]> Zhao Yanbai Git Server - minix.git/log
minix.git
11 years agocat24c256: driver for the cat24c256 i2c eeprom 79/679/2
Thomas Cort [Mon, 15 Jul 2013 14:29:31 +0000 (10:29 -0400)]
cat24c256: driver for the cat24c256 i2c eeprom

Change-Id: I682d136c80fc868d3f0a5edb0cf6c33c0f3a28ea

11 years agolibi2cdriver: library for i2c device drivers 78/678/2
Thomas Cort [Mon, 15 Jul 2013 14:29:24 +0000 (10:29 -0400)]
libi2cdriver: library for i2c device drivers

Change-Id: Ib6d5617e4b62e0bc5b25e6fa92b44baf536b1961

11 years agoImporting usr.sbin/i2cscan 77/677/1
Thomas Cort [Mon, 15 Jul 2013 14:29:16 +0000 (10:29 -0400)]
Importing usr.sbin/i2cscan

Change-Id: I7b776d6e5f3fe4f7a1af353b995ef9e2e752c7ea

11 years agoi2c: initial bus drivers for am335x and dm37xx 76/676/1
Thomas Cort [Mon, 15 Jul 2013 14:29:09 +0000 (10:29 -0400)]
i2c: initial bus drivers for am335x and dm37xx

Change-Id: I478704fbf30dbf6d3382bcbfb11e75b512c032a1

11 years agokernel: i2c padconf workaround 75/675/1
Thomas Cort [Mon, 15 Jul 2013 14:28:58 +0000 (10:28 -0400)]
kernel: i2c padconf workaround

The padconf library runs in user mode, but to actually affect the
padconf register contents, the processor has to be in privileged
mode. A full server based solution will be developed, but for now
just set the pinmux in the kernel at boot.

Change-Id: I170ed54dae64b27cd9bd8807445231598fb8e3e1

11 years agoreleasetools/release.sh: cleanup & enhancements 73/673/2
Lionel Sambuc [Fri, 12 Jul 2013 16:27:05 +0000 (18:27 +0200)]
releasetools/release.sh: cleanup & enhancements

 - use the proper  make variable to ignore file lists
   SLOPPY_FLIST instead of CHECKFLIST, which was a minix hack.

 - Add BUILDOPTIONS such that the users can activate optional
   features without editing the script.

Change-Id: Id731db96c38c2118c4c58e007f2804008a7a893f

11 years agolibc: revert inadvertent removal of O_CLOEXEC #ifndefs 74/674/1
Thomas Veerman [Fri, 12 Jul 2013 16:38:17 +0000 (18:38 +0200)]
libc: revert inadvertent removal of O_CLOEXEC #ifndefs

Change-Id: I15005181388280363f7aee546bd2ecc0cd5b37e6

11 years agoreleasetools/arm_sdimage.sh 72/672/2
Lionel Sambuc [Fri, 28 Jun 2013 18:27:53 +0000 (20:27 +0200)]
releasetools/arm_sdimage.sh

 - adapt usr partition size
 - reformulate partition sizes for easier adaptations
 - ensure 4k-block alignement of the partitions

Change-Id: I88787614163a7be177acb5dbf454b26387376a47

11 years agoEnable optional GCC install and GCC improvements 71/671/2
Lionel Sambuc [Wed, 24 Apr 2013 16:57:59 +0000 (18:57 +0200)]
Enable optional GCC install and GCC improvements

 -By adding MKGCC=yes and MKGCCCMDS=yes on the make commandline
   it is now possible to compile and install GCC on the system.

   Before doing this, if you are not using the build.sh script,
   you will need to call the fetch scripts in order to retrieve
   the sources of GCC and its dependencies.

 -Reduce difference with NetBSD share/mk

   Move Minix-specific parameters from bsd.gcc.mk to bsd.own.mk,
   which is anyway patched, so that bsd.gcc.mk is now aligned
   on the NetBSD version.

 -Clean libraries dependencies, compiles stdc++ only if gcc is
   also compiled (it is part of the gcc sources)

 -Correct minix.h header sequence, cleanup spec headers.

 -Fix cross-compilation from a 32bit host targeting MINIX/arm

Change-Id: I1b234af18eed4ab5675188244e931b2a2b7bd943

11 years agoAllow install of binutils and import texinfo 82/482/9
Lionel Sambuc [Wed, 10 Apr 2013 16:13:48 +0000 (18:13 +0200)]
Allow install of binutils and import texinfo

 - Enable installing binutils from the base system.

 - Import texinfo which is required for the binutils tools
   to be compiled.

 - Also adapted the fetch rules to correctly generate the
   gitignore files for gcc, and allow the case of multiple
   modules in the same directory, as found in gnu/dist.

Warning: This patch has an entry in docs/UPDATING

Change-Id: Ib781734e8fd7f9c6265fa65d62ba2cf3fccbc5ba

11 years agoUpgrading to binutils-2.22 83/483/6
Lionel Sambuc [Mon, 8 Apr 2013 21:11:36 +0000 (23:11 +0200)]
Upgrading to binutils-2.22

 * Added --with-ld=default
 * Added --with-lto
 * Fixed ld script for ARM
 * Options are prepared for future activation of gold.

Change-Id: Id9618904055e18d1b37f5e9585f775c367ff356e

11 years agoImplement getrusage 69/669/2
Xiaoguang Sun [Tue, 25 Jun 2013 12:41:01 +0000 (20:41 +0800)]
Implement getrusage

Implement getrusage.
These fields of struct rusage are not supported and always set to zero at this time
long ru_nswap;           /* swaps */
long ru_inblock;         /* block input operations */
long ru_oublock;         /* block output operations */
long ru_msgsnd;          /* messages sent */
long ru_msgrcv;          /* messages received */
long ru_nvcsw;           /* voluntary context switches */
long ru_nivcsw;          /* involuntary context switches */

test75.c is the unit test for this new function

Change-Id: I3f1eb69de1fce90d087d76773b09021fc6106539

11 years agoinclude: Import dev/i2c/i2c_io.h 70/670/1
Thomas Cort [Wed, 29 May 2013 18:40:16 +0000 (14:40 -0400)]
include: Import dev/i2c/i2c_io.h

Change-Id: Icb754891349b7f571863f3a87b0318c28c35ac79

11 years agotest42: skip single step test 68/668/1
Ben Gras [Tue, 25 Jun 2013 11:31:40 +0000 (11:31 +0000)]
test42: skip single step test

Change-Id: If57fc467b0ad28e0e9db27c6dadbe49e59dbc463

11 years agokernel: high-hz workaround for do_settime
Ben Gras [Mon, 24 Jun 2013 13:46:20 +0000 (15:46 +0200)]
kernel: high-hz workaround for do_settime

. with hz=1000, clock_t only lasts a few years.
  whenever we can't express the desired realtime
  in ticks because the distance with boottime is
  too high, simply adjust bootime like we do for
  otherwise negative values.
. fixes test 2 on ARM

11 years agokernel: maintain stack alignment
Ben Gras [Mon, 24 Jun 2013 12:35:25 +0000 (14:35 +0200)]
kernel: maintain stack alignment

. 'fixes' 64-bit varargs (i.e. printf("%llu", ..)) argument
  retrieval bug

11 years agore-import makewhatis 67/667/2
Ben Gras [Mon, 24 Jun 2013 14:41:30 +0000 (14:41 +0000)]
re-import makewhatis

. to benefit from the contribution found by rjkovacic@gmail.com
  in the minix makewhatis and submitted upstream, see NetBSD
  PR/47958

Change-Id: I147e1afe0088ab16e7b171f74330d1c6f44a13cf

11 years agokernel, arm ucontext: ARM DBG=-g run fixes 64/664/3
Ben Gras [Sun, 23 Jun 2013 16:37:57 +0000 (18:37 +0200)]
kernel, arm ucontext: ARM DBG=-g run fixes

kernel:
. modules can be as big as the space (8MB) between them
  instead of 4MB; memory is slightly bigger with DBG=-g

arm ucontext:
. r4 is clobbered by the restore function, as it's
  used as a scratch register, causing problems for the
  DBG=-g build
. r1-r3 are safe for scratch registers, as they are
  caller-save, so use r3 instead; and don't bother
  restoring r1-r3, but preserve r4

vfs:
. improve TLL pointer sanity check a bit

Change-Id: I0e3cfc367fdc14477e40d04b5e044f288ca4cc7d

11 years agoretire 64-bit conversion functions 51/651/2
Ben Gras [Mon, 17 Jun 2013 08:31:12 +0000 (10:31 +0200)]
retire 64-bit conversion functions

Change-Id: Ib6b81403f877c363a286c654e0524fa1cb781b80

11 years agoarm:use the 4 bit mmc data bus if supported by the card. 63/663/2
Kees Jongenburger [Fri, 21 Jun 2013 13:52:52 +0000 (15:52 +0200)]
arm:use the 4 bit mmc data bus if supported by the card.

Change-Id: I55de05db23ad1766e5faec93252e8410cbf59001

11 years agobuild:arm remove "start usb" from u-boot command line. 62/662/2
Kees Jongenburger [Fri, 21 Jun 2013 11:20:53 +0000 (13:20 +0200)]
build:arm remove "start usb" from u-boot command line.

Removed "start usb" from u-boot command line as it is not needed and
speeds up tftp booting.

Change-Id: I416f7eb0467888d1b69ef3dad09df0639fb63f10

11 years agoarm:increase the mmc driver process priority. 61/661/2
Kees Jongenburger [Fri, 21 Jun 2013 13:49:09 +0000 (15:49 +0200)]
arm:increase the mmc driver process priority.

Increase driver process priority to prevent starvation. Without
special attribute drivers will have the same priority as other
userland processes this. Posix test 1 was taking so much resources
that it triggered alarms in the MMC driver.

Change-Id: Icd3295d8f2a4a284418327a3715641fe9a3b3043

11 years agoarm:mmc driver refactor and timeout fix for BeagleBoneBlack. 60/660/3
Kees Jongenburger [Tue, 18 Jun 2013 19:39:04 +0000 (21:39 +0200)]
arm:mmc driver refactor and timeout fix for BeagleBoneBlack.

When we send MMC commads that contain data the controller provides no
description of the order of events and we need to be able to send data
from and to the controller "when needed". Changed the code to react
on buffer read and buffer write ready based on interrupts.

Change-Id: I60c9140bf0e45b74be6475054564d4e1bd89f21e

11 years agopadconf: remove duplicate code. 59/659/2
Thomas Cort [Wed, 19 Jun 2013 11:00:44 +0000 (07:00 -0400)]
padconf: remove duplicate code.

The same code block was repeated twice (copy & paste error).

Change-Id: I2035c9fff46b3f0a52084c2d6c871bbda6f8f07f

11 years agoarm timer fix 58/658/2
Ben Gras [Wed, 19 Jun 2013 09:39:05 +0000 (11:39 +0200)]
arm timer fix

. set 'done' once initialized so 32-bit read frc works,
  thanks to keesj
. make sure the software-implemented upper 32 bit of the 64-bit
  "tsc" value works OK by adding an assert in one of its calls

Change-Id: I5ce24fea919f4610c6a86ac7ec9f04b1815620c2

11 years agoarm:mmc driver refactor. 57/657/1
Kees Jongenburger [Thu, 13 Jun 2013 12:53:59 +0000 (14:53 +0200)]
arm:mmc driver refactor.

Generalize the usage of mmc_send_cmd function to allow
it to transfer data and remove direct invocations to
mmchs_send_cmd.

Change-Id: Iabb9a7d3f4ec57536407b369531ded2b8b649cce

11 years agoLLVM bitcode for almost the whole source tree 55/655/3
Anton Kuijsten [Mon, 17 Jun 2013 14:51:49 +0000 (14:51 +0000)]
LLVM bitcode for almost the whole source tree

Change-Id: Ibc6f89de112e0649c3a3bfe49c2baa2ed0746212

11 years agoclkconf: move clkconf from libgpio to it's own lib 54/654/2
Thomas Cort [Mon, 17 Jun 2013 12:23:31 +0000 (08:23 -0400)]
clkconf: move clkconf from libgpio to it's own lib

Clock configuration will be needed by other/future subsystems such as i2c.
Extract the functionality from libgpio and put it into it's own library.

Change-Id: I6f6de0b3fb4d305ddfeac74123b78c983d1318dd

11 years agofix for tap tests 53/653/2
Ben Gras [Mon, 17 Jun 2013 09:12:17 +0000 (09:12 +0000)]
fix for tap tests

Change-Id: Ic79cfdaf420cbf15a1aba6e30c821f35b726d0da

11 years agolibsys: correct cast priority 52/652/1
Ben Gras [Mon, 17 Jun 2013 08:40:28 +0000 (10:40 +0200)]
libsys: correct cast priority

Change-Id: I223482884c0e27726e8100209b203a33ded66f59

11 years agokernel, libsys: make it arm-target-independent 49/649/2
Ben Gras [Sun, 16 Jun 2013 22:33:47 +0000 (00:33 +0200)]
kernel, libsys: make it arm-target-independent

. by making the address and frequency of the
  free running clock kinfo members, set at runtime
  in the kernel, instead of compile time constants
  in libsys

Change-Id: I4a8387302d4d3ffd47d2448525725683a74c9a4f

11 years agomake test run script produce TAP if desired 48/648/2
Ben Gras [Sun, 16 Jun 2013 17:53:18 +0000 (17:53 +0000)]
make test run script produce TAP if desired

. primarily intended for jenkins
. see http://testanything.org/ for more info

Change-Id: If0e942f75898eed640f9edd89ef8ffd0bab9726e

11 years agoa few DBG=-g build fixes, both arches 47/647/3
Ben Gras [Sat, 15 Jun 2013 23:43:39 +0000 (23:43 +0000)]
a few DBG=-g build fixes, both arches

- setlogin() not available
- softfloat needed in -lminc for arm
- libminc: we have to set LIBCDIR for the included files,
  so use that instead of LIBCSRCDIR

Change-Id: I7f92621ebbca9ce08dc377b3fa61dee089071757

11 years agosome arm uart fixes 46/646/2
Ben Gras [Thu, 13 Jun 2013 13:08:26 +0000 (15:08 +0200)]
some arm uart fixes

. always keep reading data from uart so the interrupt is
  not continually asserted if data is sent but no process
  reads it
. increase tx & rx fifo trigger levels -> reduces the number
  of interrupts necessary
. bigger rx/tx buffers

Change-Id: I3cf7c73b22ae2fc091b845d516ba4aa53e892cda

11 years agovfs: patch for unpause()/revive() race condition 45/645/2
Ben Gras [Wed, 12 Jun 2013 18:42:41 +0000 (18:42 +0000)]
vfs: patch for unpause()/revive() race condition

. unpause() and revive() can race - revive() can run during
  a device i/o unblock, causing two sendnb()s to occur, and the
  2nd one to fail
. this can easily happen when a process is blocking on tty and
  is then killed by a signal - tty cancels the i/o and then
  kills the process by a signal

Change-Id: Ia319acaedfa336b78c030a2c4af7246959bdcf87

11 years agoarm:remove pre 1:1 mapping workarounds. 32/632/3
Kees Jongenburger [Wed, 12 Jun 2013 08:19:22 +0000 (10:19 +0200)]
arm:remove pre 1:1 mapping workarounds.

Change-Id: I5a690cf5a561cdca9b9c1f031402f80fd203c92d

11 years agoarm:no longer rely on a 1:1 phys mapping for device memory. 31/631/3
Kees Jongenburger [Tue, 11 Jun 2013 14:07:43 +0000 (16:07 +0200)]
arm:no longer rely on a 1:1 phys mapping for device memory.

Change-Id: Ie3f61069f882c37dbb81dee813fdfd883e7468cf

11 years agoallow ldd mmap fallback 30/630/2
Anton Kuijsten [Wed, 12 Jun 2013 05:58:46 +0000 (05:58 +0000)]
allow ldd mmap fallback

Change-Id: Ia7b8aa88c3f063c804869065719ac0cbf049b7dd

11 years agois: leave newline for stacktrace for SF9 dump 29/629/2
Ben Gras [Tue, 11 Jun 2013 15:19:38 +0000 (15:19 +0000)]
is: leave newline for stacktrace for SF9 dump

. otherwise the standard console screen can't hold them

Change-Id: Iea2c047b715bc456747aace8957d8c5eeb94d0fa

11 years agoinline sendnb should not call send vector 28/628/2
Ben Gras [Tue, 11 Jun 2013 15:13:52 +0000 (15:13 +0000)]
inline sendnb should not call send vector

. also vfs has to reply to a vm call - so use asynsend for that

Change-Id: I30ac1e591191dea5c99e25b03151a4415d1151b0

11 years agoinit: handle getty string of "" (default) 27/627/2
Ben Gras [Tue, 11 Jun 2013 14:30:31 +0000 (14:30 +0000)]
init: handle getty string of "" (default)

. the default entry for disabled ttys for getty
  in /etc/ttys is "", but init crashed on handling that
  string.

Change-Id: Ib7cd6c6869e338f47df0aa5abed36f15eda4f6ff

11 years agotty: i386: rs232 fix 26/626/2
Ben Gras [Tue, 11 Jun 2013 14:11:57 +0000 (14:11 +0000)]
tty: i386: rs232 fix

. ignore interrupt (stop interrupt check loop) if
  interrupt bit not set; limit loop too
. mask off other bits when testing bits in the status register
. this fixes rs232 output that would otherwise never get re-triggered
  as too many bits were set in the status byte to match the
  possibilities.

Change-Id: I311c93377fa8fb477ee9a756455fdeda780e6ba1

11 years agolibc syslog, syslogd, logger, uds fixes 25/625/2
Ben Gras [Mon, 10 Jun 2013 23:47:42 +0000 (23:47 +0000)]
libc syslog, syslogd, logger, uds fixes

changes necessary for libc syslog() using a unix domain socket.

. libc syslog: don't use send() connect() for unix
  datagram sockets, minix wants write() and ioctl()
. syslogd: listen on _PATH_LOG unix domain socket
. logger: warnings fixes
. pfs: make uds dgram socket type nonblocking so
  syslog() doesn't block
. vfs: add sanity check for empty fd in unpause()

Change-Id: Ied136c6fe0cc288f5a53478f1eebccc1ab1f39fb

11 years agoGold linker support for entire source tree
Anton Kuijsten [Tue, 11 Jun 2013 19:08:05 +0000 (19:08 +0000)]
Gold linker support for entire source tree

11 years agoImporting usr.bin/shuffle
Thomas Cort [Sun, 2 Jun 2013 19:45:38 +0000 (19:45 +0000)]
Importing usr.bin/shuffle

11 years agoarm:tty enable FIFO 19/619/4
Kees Jongenburger [Fri, 31 May 2013 14:20:38 +0000 (16:20 +0200)]
arm:tty enable FIFO

Change-Id: Iaaa7c0bcb40c2cce2423c1defc14809ed6c70e9c

11 years agolibexec/fptd/ftpcmd.c: Removed as it is generated. 17/617/2
Lionel Sambuc [Sun, 2 Jun 2013 16:36:15 +0000 (18:36 +0200)]
libexec/fptd/ftpcmd.c: Removed as it is generated.

Change-Id: Id700ed9c2462e998f9c74482f80ee95b644a1399

11 years agovm: bugfix: cover vfs_request-in-callback case 21/621/1
Ben Gras [Wed, 5 Jun 2013 15:02:56 +0000 (15:02 +0000)]
vm: bugfix: cover vfs_request-in-callback case

Change-Id: I16fccd9938fe8edab83c6c2e327d27d65ff20224

11 years agoImporting libexec/fingerd 16/616/1
Thomas Cort [Sat, 1 Jun 2013 14:14:33 +0000 (14:14 +0000)]
Importing libexec/fingerd

11 years agoImporting usr.bin/finger
Thomas Cort [Sat, 1 Jun 2013 14:20:33 +0000 (14:20 +0000)]
Importing usr.bin/finger

11 years agoImporting usr.bin/bdes
Thomas Cort [Fri, 31 May 2013 22:12:09 +0000 (22:12 +0000)]
Importing usr.bin/bdes

11 years agoremove vm_yield_block_get_block leftovers. 15/615/1
Qi Yong [Sat, 1 Jun 2013 12:51:31 +0000 (20:51 +0800)]
remove vm_yield_block_get_block leftovers.

11 years agofix some warnings 12/612/1
Ben Gras [Fri, 31 May 2013 20:37:25 +0000 (20:37 +0000)]
fix some warnings

Change-Id: I1f3fa582d8e3ddb3c981784dddb13e740102f4f3

11 years agovm: handle disappearing process case 11/611/1
Ben Gras [Fri, 31 May 2013 17:31:55 +0000 (17:31 +0000)]
vm: handle disappearing process case

Change-Id: Id96759883e4cdb175c79dcef7ef5ff254612101f

11 years agold.elf_so, grep, make: use mmap() 66/566/14
Ben Gras [Sat, 16 Mar 2013 03:19:47 +0000 (03:19 +0000)]
ld.elf_so, grep, make: use mmap()

. let the dynamic linker use a real file mmap(), reverting
  most of the minix-specific changes
. except for a fallback mode for systems or filesystems that
  do not support mmap() on files
. make, grep: let them use mmap() too, with fallback

Change-Id: I11c0c9ad439d713917a6f5c2d196558f6c03eabd

11 years agovm: mmap support 44/544/15
Ben Gras [Tue, 7 May 2013 12:36:09 +0000 (12:36 +0000)]
vm: mmap support

. test74 for mmap functionality
. vm: add a mem_file memory type that specifies an mmap()ped
  memory range, backed by a file
. add fdref, an object that keeps track of FD references within
  VM per process and so knows how to de-duplicate the use of FD's
  by various mmap()ped ranges; there can be many more than there can
  be FD's
. turned off for now, enable with 'filemap=1' as boot option

Change-Id: I640b1126cdaa522a0560301cf6732b7661555672

11 years agopfs, vfs: increase various limits 99/599/3
Ben Gras [Tue, 28 May 2013 19:29:08 +0000 (19:29 +0000)]
pfs, vfs: increase various limits

. pipes in pfs
. vnodes in vfs
. thread stack sizes in vfs

Change-Id: Ib27dedd42f57a90821a5b950cd7ea25cb2b42f3f

11 years agovfs: mmap support 43/543/15
Ben Gras [Tue, 7 May 2013 12:41:07 +0000 (12:41 +0000)]
vfs: mmap support

. libc: add vfs_mmap, a way for vfs to initiate mmap()s.
  This is a good special case to have as vfs is a slightly
  different client from regular user processes. It doesn't do it
  for itself, and has the dev & inode info already so the callback
  to VFS for the lookup isn't necessary. So it has different info
  to have to give to VM.
. libc: also add minix_mmap64() that accepts a 64-bit offset, even
  though our off_t is still 32 bit now.
. On exec() time, try to mmap() in the executable if available.
  (It is not yet available in this commit.)
. To support mmap(), add do_vm_call that allows VM to lookup
  (to ino+dev), do i/o from and close FD's on behalf of other
  processes.

Change-Id: I831551e45a6781c74313c450eb9c967a68505932

11 years agofix inet warnings 10/610/2
Ben Gras [Fri, 31 May 2013 14:30:13 +0000 (14:30 +0000)]
fix inet warnings

Change-Id: Ia6e761d2c649f03a49b2646387f859d6c2a646cb

11 years agoImporting usr.bin/logname
Thomas Cort [Fri, 31 May 2013 12:22:47 +0000 (12:22 +0000)]
Importing usr.bin/logname

11 years agopanic: declare as printf-style-checked 09/609/2
Ben Gras [Sat, 2 Feb 2013 00:55:35 +0000 (00:55 +0000)]
panic: declare as printf-style-checked

. and related fixes

Change-Id: I5131ac57dc53d8aec8d421a34c5ceea383404d7a

11 years agovfs: who_p fix 50/550/3
Ben Gras [Mon, 13 May 2013 19:06:12 +0000 (19:06 +0000)]
vfs: who_p fix

Change-Id: I0e04b6460907f5e67f6c90b2038d296d66b9a414

11 years agotcpd: waitpid() for any child process 06/606/1
Ben Gras [Thu, 30 May 2013 19:41:45 +0000 (19:41 +0000)]
tcpd: waitpid() for any child process

. prevents zombies from accumulating for as long as tcpd exists
  for child processes that create their own process group, e.g. ftpd.

Change-Id: Ic56597810f7d17e1cc28b755e2d1998ecf49a27c

11 years agoimport netbsd ftpd 05/605/3
Ben Gras [Thu, 30 May 2013 14:47:40 +0000 (14:47 +0000)]
import netbsd ftpd

Change-Id: Id7a3dbd40a6f37c55bcbb0d1456301f60626298f

11 years agobuild:arm allow to store build settings in a .settings file. 04/604/4
Kees Jongenburger [Thu, 30 May 2013 06:47:39 +0000 (08:47 +0200)]
build:arm allow to store build settings in a .settings file.

If present a file called .settings will be sourced when
calling the arm_sdimage.sh script. This allows to configure
the build setting for arm build without the need to specify
the flags each time arm_sdimage.sh is called.

For example to configure arm_sdimage.sh to target the
beaglebone put the following content in your .settings file.

BASE_URL=http://www.minix3.org/arm/beaglebone
FLAG=-DAM335X
CONSOLE=tty00

Change-Id: Ide0341a255bc7f41dc5c30d0a43e71b01a8ea55c

11 years agoAdd acpi poweroff
Xiaoguang Sun [Thu, 23 May 2013 11:06:13 +0000 (19:06 +0800)]
Add acpi poweroff

Use acpi poweroff if it's possible.

Change-Id: I103cc288523bf63fa536750b1d408ac88bbe35fb
Signed-off-by: Ben Gras <ben@minix3.org>
Signed-off-by: Tomas Hruby <tom@minix3.org>
11 years agovm: a bit more informative about failed pagefaults 00/600/2
Ben Gras [Wed, 29 May 2013 11:48:54 +0000 (11:48 +0000)]
vm: a bit more informative about failed pagefaults

Change-Id: I2b72dfb9291670cb837dfdb279f519892575d4a6

11 years agogen_uEnv.txt.sh shell neutraling 01/601/1
Ben Gras [Wed, 29 May 2013 12:10:22 +0000 (12:10 +0000)]
gen_uEnv.txt.sh shell neutraling

Change-Id: I52d5ac548373db1c4db9af79a24c2c9e2c1c1dcb

11 years agosyslog.conf: correct PID file path in comments. 98/598/2
Thomas Cort [Tue, 28 May 2013 19:07:10 +0000 (15:07 -0400)]
syslog.conf: correct PID file path in comments.

The syslogd.pid file has moved from /usr/run to /var/run. The comments
in syslog.conf direct the user to restart syslogd using the old PID
file location. This patch updates the comment to match the code.

Change-Id: Ib40cd1199ce201e1cbc74e1f1393d55a166ef343

11 years agoservers/pm: Fix EFAULT check for SETGROUPS 97/597/1
Lionel Sambuc [Tue, 28 May 2013 19:08:15 +0000 (19:08 +0000)]
servers/pm: Fix EFAULT check for SETGROUPS

The previous test would return EFAULT as soon as the group pointer
was NULL, while it is sensible when the count is also 0.

In that case, the SETGROUP syscall is expected to clear all the
group entries as the new set is empty.

Change-Id: I07b7e1d1f023a52e3035d53f7d9b42b660e039e8

11 years agofix some comment typo's
Qi Yong [Sun, 26 May 2013 14:11:59 +0000 (14:11 +0000)]
fix some comment typo's

Change-Id: Ic9f4333613abb858bfc28f72685424557cf2cd40

11 years agoarm:build add flags for AM335X 94/594/3
Kees Jongenburger [Fri, 24 May 2013 09:04:53 +0000 (11:04 +0200)]
arm:build add flags for AM335X

Change-Id: I1fd25d5ada52972fd2f6ec5057ff273544c65390

11 years agoarm:also use 1MB sections for mapping AM335X device memory. 93/593/3
Kees Jongenburger [Fri, 24 May 2013 08:28:00 +0000 (10:28 +0200)]
arm:also use 1MB sections for mapping AM335X device memory.

Change-Id: Idc0b285fcbabe8ec4c0be9a600b6a720c0bd3ffc

11 years agoarm:adapt MMC driver to also work on the AM335X platform. 92/592/3
Kees Jongenburger [Fri, 24 May 2013 08:26:51 +0000 (10:26 +0200)]
arm:adapt MMC driver to also work on the AM335X platform.

Change-Id: I30e69b7bfd377d1dd2b0b458dd00ee411b060285

11 years agoarm:mmc only compile mmchs when targeting arm. 95/595/1
Kees Jongenburger [Fri, 24 May 2013 13:39:41 +0000 (15:39 +0200)]
arm:mmc only compile mmchs when targeting arm.

Change-Id: Ie8d0b6461b58c2a7becda98401dece81be607a01

11 years agoarm:adapt tty driver for AM335X. 91/591/2
Kees Jongenburger [Fri, 24 May 2013 08:25:30 +0000 (10:25 +0200)]
arm:adapt tty driver for AM335X.

Change-Id: Ic861d68d0576c2abcdcb265a450c013ea2b28499

11 years agoarm:build also generate cmdline.txt 90/590/2
Kees Jongenburger [Fri, 24 May 2013 08:23:44 +0000 (10:23 +0200)]
arm:build also generate cmdline.txt

Also generate cmdline.txt as we need to use a different serial
for the AM335X boards.

Change-Id: I18c1b5911b5d7238d4a765e0aa01613908d9f2e6

11 years agoarm:omap timers remove hardcoded base address. 86/586/4
Kees Jongenburger [Thu, 23 May 2013 16:02:23 +0000 (18:02 +0200)]
arm:omap timers remove hardcoded base address.

Omap timers remove hardcoded base address and add some initial
support for the beaglebone's timers. Frclock_util will need
refactoring to remain independent of the ARM flavour.

Change-Id: I2b5d04e930364262c81b5686de634c0a51796b23

11 years agoarm:configure tick per ms to reflect the free running clock. 85/585/4
Kees Jongenburger [Thu, 23 May 2013 15:58:51 +0000 (17:58 +0200)]
arm:configure tick per ms to reflect the free running clock.

Change-Id: Ifc02658d6ae48dd01a868bfaa3d60f77bc6cc800

11 years agoarm:interupt handling remove hardcoded base address. 84/584/4
Kees Jongenburger [Thu, 23 May 2013 12:25:14 +0000 (14:25 +0200)]
arm:interupt handling remove hardcoded base address.

Remove hardcoded base address for the omap interrupt handler and add
interrupt names for AM335X in omap_intr.h.

Change-Id: Ie606d8612f55990d55f9db655583052f53950e8e

11 years agoarm:add AM335X serial 83/583/3
Kees Jongenburger [Thu, 23 May 2013 11:16:24 +0000 (13:16 +0200)]
arm:add AM335X serial

Change-Id: I728622ddd4f59529d15e8ac2499c41fa815eee3b

11 years agoarm:keep kernel and modules in the first 256MB of memory. 77/577/4
Kees Jongenburger [Thu, 23 May 2013 08:45:38 +0000 (10:45 +0200)]
arm:keep kernel and modules in the first 256MB of memory.

Keep kernel and modules in the first 256MB of memory in preparation
for the beaglebone. That target only has 256 MB of memory.

Change-Id: I3d92247b5d4e5d3aab7388fe01c2f5713d6a4593

11 years agoarm:remove hardcoded base address for the in kernel serial. 82/582/3
Kees Jongenburger [Thu, 23 May 2013 08:03:09 +0000 (10:03 +0200)]
arm:remove hardcoded base address for the in kernel serial.

Removed hardcoded base address for in kernel serial. This will ease
porting to different boards and allow us to remap i/o at later stage.

Change-Id: I4a4e00ed2aa2f94dfe928dc43a6816d3b94576b7

11 years agoarm:kernel simplify the in kernel serial. 81/581/3
Kees Jongenburger [Wed, 22 May 2013 14:40:45 +0000 (16:40 +0200)]
arm:kernel simplify the in kernel serial.

Simplify the in kernel serial header to a minimum. The driver doing
the real handling is the tty driver.

Change-Id: I5d487d71a3d22906313aa8af5e9d84b0751a6868

11 years agoarm:add DDM37XX cflag 80/580/3
Kees Jongenburger [Wed, 22 May 2013 09:25:59 +0000 (11:25 +0200)]
arm:add DDM37XX cflag

We are planning on using the NetBSD config system to perform build
system configuration when targeting different hardware variants.
This functionality however is not integrated yet and we currently
rely on using ifdef's in the code. This commit add the DDM37XX flag.

Change-Id: Ia25dfa520c7df4b648f4d1a583808de774acff63

11 years agoarm:build generate uEnv.txt 79/579/3
Kees Jongenburger [Wed, 22 May 2013 07:12:13 +0000 (09:12 +0200)]
arm:build generate uEnv.txt

Generate uEnv.txt to allow to generate an uEnv containing commands
to boot over the network.

Change-Id: I8bcd831ee474d837411568bf1c1e2c2d96ec931b

11 years agoreleasetools/arm_sdimage.sh: Improvements 89/589/2
Lionel Sambuc [Thu, 23 May 2013 16:09:19 +0000 (18:09 +0200)]
releasetools/arm_sdimage.sh: Improvements

 - Added the possibility to define the number of parallel jobs from the
   command line with (example, default is 1):
   $ JOBS=8 ./releasetools/arm_sdimage.sh

 - Generate filesystem filling the reserved partition area, instead of
   being smaller.

 - Added a possibility to specify additional parameter to the build.sh
   script from the command line with (example):
   $ BUILDVARS='-N0 -V MYVAR=yes' ./releasetools/arm_sdimage.sh

Change-Id: I5c34268ebf7d45a8d3408b45b1dc5051d775826c

11 years agousr.sbin/mkfs: Allow -b SIZE with proto file 88/588/2
Lionel Sambuc [Thu, 23 May 2013 16:02:58 +0000 (18:02 +0200)]
usr.sbin/mkfs: Allow -b SIZE with proto file

Previously, if a proto file was given, the -b argument would be ignored.

Also added:
 - if the number of inodes is not given, then an estimation is used.
 - sanity checks where moved from one special case to the general case.

Change-Id: I2abfa52bf34206c9087c64b5bfc26af866eb47cb

11 years agoFix symlink issue when generating an ARM image 87/587/2
Lionel Sambuc [Thu, 23 May 2013 16:00:34 +0000 (18:00 +0200)]
Fix symlink issue when generating an ARM image

A few symlinks were pointing to a nonexistent file named '(null)'
instead of their intended target. This was only seen when using the
arm_sdimage.sh script.

There is two ways of specifying links and directories. the first one
using an entry in one of the mtree files, and the second one in
/etc/Makefile.

Those entries where doubled, and one of them would specify some rights,
while the other had the target.

By removing those entries, I make sure there is only one definition of
these symlinks, which solves the problem I was seeing on an ARM image.

These symlinks are still present on a generated system, as they are
required.

Change-Id: I9ced8860f72d7c4d686a09720de4d8257d6e04fa

11 years agomkfs: only complain about failing stat with -d 78/578/2
Ben Gras [Thu, 23 May 2013 00:43:33 +0000 (02:43 +0200)]
mkfs: only complain about failing stat with -d

mkfs -d is a feature that needs a stat() of the mkfs binary, which
often fails (as mkfs is often not invoked with a full path or from
the same directory). it makes setup look a bit messy as the error
is always printed while installing the system, for each created FS,
even though the situation is harmless. This change only complains
when the stat is actually needed (-d).

Change-Id: I54ac01505aa97c1cbe40456c04a35aed5a7ee953

11 years agoarm: recognize instruction fetch pagefault 68/568/1
Ben Gras [Tue, 21 May 2013 12:57:57 +0000 (14:57 +0200)]
arm: recognize instruction fetch pagefault

. an instruction fetch pagefault generates a prefetch
  abort exception, this should also be a source of
  a pagefault event
. the saved_lr argument to the C exception handler was
  sometimes an address, sometimes a pointer to that address;
  the kernel sometimes dereferences it, so it should always be
  an in-kernel address, never a userspace address. fix in mpx.S
  makes it always a pointer.
. move dumping of all processes over serial out of
  the arch-specific arch_system.c

11 years agoat_wini: enable busmastering in pci if necessary
Ben Gras [Sat, 18 May 2013 13:24:23 +0000 (13:24 +0000)]
at_wini: enable busmastering in pci if necessary

11 years agovirtio_blk: always start if found 59/559/2
Ben Gras [Wed, 15 May 2013 15:45:33 +0000 (15:45 +0000)]
virtio_blk: always start if found

Change-Id: Iec70c67e1b675297b5ffb800a9fb263feffa9271

11 years agoarm:vm caching fix. 65/565/1
Kees Jongenburger [Thu, 16 May 2013 08:11:12 +0000 (10:11 +0200)]
arm:vm caching fix.

Improve reliability by using write trough cache.

11 years agoarm:vm header cleanup. 64/564/1
Kees Jongenburger [Thu, 16 May 2013 08:11:12 +0000 (10:11 +0200)]
arm:vm header cleanup.

11 years agoarm:make no assumptions about TRE and AFE 63/563/1
Kees Jongenburger [Thu, 16 May 2013 07:50:55 +0000 (09:50 +0200)]
arm:make no assumptions about TRE and AFE

The bootloader can leave the system control register
in at state that doesn't match our setup. make no assumptions
and configure TRE and AFE.

11 years agobuild:arm enable verbose mode by default. 62/562/1
Kees Jongenburger [Thu, 16 May 2013 08:20:01 +0000 (10:20 +0200)]
build:arm enable verbose mode by default.

11 years agobuild:arm disable I and D cache before loading the kernel. 61/561/1
Kees Jongenburger [Thu, 16 May 2013 08:18:07 +0000 (10:18 +0200)]
build:arm disable I and D cache before loading the kernel.

11 years agobuild:copy uEnv and cmdline.txt onto OBJ 60/560/1
Kees Jongenburger [Sat, 27 Apr 2013 07:50:04 +0000 (09:50 +0200)]
build:copy uEnv and cmdline.txt onto OBJ

11 years agobuild:make sdimage more silent. 57/557/2
Kees Jongenburger [Sat, 27 Apr 2013 20:58:49 +0000 (22:58 +0200)]
build:make sdimage more silent.

Change-Id: I9bba37c93c27c599ec914d1f069ca21e6cbfe111