Commit graph

3770 commits

Author SHA1 Message Date
Ben Gras
467a01024e bump trunk version to 3.1.9
make issue.install not include version
2010-09-15 08:47:42 +00:00
Ben Gras
2065c9982b include - minor compile fixes (including <ansi.h> in files that use it)
workaround for kernel/debug.h that causes <ansi.h> to be included in mpx.S
indirectly.
2010-09-15 08:47:10 +00:00
David van Moolenbroek
354da24f5b make getsysinfo() a system-land call 2010-09-14 21:50:05 +00:00
David van Moolenbroek
d299a6422b make all other commands use ProcFS 2010-09-14 21:31:56 +00:00
David van Moolenbroek
33b131766f make top(1) use ProcFS 2010-09-14 21:30:23 +00:00
David van Moolenbroek
40c8a7062c make ps(1) use ProcFS 2010-09-14 21:29:40 +00:00
David van Moolenbroek
d7b10963c2 make library calls use ProcFS 2010-09-14 21:28:34 +00:00
David van Moolenbroek
d269f5fb90 ProcFS integration 2010-09-14 21:27:31 +00:00
David van Moolenbroek
2c5c5c06ea ProcFS server, by Alen Stojanov and David van Moolenbroek 2010-09-14 21:25:25 +00:00
David van Moolenbroek
3eb65448a6 VM: expose secondary cache size 2010-09-14 21:22:56 +00:00
David van Moolenbroek
7763fd67c7 PM: save process frame length 2010-09-14 21:21:54 +00:00
Ben Gras
a5a8c57fe7 add EHOSTDOWN errno. 2010-09-14 11:51:41 +00:00
Ben Gras
eaa37fd21c make, lib changes to make gcc/clang format libraries in /usr/lib instead of /usr/gnu/lib. 2010-09-13 16:37:02 +00:00
Ben Gras
23311d9819 lib: fixes to make clang not error 2010-09-13 15:50:54 +00:00
Ben Gras
b7eb096088 Makefiles: remove -D__minix for assembly invocations.
They seem to be redundant (for gcc+ack+clang) and produce
very messy warnings (for clang, due to it already being defined).
2010-09-13 13:36:29 +00:00
Arun Thomas
7c3cfbde1a Update PATHs for llvm and X11 2010-09-12 17:56:53 +00:00
Arun Thomas
9375cc8e1f Remove packit/packman from Makefile 2010-09-12 16:53:45 +00:00
Arun Thomas
42a522f0e8 Remove packman and packit manpages 2010-09-12 13:55:00 +00:00
Arun Thomas
207e88d3c4 Update release script for pkgsrc 2010-09-10 17:00:30 +00:00
David van Moolenbroek
d29833c5aa remove hardcoding of system.conf path in various service calls 2010-09-10 09:22:22 +00:00
David van Moolenbroek
6b902492ae memory driver: propagate data copy error to caller 2010-09-09 16:13:02 +00:00
David van Moolenbroek
2bda06bbb5 shutdown(8): buffer accessed beyond end during option parsing 2010-09-09 08:48:47 +00:00
David van Moolenbroek
2e209097b6 IS: delete obsolete is.h 2010-09-09 08:47:39 +00:00
Tomas Hruby
e6ebac015d APIC mode uses IO APICs
- kernel turns on IO APICs if no_apic is _not_ set or is equal 0

- pci driver must use the acpi driver to setup IRQ routing otherwise
  the system cannot work correctly except systems like KVM that use
  only legacy (E)ISA IRQs 0-15
2010-09-07 07:18:11 +00:00
Erik van der Kouwe
5cc29a6c7e Make realpath deal with double slashes 2010-09-06 07:45:32 +00:00
Tomas Hruby
ad5ecf906b asm.h not installed 2010-09-03 17:01:18 +00:00
Erik van der Kouwe
18ad0b52d3 Make hexdump ACK-compilable and add it to the base system 2010-09-03 07:37:31 +00:00
Arun Thomas
9e4e26e0c8 No more packman, easypack, packit
-Create a dummy packman script that tells users about pkgin
2010-09-03 07:27:39 +00:00
Tomas Hruby
99d9144556 PCI driver uses ACPI if APIC is used.
-  PCI must query ACPI, if (IO)APIC is in use, for the routing
   information and change the ILR (interrupt line register) of each
   device accordingly so drivers use the right IRQ.
2010-09-02 15:44:38 +00:00
Tomas Hruby
7ae6f8740d ACPI driver loaded at boot time
- the acpi driver is part of the ramdisk and if present the rc script
  starts it before pci is started as pci needs if APIC mode is turned
  on.
2010-09-02 15:44:36 +00:00
Tomas Hruby
9560b6dea8 ACPI driver
- 99% of the code is Intel's ACPICA. The license is compliant with BSD
  and GNU and virtually all systems that use ACPI use this code, For
  instance it is part of the Linux kernel.

- The only minix specific files are

  acpi.c
  osminixxf.c
  platform/acminix.h

  and

  include/minix/acpi.h

- At the moment the driver does not register interrupt hooks which I
  believe is mainly for handling PnP, events like "battery level is
  low" and power management. Should not be difficult to add it if need
  be.

- The interface to the outside world is virtually non-existent except
  a trivial message based service for PCI driver to query which device
  is connected to what IRQ line. This will evolve as more components
  start using this driver. VM, Scheduler and IOMMU are the possible
  users right now.

- because of dependency on a native 64bit (long long, part of c99) it
  is compiled only with a gnu-like compilers which in case of Minix
  includes gcc llvm-gcc and clang
2010-09-02 15:44:04 +00:00
Tomas Hruby
43a4725423 pci_*.h headers moved from drivers/pci to include/machine 2010-09-02 15:43:59 +00:00
Tomas Hruby
2440ffae49 Kernel exports DSDP and apic_enabled in machine structure
- kernel exports DSDP (the root pointer where ACPI parsing starts) and
  apic_enabled in the machine structure.

- ACPI driver uses DSDP to locate ACPI in memory. acpi_enabled tell
  PCI driver to query ACPI for IRQ routing information.
2010-09-02 15:43:56 +00:00
Tomas Hruby
344e9221ec Kernel supports up to 64 IRQs
- enough for 2 io apics (usually with 24 pins)
2010-09-02 15:43:54 +00:00
Tomas Hruby
45badf4c05 ACPI in kernel
- the ability for kernel to use ACPI tables to detect IO APICs. It is
  the bare minimum the kernel needs to know about ACPI tables.

- it will be used to find out about processors as the MPS tables are
  deprecated by ACPI and not all vendorsprovide them.
2010-09-02 15:43:51 +00:00
Thomas Veerman
e7252adc1e Clean up the created temp file after running the test. 2010-09-02 09:33:37 +00:00
Ben Gras
eca25a6cab removed redundant includes dependency from gnu-libraries. 2010-09-01 18:48:47 +00:00
Ben Gras
c6cfdbffe4 also make gcc includes before compiling gcc libraries. 2010-09-01 18:28:55 +00:00
Thomas Veerman
0b00cf70b6 - Return ENOENT when trying to add files to removed (but open) directories.
- Add test58 to test this behavior.
2010-09-01 09:07:18 +00:00
Tomas Hruby
bf6f0216d5 tools/Makefile uses padtext only when building with ack 2010-09-01 07:46:51 +00:00
Ben Gras
b3c3a1cb1e gcc compiles - add -fno-builtin so that gcc optimisations don't break things.
- kernel compile was broken with gcc as putchar() was added by gcc in
    stacktrace.c
  - add -fno-builtin everywhere to avoid such problems in the future
  - -fno-builtin in kernel now redundant
2010-08-31 16:42:44 +00:00
Ben Gras
b1847ae244 make service look in /etc/system.conf.d/<progname> first for config file.
This makes it easier to
  - have non-base system drivers (get clobbered by global system.conf)
  - have drivers as packages (can't touch global system.conf)
  - make configs part of the drivers/servers instead of in global file
    (makes system parts more self-contained)
2010-08-31 14:33:31 +00:00
Erik van der Kouwe
1f2054c89c Shutdown changes for multiboot: CTRL-ALT-DEL resets, panic halts 2010-08-30 19:01:58 +00:00
David van Moolenbroek
4448836263 ramdisk: more inodes 2010-08-30 17:08:19 +00:00
Arun Thomas
36fb30dfef cdefs.h: Allow __CONCAT() macro to be nested 2010-08-30 14:34:58 +00:00
Thomas Veerman
a4c30acd9d Improve UDS testing. Fix a few bugs. Add testing of passing file descriptors.
Contributed by Thomas Cort.
2010-08-30 13:49:49 +00:00
Thomas Veerman
e8ddc0f46e - Add support for file descriptor passing to PFS.
- For security reasons move some libc code to PFS.
- Fix a few bugs in PFS.
Contributed by Thomas Cort.
2010-08-30 13:46:44 +00:00
Thomas Veerman
13ef7f1f38 Prepare VFS to support back calls from PFS. For security reasons and to support
file descriptor passing, PFS does some back calls to VFS. For example, to
verify the validity of a path provided by a process and to tell VFS it must
copy file descriptors from one process to another.
2010-08-30 13:44:07 +00:00
Thomas Veerman
2297e26660 Prepare headers to support file descriptor passing over UNIX Domain Sockets.
Contributed by Thomas Cort.
2010-08-30 13:41:22 +00:00
Erik van der Kouwe
a0bb199f69 Release script improvements, moving towards having the ramdisk working again 2010-08-29 19:22:42 +00:00