Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
162,846 Commits 606 Branches 0 Tags 3.1 GiB
ae8068f5f7
Commit Graph

90104 Commits

Author SHA1 Message Date
nisimura
ae8068f5f7 still remain crude, work in progress. 2007-10-28 03:15:04 +00:00
dsl
b279dc62eb Instead of having 2 copies of the sycall entry code, reprocess the file 
to generate syscall_plain() with #defines to remove the trace calls.
Join the club of making micro-optimisations to this code to speed up
benchmarks of empty system calls (eg getpid()):
- Use __predict_false() to get fewer branches in the 'normal' path.
- Just copy all the arguments from the stack frame to 'args', since the
  jump indirect for the switch statement is basically unpredicatable.
- Keep the (rather pointless) test for 0 arguments - just for higher
  benchmark scores.
NB system call timing are dominated by other factors! I measured getppid()
  as faster than getpid(), sometimes open("foo", -1) was also faster.
I also suspect that if trace_is_enabled(p) were a simple bit test (that
could be inlined), then the cost of the tests would be unmeasurable and
the syscall_plain/fancy fiasco could be expunged completely.
 2007-10-27 22:56:41 +00:00
plunky
60822ed9d3 When no estimate of battery minutes_left can be made, apm should 
return a value of 0.
 2007-10-27 19:51:29 +00:00
pooka
eff8fdc9d1 Prevent recursion in INACTIVE -> vrecycle -> inactive -> ... 2007-10-27 19:36:34 +00:00
dsl
7cfa26e69a Remove some gratuitous differences between syscall_fancy and syscall_plain. 2007-10-27 18:41:57 +00:00
tsutsui
7eee2a15d6 Use PCI bus/dev/function numbers rather than device unit 
to check first tlp in device_register().
 2007-10-27 17:23:37 +00:00
nisimura
b08d93761c stop to generate SREC kernel image. netboot is now considered a 
standard path to boot off. DINK32 and ppcboot can handle it as a
staging bootloader.
 2007-10-27 15:36:29 +00:00
tsutsui
3d88d4ee08 Assume netboot if "nfsroot=/path" is specified in bootstring. 2007-10-27 15:12:09 +00:00
tsutsui
683d044ea4 In device_register(), check booted devices more properly: 
- check "tlp" name on netboot
- use ata channel and drive numbers to check wd unit
 2007-10-27 14:47:04 +00:00
tsutsui
d1b230ab5a Include <lib/libsa/dev_net.h> for "extern int try_bootp" decl. 2007-10-27 12:23:44 +00:00
tsutsui
fe9146d494 Add an extern decl of try_bootp variable here since MD sources could set it. 2007-10-27 12:21:17 +00:00
tsutsui
03f68714d2 Include "bootp.h" for a function declaration of bootp(). 2007-10-27 12:19:41 +00:00
njoly
d57ae6a62d Regen for chown additions. 2007-10-27 09:17:50 +00:00
njoly
97151f15f1 Add missing chown's syscalls family members. 
This makes chown(1)/chgrp(1) works under compat linux32.

Problem reported and fix tested by Arto Huusko. Thanks.
 2007-10-27 09:16:24 +00:00
nisimura
f1a2c1395e arrange cache line friendly allocation where applicapable. Not 
immutune from careless change of struct local. Away from
compiler extention.
 2007-10-27 06:34:19 +00:00
nisimura
f74459a106 redo again the arrangement. 2007-10-27 05:35:55 +00:00
nisimura
94ca8c41e4 - remove ppb which is unlikey used. 
- a comment about console selection.
 2007-10-27 05:24:29 +00:00
nisimura
e7cbc24fab rearrange lines to make the previous change easy to understand. 2007-10-27 05:10:19 +00:00
nisimura
653a3ff7de - an inch forward southbridge-less configuration. 
- make sure to have a fallback default for the case when no bootinfo
  was prepared by a bootloader.
 2007-10-27 04:32:39 +00:00
nisimura
2cbfeb1621 - make sure to have Rx poll demand issued. 
- more vtophys() fallout.
- dismiss PHY stuff until MII r/w ops written.
 2007-10-27 02:51:59 +00:00
nisimura
a7153f09ae - copy intr.h from prep and make mods. 
- ANSI-fy pci_machdep.c
 2007-10-27 02:06:04 +00:00
nisimura
80011a5b90 off-by-one error to detect MSB of SEEPROM index. 2007-10-27 02:01:42 +00:00
ad
dde5d75e78 Fix spelling. 2007-10-27 01:23:25 +00:00
ad
f7b1ea2b3f Tweak comments. 2007-10-27 01:22:53 +00:00
xtraeme
358f1bd73f Comment out piixpcib(4) until the multiple PRs are fixed. 2007-10-26 22:42:43 +00:00
xtraeme
a2da5d58f1 xen/i386/cpu.c uses i686_mtrr_init_first if MTRR is defined, but 
mtrr_i686.c wasn't built...
 2007-10-26 22:33:23 +00:00
uwe
538e94b48f We don't need MIN_LWP_ALIGNMENT. 2007-10-26 22:24:51 +00:00
xtraeme
72b98e9287 Add amdpcib(4). 2007-10-26 22:21:20 +00:00
xtraeme
19affc8a1a - Add amdpcib(4) commented out. 
- Comment out piixpcib(4) until the multiple PRs are fixed, this driver
  has been broken for long time in some machines.
 2007-10-26 22:20:43 +00:00
xtraeme
7e8be191f1 Share pcib(4) and amdpcib(4) between i386 and amd64; one copy is enough. 2007-10-26 22:17:13 +00:00
xtraeme
66ba6aa324 ichlpcib(4) lives in x86/pci these days... 2007-10-26 22:11:45 +00:00
xtraeme
8129e894df Apply some KNF. 2007-10-26 21:58:17 +00:00
xtraeme
3338a139da - Share pchb(4) between i386 and amd64; one copy is enough for both. 
- Move some of the x86 PCI devices into x86/pci/files.pci.
- Add more x86 stuff into x86/conf/files.x86.

ok joerg.
 2007-10-26 21:49:50 +00:00
pooka
87ffdd1649 RUMUSEROBJDIR -> RUMPUSEROBJDIR (those cold-blooded old times ...) 2007-10-26 18:03:26 +00:00
pooka
6640f4d32a It's non-trivial to say that "no, we don't have cpu counter on this 
arch (rump)", so emulate it.  But this is suboptimal.  The crux of
the problem seems to be that types.h contains both information on
the machine architecture (which we want) as well as other defines
such as __HAVE_CPU_COUNTER (which we don't want).
 2007-10-26 17:59:38 +00:00
pooka
ff9a260335 few stubs 2007-10-26 17:56:24 +00:00
pooka
bf494d7f8d more carefully emulate vget() and vrecycle() 2007-10-26 17:55:43 +00:00
pooka
90f8810031 In threading where available. 2007-10-26 17:36:23 +00:00
ad
630552ea77 ras_sync: deal with spurious wakeup and add comments. 2007-10-26 17:28:37 +00:00
pooka
f977a9f9b5 Read/write can reuse message memory if operating uncached. This 
will change evetually, but for now just appease a KASSERT by
resetting the message header to 0 after each loop.
 2007-10-26 16:54:50 +00:00
nisimura
eae2367539 make sure to have vtophys conversion to feed memory ptr to HW, 
and some consistency among them.
 2007-10-26 14:30:03 +00:00
nisimura
9c0670836c add missing rxstore[] allocation. 2007-10-26 13:56:18 +00:00
nisimura
16ea8193e3 sprinkle copyright notice. 2007-10-26 13:32:57 +00:00
joerg
930e1f6e1d Match delay/DELAY on x86 with delay(9). It takes an unsigned int as 
argument. Use this and replace the inline assembly (mul + div using the
64bit intermediate result) with normal 32bit multiplication and
division. The compiler can turn the division into a multiplication and
shift, making it even cheaper then the original assembly. For extreme
long delays, just use 64bit arithmetic.
 2007-10-26 13:24:39 +00:00
he
c71c45f606 After the machine's bus.h #ifdef was updated, this one needs a tweak as well. 2007-10-26 11:07:32 +00:00
pooka
4f8627af5f define _RUMPKERNEL 2007-10-26 10:44:27 +00:00
peter
f27028f22e Call callout_init before callout_stop. 2007-10-26 10:39:40 +00:00
he
27b6c6515d Correct the symbol used for multiple include protection, we can't 
use _SYS_BUS_H_ here and expect good results when <sys/bus.h> is
included.
 2007-10-26 09:15:54 +00:00
garbled
670c8a4f93 Fix the ranges_bitmap function so it correctly finds the various ranges 
on OFW, and passes them on to ofwoea_batmap correctly.  With this fix,
ofppc now properly initializes all of it's batmappings.

Kept the macppc specific override of the bitmap, but surrounded it with
an #ifdef macppc.  It would probably be worth testing macppc without this
ifdef to see if it works there as well.
 2007-10-26 08:41:24 +00:00
nisimura
b9c06c9fc0 add one more 10/100 Ethernet PCI NIC and fix some others. Needs 
innovation to handle multiple device configuration with match/attach/init
sequence like NB auto-configuration machine.
 2007-10-26 01:16:27 +00:00