecture-dependent functions in arch/$(ARCH)/.
 . architecture-dependent constants and types defined in arch/$(ARCH)/include.
 . <ibm/portio.h> moved to <minix/portio.h>, as they have become, for now,
   architecture-independent functions.
 . int86, sdevio, readbios, and iopenable are now i386-specific kernel calls
   and live in arch/i386/do_* now.
 . i386 arch now supports even less 86 code; e.g. mpx86.s and klib86.s have
   gone, and 'machine.protected' is gone (and always taken to be 1 in i386).
   If 86 support is to return, it should be a new architecture.
 . prototypes for the architecture-dependent functions defined in
   kernel/arch/$(ARCH)/*.c but used in kernel/ are in kernel/proto.h
 . /etc/make.conf included in makefiles and shell scripts that need to
   know the building architecture; it defines ARCH=<arch>, currently only
   i386.
 . some basic per-architecture build support outside of the kernel (lib)
 . in clock.c, only dequeue a process if it was ready
 . fixes for new include files

files deleted:
 . mpx/klib.s - only for choosing between mpx/klib86 and -386
 . klib86.s - only for 86

i386-specific files files moved (or arch-dependent stuff moved) to arch/i386/:
 . mpx386.s (entry point)
 . klib386.s
 . sconst.h
 . exception.c
 . protect.c
 . protect.h
 . i8269.c

17 years ago . fsck needs more memory
Ben Gras [Fri, 22 Dec 2006 13:40:37 +0000 (13:40 +0000)]
 . fsck needs more memory
 . fsck must deal with s_max_file_size in superblock being larger
   than LONG_MAX now that off_t is signed

17 years agoMore detailed upgrade procedure.
Philip Homburg [Fri, 22 Dec 2006 13:36:18 +0000 (13:36 +0000)]
More detailed upgrade procedure.

17 years agoRemoved verbose statements from vfs and mfs
Ben Gras [Fri, 22 Dec 2006 11:54:42 +0000 (11:54 +0000)]
Removed verbose statements from vfs and mfs

17 years agoBSD versions of strcasecmp() and strncasecmp() after a bug
Ben Gras [Tue, 19 Dec 2006 13:15:16 +0000 (13:15 +0000)]
BSD versions of strcasecmp() and strncasecmp() after a bug
in strcasecmp() reported by Rogier Meurs <rogier@meurs.org>.

17 years agomake pci driver run as root, so it can call procstat
Ben Gras [Fri, 15 Dec 2006 15:56:36 +0000 (15:56 +0000)]
make pci driver run as root, so it can call procstat

17 years agoSetjmp/longjmp updates and cleanup.
David van Moolenbroek [Fri, 8 Dec 2006 16:23:18 +0000 (16:23 +0000)]
Setjmp/longjmp updates and cleanup.

17 years agoMissing lseek64 library functions.
Philip Homburg [Thu, 7 Dec 2006 20:01:43 +0000 (20:01 +0000)]
Missing lseek64 library functions.

17 years agoSome 64-bit file offset changes that were left out accidentally in the first
Philip Homburg [Wed, 6 Dec 2006 15:21:27 +0000 (15:21 +0000)]
Some 64-bit file offset changes that were left out accidentally in the first
commit.

17 years agoFirst cut at 64-bit file offsets in block devices for mkfs/fsck.
Philip Homburg [Mon, 27 Nov 2006 14:21:43 +0000 (14:21 +0000)]
First cut at 64-bit file offsets in block devices for mkfs/fsck.

17 years agoReplaced <servers/fs/...> with <servers/mfs/...>
Philip Homburg [Fri, 24 Nov 2006 14:04:01 +0000 (14:04 +0000)]
Replaced <servers/fs/...> with <servers/mfs/...>

17 years agoBetter error handling in _mount.c
Philip Homburg [Fri, 24 Nov 2006 14:01:14 +0000 (14:01 +0000)]
Better error handling in _mount.c

17 years agochmod needs more space
Philip Homburg [Fri, 24 Nov 2006 13:59:13 +0000 (13:59 +0000)]
chmod needs more space

17 years agoSmall image doesn't fit on boot floppy.
Philip Homburg [Fri, 24 Nov 2006 13:57:37 +0000 (13:57 +0000)]
Small image doesn't fit on boot floppy.

17 years agoRevert back to previous minised
Ben Gras [Fri, 17 Nov 2006 16:12:08 +0000 (16:12 +0000)]
Revert back to previous minised

17 years agoFix warnings (Ingmar Alting)
Ben Gras [Fri, 10 Nov 2006 18:19:38 +0000 (18:19 +0000)]
Fix warnings (Ingmar Alting)

17 years agoFix some warnings (by Ingmar Alting)
Ben Gras [Fri, 10 Nov 2006 18:13:02 +0000 (18:13 +0000)]
Fix some warnings (by Ingmar Alting)

17 years agoAdded missing defines in include/minix/vfsif.h
Philip Homburg [Fri, 10 Nov 2006 15:05:32 +0000 (15:05 +0000)]
Added missing defines in include/minix/vfsif.h

17 years agoGetdents implementation in library/vfs/mfs.
Philip Homburg [Thu, 9 Nov 2006 16:22:54 +0000 (16:22 +0000)]
Getdents implementation in library/vfs/mfs.
Changed readdir, etc. to use getdents

17 years agoTypo reported by ville.solarius@gmail.com
Ben Gras [Mon, 6 Nov 2006 18:02:22 +0000 (18:02 +0000)]
Typo reported by ville.solarius@gmail.com

17 years agoset $PATH so /usr/local/egrep isn't used; also exclude /var/log from package
Ben Gras [Fri, 3 Nov 2006 14:47:28 +0000 (14:47 +0000)]
set $PATH so /usr/local/egrep isn't used; also exclude /var/log from package

17 years agoNew sed manpage
Ben Gras [Thu, 2 Nov 2006 16:51:04 +0000 (16:51 +0000)]
New sed manpage

17 years agoold sed out, new sed in
Ben Gras [Thu, 2 Nov 2006 16:48:47 +0000 (16:48 +0000)]
old sed out, new sed in

17 years agoNew sed
Ben Gras [Thu, 2 Nov 2006 16:48:15 +0000 (16:48 +0000)]
New sed

17 years agoobsolete server.
Ben Gras [Wed, 1 Nov 2006 16:53:45 +0000 (16:53 +0000)]
obsolete server.

17 years agoat_wini now calls pci_reserve.
Philip Homburg [Wed, 1 Nov 2006 14:55:00 +0000 (14:55 +0000)]
at_wini now calls pci_reserve.

17 years agoExplicit conversion from O_ACCMODE to minix [RW]_BIT is clearer
Ben Gras [Wed, 1 Nov 2006 14:17:47 +0000 (14:17 +0000)]
Explicit conversion from O_ACCMODE to minix [RW]_BIT is clearer
and catches bogus values (such as 3).

17 years agors changes (also use driver configurations in the image ramdisk)
Philip Homburg [Tue, 31 Oct 2006 13:35:04 +0000 (13:35 +0000)]
rs changes (also use driver configurations in the image ramdisk)

17 years agoIgnore .svn dir as well as CVS dir.
Ben Gras [Mon, 30 Oct 2006 16:11:57 +0000 (16:11 +0000)]
Ignore .svn dir as well as CVS dir.

17 years agoSystem statistical and call profiling
Ben Gras [Mon, 30 Oct 2006 15:53:38 +0000 (15:53 +0000)]
System statistical and call profiling
support by Rogier Meurs <rogier@meurs.org>.

17 years agoMerge of VFS by Balasz Gerofi with Minix trunk.
Ben Gras [Wed, 25 Oct 2006 13:40:36 +0000 (13:40 +0000)]
Merge of VFS by Balasz Gerofi with Minix trunk.

17 years agoMore memory for these drivers
Ben Gras [Wed, 25 Oct 2006 13:39:53 +0000 (13:39 +0000)]
More memory for these drivers

17 years agoPart of wait()/waitpid() fix in revision 2629:2634 in PM - kernel
Ben Gras [Wed, 25 Oct 2006 13:38:31 +0000 (13:38 +0000)]
Part of wait()/waitpid() fix in revision 2629:2634 in PM - kernel
unblocks a process that has no priority (stopped with sys_nice()
and PRIO_STOP) and reschedules it if that gives it RTS flags of 0.

17 years agoDefault amount of memory for tests (some ran out of memory)
Ben Gras [Wed, 25 Oct 2006 13:35:03 +0000 (13:35 +0000)]
Default amount of memory for tests (some ran out of memory)

17 years agoInclude kernel/priv.h for rs/manager.h
Philip Homburg [Wed, 25 Oct 2006 13:07:58 +0000 (13:07 +0000)]
Include kernel/priv.h for rs/manager.h

17 years ago. processes stay ZOMBIE, even after wait(), to avoid wrongly seeing them
Ben Gras [Wed, 25 Oct 2006 11:29:43 +0000 (11:29 +0000)]
. processes stay ZOMBIE, even after wait(), to avoid wrongly seeing them
  as living processes  before they are cleaned up (fixes
  wait()/waitpid() hanging forever on previously-ZOMBIE processes)

. stop processes from running using sys_nice() with PRIO_STOP
  when a handled signal is delivered, before computing
  stack locations for sys_sigsend(). (fixes race condition
  when runnable processes get signals, and e.g. get scheduled
  before FS sends a reply to unpause(), which can make the
  signal stack location wrong.)

17 years agoDon't do make clean if flex Makefile doesn't exist.
Ben Gras [Tue, 24 Oct 2006 14:20:59 +0000 (14:20 +0000)]
Don't do make clean if flex Makefile doesn't exist.

17 years ago(Incomplete) support for access control in PCI (pci_set_acl).
Philip Homburg [Fri, 20 Oct 2006 15:01:32 +0000 (15:01 +0000)]
(Incomplete) support for access control in PCI (pci_set_acl).
-script argument to service for crash recovery scripts
-config argument to service for driver resource configuration
restart command in service to restart a driver after a crash (for use in
crash recovery scripts).
down and refresh now take labels instead of pids.
verious changes in rs to make this work.

17 years agoAccess control in do_sdevio and do_vdevio
Philip Homburg [Fri, 20 Oct 2006 14:46:55 +0000 (14:46 +0000)]
Access control in do_sdevio and do_vdevio

17 years agoInitialize priv from user supplied priv structure in SYS_PRIV_INIT.
Philip Homburg [Fri, 20 Oct 2006 14:42:48 +0000 (14:42 +0000)]
Initialize priv from user supplied priv structure in SYS_PRIV_INIT.
Added SYS_PRIV_USER call to downgrade a privileged process to a user process.

17 years agoAdditional ptrace commands for fault injection.
Philip Homburg [Fri, 20 Oct 2006 14:14:08 +0000 (14:14 +0000)]
Additional ptrace commands for fault injection.

17 years agoNice(3) implementation
Philip Homburg [Fri, 20 Oct 2006 14:10:53 +0000 (14:10 +0000)]
Nice(3) implementation

17 years agoBigger ramdisk
Philip Homburg [Fri, 20 Oct 2006 13:59:42 +0000 (13:59 +0000)]
Bigger ramdisk

17 years agoUse /usr/tmp as temp dir for compiling imgrd_s.s
Philip Homburg [Fri, 20 Oct 2006 13:58:45 +0000 (13:58 +0000)]
Use /usr/tmp as temp dir for compiling imgrd_s.s

17 years agoMore stack for the log driver.
Philip Homburg [Fri, 20 Oct 2006 13:54:14 +0000 (13:54 +0000)]
More stack for the log driver.

17 years agoUse size-dependant no. of words, with OPEN_MAX as default for fd_set size.
Ben Gras [Fri, 6 Oct 2006 15:45:13 +0000 (15:45 +0000)]
Use size-dependant no. of words, with OPEN_MAX as default for fd_set size.

17 years agoFix for clean target.
Ben Gras [Fri, 6 Oct 2006 15:44:41 +0000 (15:44 +0000)]
Fix for clean target.

17 years agoRemove objects in aes subdir too
Ben Gras [Thu, 5 Oct 2006 09:58:50 +0000 (09:58 +0000)]
Remove objects in aes subdir too

17 years agomake clean in ibm subdir too
Ben Gras [Thu, 5 Oct 2006 09:56:17 +0000 (09:56 +0000)]
make clean in ibm subdir too

17 years agoLeftover junk
Ben Gras [Thu, 5 Oct 2006 09:55:29 +0000 (09:55 +0000)]
Leftover junk

17 years agoDon't include this in the distribution.
Ben Gras [Thu, 5 Oct 2006 09:53:42 +0000 (09:53 +0000)]
Don't include this in the distribution.

17 years agoChange select() so that only as many words as necessary for the 'nfds'
Ben Gras [Wed, 4 Oct 2006 13:28:38 +0000 (13:28 +0000)]
Change select() so that only as many words as necessary for the 'nfds'
argument from the fd bitmasks are copied from and back to userspace. This
solves an ABI dependency on OPEN_MAX. If nfds is too big for the current
OPEN_MAX, select() fails (but that's relatively easy to fix by 'just'
recompiling the system and not the application binaries), but if it's
smaller, binaries keep working.

17 years agoDon't install progressbar suid.
Ben Gras [Wed, 27 Sep 2006 15:40:10 +0000 (15:40 +0000)]
Don't install progressbar suid.

17 years agoReorder for port order, and add svn (Subversion) as service (port 3690)
Ben Gras [Wed, 27 Sep 2006 13:55:54 +0000 (13:55 +0000)]
Reorder for port order, and add svn (Subversion) as service (port 3690)

17 years ago_NSIG is supposed to be highest signal number plus one, not highest
Ben Gras [Wed, 27 Sep 2006 08:56:37 +0000 (08:56 +0000)]
_NSIG is supposed to be highest signal number plus one, not highest
signal number.

18 years agoMissing negation in comment
Philip Homburg [Thu, 21 Sep 2006 13:33:23 +0000 (13:33 +0000)]
Missing negation in comment

18 years agoAdded EOPNOTSUPP and better error handling in accept.
Philip Homburg [Thu, 14 Sep 2006 13:48:41 +0000 (13:48 +0000)]
Added EOPNOTSUPP and better error handling in accept.

18 years agoDisabled /etc/rc.rescue
Philip Homburg [Fri, 8 Sep 2006 13:20:57 +0000 (13:20 +0000)]
Disabled /etc/rc.rescue

18 years agoNUL terminate timingdata[cat].names instead of timingdata[0].names. Reported
Philip Homburg [Mon, 28 Aug 2006 15:30:14 +0000 (15:30 +0000)]
NUL terminate timingdata[cat].names instead of timingdata[0].names. Reported
by <devel@pop3.ru>.

18 years agoAlso clear word 6 in the ATAPI SCSI_READ10 command packet (in atapi_transfer).
Philip Homburg [Mon, 28 Aug 2006 15:10:10 +0000 (15:10 +0000)]
Also clear word 6 in the ATAPI SCSI_READ10 command packet (in atapi_transfer).

18 years agofirst commit of extra ptrace code for fault injection.
Philip Homburg [Mon, 28 Aug 2006 15:03:03 +0000 (15:03 +0000)]
first commit of extra ptrace code for fault injection.