]> Zhao Yanbai Git Server - minix.git/log
minix.git
15 years agofp might be NULL.
Ben Gras [Mon, 21 Sep 2009 15:23:08 +0000 (15:23 +0000)]
fp might be NULL.

15 years agoipc tests by GQ
Ben Gras [Mon, 21 Sep 2009 14:53:53 +0000 (14:53 +0000)]
ipc tests by GQ

15 years agonew ipc server that implements shared memory and semaphores.
Ben Gras [Mon, 21 Sep 2009 14:53:13 +0000 (14:53 +0000)]
new ipc server that implements shared memory and semaphores.

this server, tests, vm support, library stubs and other contributions
are the work of Guanqun Lu, a 2009 GSOC student.

15 years ago - pages that points to page directory values of all processes,
Ben Gras [Mon, 21 Sep 2009 14:49:49 +0000 (14:49 +0000)]
 - pages that points to page directory values of all processes,
   shared with the kernel, mapped into kernel address space;
   kernel is notified of its location. kernel segment size is
   increased to make it fit.
 - map in kernel and other processes that don't have their
   own page table using single 4MB (global) mapping.
 - new sanity check facility: objects that are allocated with
   the slab allocator are, when running with sanity checking on,
   marked readonly until they are explicitly unlocked using the USE()
   macro.
 - another sanity check facility: collect all uses of memory and
   see if they don't overlap with (a) eachother and (b) free memory
 - own munmap() and munmap_text() functions.
 - exec() recovers from out-of-memory conditions properly now; this
   solves some weird exec() behaviour
 - chew off memory from the same side of the chunk as where we
   start scanning, solving some memory fragmentation issues
 - use avl trees for freelist and phys_ranges in regions
 - implement most useful part of munmap()
 - remap() stuff is GQ's for shared memory

15 years ago - some exec debugging prints when errors happen
Ben Gras [Mon, 21 Sep 2009 14:49:26 +0000 (14:49 +0000)]
 - some exec debugging prints when errors happen
 - lookup mounted_on check to avoid NULL dereference
 - some errors in exec() shouldn't be fatal

15 years agosupport for vm priv system.
Ben Gras [Mon, 21 Sep 2009 14:49:04 +0000 (14:49 +0000)]
support for vm priv system.

fix NULL envp ptr.

15 years ago - Introduce some macros for field names, so that endpt, pendpt,
Ben Gras [Mon, 21 Sep 2009 14:48:19 +0000 (14:48 +0000)]
  - Introduce some macros for field names, so that endpt, pendpt,
    addr and taddr don't have to be defined any more, so that <sys/mman.h>
    can be included for proper prototypes of munmap() and friends.
  - rename our GETPID to MINIX_GETPID to avoid a name conflict with
    other sources
  - PM needs its own munmap() and munmap_text() to avoid sending messages
    to VM at the startup phase. It *does* want to do that, but only
    after initialising. So they're called again with unmap_ok set to 1
    later.
  - getnuid(), getngid() implementation

15 years ago - No maximum block size any more.
Ben Gras [Mon, 21 Sep 2009 14:47:51 +0000 (14:47 +0000)]
 - No maximum block size any more.
 - If allocation of a new buffer fails, use an already-allocated
   unused buffer if available (low memory conditions)
 - Allocate buffers dynamically, so memory isn't wasted on wrong-sized
   buffers.
 - No more _MAX_BLOCK_SIZE.

15 years agono more scheduling queue dumps (kernel will print this over serial);
Ben Gras [Mon, 21 Sep 2009 14:47:24 +0000 (14:47 +0000)]
no more scheduling queue dumps (kernel will print this over serial);
initial vm stats retrieve support

15 years agoclever hack to avoid starting from 0 disabled because
Ben Gras [Mon, 21 Sep 2009 14:46:42 +0000 (14:46 +0000)]
clever hack to avoid starting from 0 disabled because
when own 0-pages are mapped out, this doesn't work and
isn't necessary.

15 years agovarious syscall entry points
Ben Gras [Mon, 21 Sep 2009 14:46:10 +0000 (14:46 +0000)]
various syscall entry points

15 years agorename GETPID to MINIX_GETPID (for now); macro for field name
Ben Gras [Mon, 21 Sep 2009 14:45:46 +0000 (14:45 +0000)]
rename GETPID to MINIX_GETPID (for now); macro for field name

15 years agomore vm stubs
Ben Gras [Mon, 21 Sep 2009 14:45:09 +0000 (14:45 +0000)]
more vm stubs

15 years agoftime()
Ben Gras [Mon, 21 Sep 2009 14:44:35 +0000 (14:44 +0000)]
ftime()

15 years agomore pid queries; PM field macros; vm stubs
Ben Gras [Mon, 21 Sep 2009 14:44:00 +0000 (14:44 +0000)]
more pid queries; PM field macros; vm stubs

15 years agovm stubs; no more sys_vm_setbuf
Ben Gras [Mon, 21 Sep 2009 14:42:58 +0000 (14:42 +0000)]
vm stubs; no more sys_vm_setbuf

15 years agosysvipc dir
Ben Gras [Mon, 21 Sep 2009 14:42:13 +0000 (14:42 +0000)]
sysvipc dir

15 years agochange startup code to call munmap()/munmap_text() to map out
Ben Gras [Mon, 21 Sep 2009 14:41:47 +0000 (14:41 +0000)]
change startup code to call munmap()/munmap_text() to map out
own 0-pages.

15 years agonew sysv ipc call stubs
Ben Gras [Mon, 21 Sep 2009 14:41:03 +0000 (14:41 +0000)]
new sysv ipc call stubs

15 years ago - remove unused kmalloc
Ben Gras [Mon, 21 Sep 2009 14:39:15 +0000 (14:39 +0000)]
 - remove unused kmalloc
 - initial implementation of _read_host_time_ns, a function to retrieve
   real-time in ns in vmware virtual machines

15 years agoPrimary goal for these changes is:
Ben Gras [Mon, 21 Sep 2009 14:31:52 +0000 (14:31 +0000)]
Primary goal for these changes is:
  - no longer have kernel have its own page table that is loaded
    on every kernel entry (trap, interrupt, exception). the primary
    purpose is to reduce the number of required reloads.
Result:
  - kernel can only access memory of process that was running when
    kernel was entered
  - kernel must be mapped into every process page table, so traps to
    kernel keep working
Problem:
  - kernel must often access memory of arbitrary processes (e.g. send
    arbitrary processes messages); this can't happen directly any more;
    usually because that process' page table isn't loaded at all, sometimes
    because that memory isn't mapped in at all, sometimes because it isn't
    mapped in read-write.
So:
  - kernel must be able to map in memory of any process, in its own
    address space.
Implementation:
  - VM and kernel share a range of memory in which addresses of
    all page tables of all processes are available. This has two purposes:
      . Kernel has to know what data to copy in order to map in a range
      . Kernel has to know where to write the data in order to map it in
    That last point is because kernel has to write in the currently loaded
    page table.
  - Processes and kernel are separated through segments; kernel segments
    haven't changed.
  - The kernel keeps the process whose page table is currently loaded
    in 'ptproc.'
  - If it wants to map in a range of memory, it writes the value of the
    page directory entry for that range into the page directory entry
    in the currently loaded map. There is a slot reserved for such
    purposes. The kernel can then access this memory directly.
  - In order to do this, its segment has been increased (and the
    segments of processes start where it ends).
  - In the pagefault handler, detect if the kernel is doing
    'trappable' memory access (i.e. a pagefault isn't a fatal
     error) and if so,
       - set the saved instruction pointer to phys_copy_fault,
 breaking out of phys_copy
       - set the saved eax register to the address of the page
 fault, both for sanity checking and for checking in
 which of the two ranges that phys_copy was called
 with the fault occured
  - Some boot-time processes do not have their own page table,
    and are mapped in with the kernel, and separated with
    segments. The kernel detects this using HASPT. If such a
    process has to be scheduled, any page table will work and
    no page table switch is done.

Major changes in kernel are
  - When accessing user processes memory, kernel no longer
    explicitly checks before it does so if that memory is OK.
    It simply makes the mapping (if necessary), tries to do the
    operation, and traps the pagefault if that memory isn't present;
    if that happens, the copy function returns EFAULT.
    So all of the CHECKRANGE_OR_SUSPEND macros are gone.
  - Kernel no longer has to copy/read and parse page tables.
  - A message copying optimisation: when messages are copied, and
    the recipient isn't mapped in, they are copied into a buffer
    in the kernel. This is done in QueueMess. The next time
    the recipient is scheduled, this message is copied into
    its memory. This happens in schedcheck().
    This eliminates the mapping/copying step for messages, and makes
    it easier to deliver messages. This eliminates soft_notify.
  - Kernel no longer creates a page table at all, so the vm_setbuf
    and pagetable writing in memory.c is gone.

Minor changes in kernel are
  - ipc_stats thrown out, wasn't used
  - misc flags all renamed to MF_*
  - NOREC_* macros to enter and leave functions that should not
    be called recursively; just sanity checks really
  - code to fully decode segment selectors and descriptors
    to print on exceptions
  - lots of vmassert()s added, only executed if DEBUG_VMASSERT is 1

15 years agominor change to panic code
Ben Gras [Mon, 21 Sep 2009 14:28:16 +0000 (14:28 +0000)]
minor change to panic code

15 years ago - new entry for bios_wini and ipc
Ben Gras [Mon, 21 Sep 2009 14:26:20 +0000 (14:26 +0000)]
 - new entry for bios_wini and ipc
 - start ipc server on botting

15 years ago - tty: only report unrecognized scancodes once; forget about
Ben Gras [Mon, 21 Sep 2009 14:25:54 +0000 (14:25 +0000)]
 - tty: only report unrecognized scancodes once; forget about
   remembering the origin and cursor position as that feature didn't
   really work properly anyway
 - tty: map in video and font memory using a vm call, access it from C,
   thereby eliminating pesky weird segment calls and assembly to access it,
   and unbreaks loadfont (Roman Ignatov)
 - bios_wini: fix bios_wini by allocating a <1MB buffers for it
 - memory: preallocate ramdisk, makes it a bit faster (and doesn't
   fail halfway if you allocate a huge one)
 - floppy: use <1MB buffer
 - ramdisk proto: because of the 2x1 page reservations, binaries
   got a little fatter and didn't fit on the ramdisk any more.
   increase it.

15 years ago - added 'datasizes' script that shows you the size allocated
Ben Gras [Mon, 21 Sep 2009 14:24:29 +0000 (14:24 +0000)]
 - added 'datasizes' script that shows you the size allocated
   for each symbol, usually answering those "why is does my binary have
   such a lot of BSS" questions.
 - stop binpackage looking in /var/spool for package files.
 - let makewhatis recognize .Sh as heading name
 - setup, fsck, df: allow >4kB block sizes painlessly
 - mkfs: new #-of-inodes heuristic that depends on kb, not
   on blocks; i've run out of inodes on my /usr
 - asmconv: don't silently truncate .aligns to 16 bytes
 - ipc* commands for shared memory support

15 years ago - remove unused bootdelay feauture
Ben Gras [Mon, 21 Sep 2009 14:23:47 +0000 (14:23 +0000)]
 - remove unused bootdelay feauture
 - only print a line for every boot process if 'verbose' variable set to
   nonzero; reason: with serial output, the long output
   significantly slows down frequent reboots, and causes 'scroll damage'
   that in some cases is pretty bad. also the verbose output doesn't tell
   you the one thing you might want to know about a process: how much memory
   is it using? or how much memory is everything using?
 - short format does print out total memory allocated for processes

15 years ago- VM_KERN_NOPAGEZERO feature is gone
Ben Gras [Mon, 21 Sep 2009 14:23:10 +0000 (14:23 +0000)]
- VM_KERN_NOPAGEZERO feature is gone
- sys_getbiosbuffer feature is gone (from kernel; available from vm)
- bump version number because munmap() calls that newly compiled binaries
  will do trigger an ugly (but harmless) error message in older VM's
- some new VM calls and flags, the new IPC calls
- some new CR0 register bits
- added files for shared memory

15 years agotest31 fix
David van Moolenbroek [Sun, 20 Sep 2009 18:15:24 +0000 (18:15 +0000)]
test31 fix

15 years agotest32 fix
David van Moolenbroek [Sun, 20 Sep 2009 16:47:57 +0000 (16:47 +0000)]
test32 fix

15 years agoKernel: correct a few SYSTEM source documentation comments
David van Moolenbroek [Thu, 17 Sep 2009 20:51:34 +0000 (20:51 +0000)]
Kernel: correct a few SYSTEM source documentation comments

15 years agoSome clean up of the segment selectors macros
Tomas Hruby [Tue, 15 Sep 2009 10:01:06 +0000 (10:01 +0000)]
Some clean up of the segment selectors macros

- [ABCD]_INDEX are not used anywhere

- value of *_SELECTOR is now calculated using the *_INDEX value so changing the
  index does not break the selector

- TSS is now the last of the global selectors. There will be TSS per CPU on SMP
  and the number will vary depending on the maximal supported number of CPUs
  configured

15 years agoiskernelp() and isuserp() test pointers
Tomas Hruby [Tue, 15 Sep 2009 09:58:46 +0000 (09:58 +0000)]
iskernelp() and isuserp() test pointers

- we may test even not fully initialized entries, e.g. during boot
  crash

- is we know the process number we should use iskerneln
  and isusern

15 years agopproc_addr unused and removed
Tomas Hruby [Tue, 15 Sep 2009 09:57:54 +0000 (09:57 +0000)]
pproc_addr unused and removed

15 years agoproc_addr() returns address based on location in proc array
Tomas Hruby [Tue, 15 Sep 2009 09:57:22 +0000 (09:57 +0000)]
proc_addr() returns address based on location in proc array

- pproc_addr is not neccessary to get the address of a process if we know its
  number

- local proc variables in system calls implementation (sys_task) conflicts with
  the global proc array of all process, therefore the variable were renamed to
  proc_nr as they hold the process number

15 years agoHandle SIGSTOP more consistently and according to POSIX rules; it can no longer be...
Erik van der Kouwe [Sun, 13 Sep 2009 12:24:23 +0000 (12:24 +0000)]
Handle SIGSTOP more consistently and according to POSIX rules; it can no longer be ignored

15 years agoPM: some tiny fixes
David van Moolenbroek [Sat, 12 Sep 2009 18:36:07 +0000 (18:36 +0000)]
PM: some tiny fixes

15 years agoremove unused NEW_KSIG definition
David van Moolenbroek [Mon, 7 Sep 2009 22:12:50 +0000 (22:12 +0000)]
remove unused NEW_KSIG definition

15 years agoKernel: remove unused CHECK_DEADLOCK definition
David van Moolenbroek [Mon, 7 Sep 2009 20:23:31 +0000 (20:23 +0000)]
Kernel: remove unused CHECK_DEADLOCK definition

15 years agoit's/its
David van Moolenbroek [Sun, 6 Sep 2009 17:13:08 +0000 (17:13 +0000)]
it's/its

15 years agoKernel: do_[sv]devio.c header corrections
David van Moolenbroek [Sun, 6 Sep 2009 15:54:15 +0000 (15:54 +0000)]
Kernel: do_[sv]devio.c header corrections

15 years agocleanup: sys_irqrmpolicy does not need IRQ vector
David van Moolenbroek [Sun, 6 Sep 2009 14:24:44 +0000 (14:24 +0000)]
cleanup: sys_irqrmpolicy does not need IRQ vector

15 years agoremove unused IRQ_ENDPT definition
David van Moolenbroek [Sun, 6 Sep 2009 14:14:03 +0000 (14:14 +0000)]
remove unused IRQ_ENDPT definition

15 years ago- sys_privctl: don't mix message types
David van Moolenbroek [Sun, 6 Sep 2009 12:37:13 +0000 (12:37 +0000)]
- sys_privctl: don't mix message types
- sys_privctl: remove CTL_MM_PRIV (third parameter)
- remove obsolete sys_svrctl.c library file

15 years agosyslib: make sys_nice use field aliases
David van Moolenbroek [Sat, 5 Sep 2009 12:51:00 +0000 (12:51 +0000)]
syslib: make sys_nice use field aliases

15 years agoDisabled check in test 28 which hard links a directory; this is nott required by...
Erik van der Kouwe [Fri, 4 Sep 2009 20:23:25 +0000 (20:23 +0000)]
Disabled check in test 28 which hard links a directory; this is nott required by POSIX and not supported (currently) by MINIX. Also corrected total number of tests.

15 years agoNow all tests except for 28 succeed in root mode
Erik van der Kouwe [Fri, 4 Sep 2009 14:14:49 +0000 (14:14 +0000)]
Now all tests except for 28 succeed in root mode

15 years agoRemove redundant line from previous patch
Erik van der Kouwe [Fri, 4 Sep 2009 13:36:43 +0000 (13:36 +0000)]
Remove redundant line from previous patch

15 years agoAllow setuid tests 11 and 33 to run. The former still fails (but now with a meaningfu...
Erik van der Kouwe [Fri, 4 Sep 2009 13:18:49 +0000 (13:18 +0000)]
Allow setuid tests 11 and 33 to run. The former still fails (but now with a meaningful error) while the latter succeeds. Only 2 tests are left borken on default MINIX, namely 11 and 28.

15 years agoAvoid alarm to prevent the setitimer interval from being reset
Erik van der Kouwe [Thu, 3 Sep 2009 20:35:22 +0000 (20:35 +0000)]
Avoid alarm to prevent the setitimer interval from being reset

15 years ago - Added netconf script which makes it a lot easier to change network settings.
Thomas Veerman [Thu, 3 Sep 2009 09:38:27 +0000 (09:38 +0000)]
 - Added netconf script which makes it a lot easier to change network settings.
 - Modified the setup script to use the netconf script for the network
   configuration:
    - Moved step 2 to step 8 and renamed the steps in between.
    - Autopart adapted to print step 3 instead of step 4.

15 years agotypo
David van Moolenbroek [Wed, 2 Sep 2009 21:55:26 +0000 (21:55 +0000)]
typo

15 years agopostgresql services entry.
Ben Gras [Wed, 2 Sep 2009 13:23:22 +0000 (13:23 +0000)]
postgresql services entry.

15 years agoFXP: support for 8086/1064 (reported and tested by Jose Luis Sanchez)
David van Moolenbroek [Sun, 30 Aug 2009 16:13:46 +0000 (16:13 +0000)]
FXP: support for 8086/1064 (reported and tested by Jose Luis Sanchez)

15 years agoFix time(1) for custom system HZ values (reported by Jose Luis Sanchez
David van Moolenbroek [Sun, 30 Aug 2009 16:06:59 +0000 (16:06 +0000)]
Fix time(1) for custom system HZ values (reported by Jose Luis Sanchez

15 years agosystem_init() renamed to arch_init()
Tomas Hruby [Sun, 30 Aug 2009 14:55:30 +0000 (14:55 +0000)]
system_init() renamed to arch_init()

- a better name for architecture specific init function

- some of x86 init code must execute in protected mode

- prot_init() removed from this function and still called in cstart() Imho this
  should be called from the architecture specific assembly not cstart. cstart
  perform Minix monitor specific tasks and will be touched once another
  bootloader is in use, e.g. booting via tftp, therefore we keep it as is for
  now.

- this is a backport from the SMP code which requires this. Merging will be simpler

15 years agoPrinting the trap info even if kernel crashes is handy.
Tomas Hruby [Sat, 29 Aug 2009 19:38:11 +0000 (19:38 +0000)]
Printing the trap info even if kernel crashes is handy.

15 years agosaved_proc in exception() may be NULL
Tomas Hruby [Sat, 29 Aug 2009 19:26:46 +0000 (19:26 +0000)]
saved_proc in exception() may be NULL

If an exception happens in kernel while the kernel is booting and no processes
are running yet, saved_proc == NULL and priting any process related information
results in dumping rubish.

This check is mostly useful when debugging kernel stuff. Should _never_ happen
on a production kernel.

15 years agoVFS: fetch_name() buffer underflow (reported by John Peace, bug #305)
David van Moolenbroek [Sat, 29 Aug 2009 08:22:50 +0000 (08:22 +0000)]
VFS: fetch_name() buffer underflow (reported by John Peace, bug #305)

15 years agoIDT is initialized in idt_init() not in prot_init()
Tomas Hruby [Fri, 28 Aug 2009 15:55:30 +0000 (15:55 +0000)]
IDT is initialized in idt_init() not in prot_init()

This is a backport form the SMP branch. Not required here, it only makes life
for SMP easier. And future merging too.

- filling the IDT is removed from prot_init()

- struct gate_table_s is a public type

- gate_table_pic is a global array as it is used by APIC code too

- idt_copy_vectors() is also global and used by idt_init() as well as
  apic_idt_init()

- idt_init() is called right after prot_init() in system_init()

15 years agoRe-disable complete_bars, it was not needed for PRO/100 despite the error message
Erik van der Kouwe [Thu, 27 Aug 2009 07:10:03 +0000 (07:10 +0000)]
Re-disable complete_bars, it was not needed for PRO/100 despite the error message

15 years agoExtend Intel PRO/100 support
Erik van der Kouwe [Wed, 26 Aug 2009 21:01:46 +0000 (21:01 +0000)]
Extend Intel PRO/100 support

15 years agourlget: broken ftpget call, reported by Charles Littlefield
David van Moolenbroek [Sun, 23 Aug 2009 13:53:03 +0000 (13:53 +0000)]
urlget: broken ftpget call, reported by Charles Littlefield

15 years agosupport for TCP sockets in send/sendto/recv/recvfrom
David van Moolenbroek [Fri, 21 Aug 2009 09:59:09 +0000 (09:59 +0000)]
support for TCP sockets in send/sendto/recv/recvfrom

15 years agoUse su to allow tests 3, 6, 17, 18, 19, 21 and 25 to work as root, making testing...
Erik van der Kouwe [Tue, 18 Aug 2009 19:58:34 +0000 (19:58 +0000)]
Use su to allow tests 3, 6, 17, 18, 19, 21 and 25 to work as root, making testing easier

15 years agoFix HUGE_VAL warnings
Erik van der Kouwe [Tue, 18 Aug 2009 19:10:48 +0000 (19:10 +0000)]
Fix HUGE_VAL warnings

15 years agoFix HUGE_VAL warnings
Erik van der Kouwe [Tue, 18 Aug 2009 19:10:20 +0000 (19:10 +0000)]
Fix HUGE_VAL warnings

15 years agoVFS quits gracefully if mount fails and mounted_on remains uninitialized
Tomas Hruby [Tue, 18 Aug 2009 13:30:05 +0000 (13:30 +0000)]
VFS quits gracefully if mount fails and mounted_on remains uninitialized

15 years agoAdded a hack to start binaries from the boot image only. In particular, setting
Thomas Veerman [Tue, 18 Aug 2009 11:36:01 +0000 (11:36 +0000)]
Added a hack to start binaries from the boot image only. In particular, setting
bin_img=1 in the boot monitor will make sure that during the boot procedure the
mfs binary that is part of the boot image is the only binary that is used to
mount partitions. This is useful when for some reason the mfs binary on disk
malfunctions, rendering Minix unable to boot. By setting bin_img=1, the binary
on disk is ignored and the binary in the boot image is used instead.

- 'service' now accepts an additional flag -r. -r implies -c. -r instructs RS
  to first look in memory if the binary has already been copied to memory and
  execute that version, instead of loading the binary from disk. For example,
  the first time a MFS is being started it is copied (-c) to memory and
  executed from there. The second time MFS is being started this way, RS will
  look in memory for a previously copied MFS binary and reuse it if it exists.
- The mount and newroot commands now accept an additional flag -i, which
  instructs them to set the MS_REUSE flag in the mount flags.
- The mount system call now supports the MS_REUSE flag and invokes 'service'
  with the -r flag when MS_REUSE is set.
- /etc/rc and the rc script that's included in the boot image check for the
  existence of the bin_img flag in the boot monitor, and invoke mount and
  newroot with the -i flag accordingly.

15 years agosmall errno.h corrections
David van Moolenbroek [Tue, 18 Aug 2009 09:26:51 +0000 (09:26 +0000)]
small errno.h corrections

15 years agoFixed some warnings
Erik van der Kouwe [Mon, 17 Aug 2009 18:49:15 +0000 (18:49 +0000)]
Fixed some warnings

15 years agoFixed some warnings
Erik van der Kouwe [Mon, 17 Aug 2009 18:49:07 +0000 (18:49 +0000)]
Fixed some warnings

15 years agoFixed some warnings
Erik van der Kouwe [Mon, 17 Aug 2009 18:48:57 +0000 (18:48 +0000)]
Fixed some warnings

15 years agoFixed some warnings
Erik van der Kouwe [Mon, 17 Aug 2009 18:48:27 +0000 (18:48 +0000)]
Fixed some warnings

15 years agoFix warning in panic
Erik van der Kouwe [Mon, 17 Aug 2009 17:56:34 +0000 (17:56 +0000)]
Fix warning in panic

15 years agoMissing defines preventing recursive includesion in some include/sys/*.h files
Tomas Hruby [Mon, 17 Aug 2009 14:34:14 +0000 (14:34 +0000)]
Missing defines preventing recursive includesion in some include/sys/*.h files

15 years agoRemoved obsolete declarations in include/ipc.h
Erik van der Kouwe [Sun, 16 Aug 2009 15:12:07 +0000 (15:12 +0000)]
Removed obsolete declarations in include/ipc.h

15 years agoNanosleep was implemented twice, removed from _sleep.c
Erik van der Kouwe [Sun, 16 Aug 2009 12:55:59 +0000 (12:55 +0000)]
Nanosleep was implemented twice, removed from _sleep.c

15 years agoFunction nanosleep now checks arguments
Erik van der Kouwe [Sun, 16 Aug 2009 12:13:33 +0000 (12:13 +0000)]
Function nanosleep now checks arguments

15 years agoMan page for new function nanosleep(3), updated man page for sleep(3)
Erik van der Kouwe [Sun, 16 Aug 2009 12:12:40 +0000 (12:12 +0000)]
Man page for new function nanosleep(3), updated man page for sleep(3)

15 years agoman page fix: sigaction.2 signals list
David van Moolenbroek [Sat, 15 Aug 2009 23:27:32 +0000 (23:27 +0000)]
man page fix: sigaction.2 signals list

15 years agoPM: make SIGBUS, but not SIGUSR1 or SIGUSR2, cause a core dump
David van Moolenbroek [Sat, 15 Aug 2009 23:23:24 +0000 (23:23 +0000)]
PM: make SIGBUS, but not SIGUSR1 or SIGUSR2, cause a core dump

15 years agorun test 41, too (oops)
David van Moolenbroek [Sat, 15 Aug 2009 22:31:38 +0000 (22:31 +0000)]
run test 41, too (oops)

15 years agonanosleep(3), and sleep(3) rewrite, by Erik van der Kouwe
David van Moolenbroek [Sat, 15 Aug 2009 22:14:48 +0000 (22:14 +0000)]
nanosleep(3), and sleep(3) rewrite, by Erik van der Kouwe

15 years agoSupport for setitimer(ITIMER_VIRTUAL/ITIMER_PROF). New test (41) for setitimer.
David van Moolenbroek [Sat, 15 Aug 2009 21:37:26 +0000 (21:37 +0000)]
Support for setitimer(ITIMER_VIRTUAL/ITIMER_PROF). New test (41) for setitimer.

15 years agoSupport for setitimer(ITIMER_REAL).
David van Moolenbroek [Sat, 15 Aug 2009 16:09:32 +0000 (16:09 +0000)]
Support for setitimer(ITIMER_REAL).

15 years agofix test14's "clever" (broken) filename generation
David van Moolenbroek [Sat, 15 Aug 2009 15:13:57 +0000 (15:13 +0000)]
fix test14's "clever" (broken) filename generation

15 years agoawk: support for whitespace between array name and bracket
David van Moolenbroek [Sat, 15 Aug 2009 12:05:41 +0000 (12:05 +0000)]
awk: support for whitespace between array name and bracket

15 years agoawk: support for ENVIRON
David van Moolenbroek [Sat, 15 Aug 2009 11:34:41 +0000 (11:34 +0000)]
awk: support for ENVIRON

15 years agou64_t for human beings - the high and low parts are named hi and lo, lib/sysutil...
Tomas Hruby [Thu, 13 Aug 2009 15:47:49 +0000 (15:47 +0000)]
u64_t for human beings - the high and low parts are named hi and lo, lib/sysutil/profile.c updated accordingly

15 years agomove _mount.c from lib/posix to lib/other
David van Moolenbroek [Thu, 13 Aug 2009 09:55:50 +0000 (09:55 +0000)]
move _mount.c from lib/posix to lib/other

15 years agoFixed compilation issue.
Thomas Veerman [Thu, 13 Aug 2009 09:02:43 +0000 (09:02 +0000)]
Fixed compilation issue.

15 years ago - Moved (u)mount prototypes from unistd.h to sys/mount.h.
Thomas Veerman [Wed, 12 Aug 2009 19:57:37 +0000 (19:57 +0000)]
 - Moved (u)mount prototypes from unistd.h to sys/mount.h.
 - Prepared mount system call to accept multiple mount flags
   instead of just read_only (however, it remains backwards
   compatible).
 - Updated the man mount(2) to reflect new header file usage.
 - Updated badblocks, newroot, mount, and umount commands to use the
   new header file.

15 years agocompile and install mdb(1) by default again
David van Moolenbroek [Tue, 11 Aug 2009 09:08:33 +0000 (09:08 +0000)]
compile and install mdb(1) by default again

15 years agovm_assert macros fix up
Tomas Hruby [Tue, 11 Aug 2009 07:53:32 +0000 (07:53 +0000)]
vm_assert macros fix up

15 years agoLance cleanup
Arun Thomas [Wed, 22 Jul 2009 12:36:19 +0000 (12:36 +0000)]
Lance cleanup

15 years agoFixed case error
Thomas Veerman [Wed, 22 Jul 2009 08:59:57 +0000 (08:59 +0000)]
Fixed case error

15 years agoMFS put_inode issues (reported by Maurizio Lombardi)
David van Moolenbroek [Wed, 22 Jul 2009 08:54:28 +0000 (08:54 +0000)]
MFS put_inode issues (reported by Maurizio Lombardi)

15 years agoFix VirtualBox Lance bug; ec_reinit should be called only once.
Arun Thomas [Tue, 21 Jul 2009 13:13:45 +0000 (13:13 +0000)]
Fix VirtualBox Lance bug; ec_reinit should be called only once.

15 years agoAdded select test
Thomas Veerman [Tue, 14 Jul 2009 09:43:33 +0000 (09:43 +0000)]
Added select test

15 years agoFixed a minor select bug:
Thomas Veerman [Tue, 14 Jul 2009 09:39:05 +0000 (09:39 +0000)]
Fixed a minor select bug:
 - When one does a select on a file descriptor that is meaningless for that particular file type, select shall indicate that the file descriptor is ready for that particular operation and that the file descriptor has no exceptional condition pending.