Commit Graph

171739 Commits

Author SHA1 Message Date
mrg
8754a8144b move an XXXNETBSD along with the assignment that moved 2008-07-19 22:23:14 +00:00
mrg
f9c3ff3858 merge cornflakes from git-change-04893aa99abfbed8eb6d7067a974fa1f31193c87. 2008-07-19 22:22:17 +00:00
mrg
acf7df4527 fix a merge error 2008-07-19 22:14:13 +00:00
mrg
b0c27870b5 merge cornflakes from git-change-04893aa99abfbed8eb6d7067a974fa1f31193c87. 2008-07-19 22:10:09 +00:00
mrg
60c992bef9 import mesa-drm git at change 04893aa99abfbed8eb6d7067a974fa1f31193c87. 2008-07-19 22:05:02 +00:00
rmind
103b767319 - From Paul Goyette: correct the description of mutex_spin_exit.
- Remove sentence about lockmgr(9), as it is finally replaced.
- Note atomic_ops(3) and membar_ops(3) man pages.
2008-07-19 21:21:38 +00:00
jmcneill
a8309870a5 Pull in agp_i810.h so the agp borrow hack works on i915drm. 2008-07-19 20:08:39 +00:00
jmcneill
e3f4f88866 PCI_PRODUCT works on pa_id, not pa_device 2008-07-19 17:45:53 +00:00
bjs
61ffbe1fec dev->pci_device should be assigned PCI_PRODUCT(pa->pa_id), not the id
itself.
2008-07-19 17:44:14 +00:00
bjs
79b565981a use ALIGN() instead of rolling our own. 2008-07-19 17:28:14 +00:00
bjs
e87f5b54be set dev->kdev = kdev at open time, not in drm_ioctl. 2008-07-19 17:23:04 +00:00
reinoud
3efc85e77e Limit sectorsize to max. 4096 bytes. This prevents a panic when declaring
the node pool since standard pools are limited upto 4096 aparently.
Pagesize limit even?
2008-07-19 16:36:40 +00:00
reinoud
aa9f26d0bd Use the allocation descriptor types selected on mount when growing from an
internal to an extent recording.
2008-07-19 16:23:09 +00:00
reinoud
a2718fb081 Use ecma-167 constants instead of homegrown to prevent double mapping. 2008-07-19 16:14:09 +00:00
reinoud
a1f7916969 Choose which time of allocation descriptors to use on new nodes based on
the number of partition mappings.
2008-07-19 16:00:35 +00:00
jmcneill
4a648a5b9a Implement drm_drawable gunk and drm_locked_task, untested. i915drm links now. 2008-07-19 14:15:11 +00:00
kardel
b6ecf59844 buf_destroy() an embedded buffer before returning memory to pool
issue detected by LOCKDEBUG panicing about "allocation contains active lock"
2008-07-19 13:58:08 +00:00
kardel
ccab995bc6 Initialize timecounter_lock at IPL_HIGH. Prevents a "locking against myself"
when com.c (options PPS_SYNC) delivers a PPS time stamp during clock
interupt processing.
2008-07-19 10:33:58 +00:00
bjs
d63b70db6d Change IPL of DRM_SPININIT macro to IPL_NONE. Currently, the only
mutex initialized by this macro is the dma lock, which certainly
should not be IPL_VM.  The priorities of the other locks should be
revisited.

Also, in DRM_WAIT_ON, enforce the policy of "drm returns -errno
to shared code" (this used to be implemented via a macro).
2008-07-19 10:09:01 +00:00
mrg
3cbb0df829 bare-bones entry for mesa-drm 2008-07-19 07:37:52 +00:00
mrg
41097c94f6 support new drm.
XXX: not fully enabled yet.  config(1) has a makeoptions bug
with complex conditionals.  see files.drm / files.pcidrm for
some instructions on how to enable it.
2008-07-19 07:26:54 +00:00
mrg
242955a02d XXX: remove the (volatile uint32_t *) cast from atomic_read() macro for now.
now all bsd drm drivers build except for i915 which fails to link.
2008-07-19 07:03:07 +00:00
mrg
cc28125d9e XXX: always return DRM_MIGHT_BE_AGP for now -- need to implement the
XXX: releant bus controller grovelling for netbsd pci.
2008-07-19 07:02:15 +00:00
mrg
310b3512f5 XXX rename some macro internal variable names to not clash with users. 2008-07-19 06:48:44 +00:00
mrg
c97fa01b94 rename some inner local variables away from outer scope name conflicts 2008-07-19 06:42:07 +00:00
mrg
215f2c6a8b rename a local variable from free to myfree 2008-07-19 06:40:01 +00:00
mrg
8ff8b7b272 pull over many many changes from netbsd-current sys/dev/{,pci/}drm/.
make the mesa-drm style Makefile's actually run on netbsd (but i doubt
they work yet.)

this still has a lot to go.  it doesn't work yet, but it's getting
close.

most of this is from other authors than myself:  yorrick, drochner,
jared, bjs, all contributed a lot more than i have so far, but there's
a bunch of clean up from myself in here as well.
2008-07-19 06:18:23 +00:00
mrg
3a3ec8ba5f include bsd.subdir.mk 2008-07-19 06:14:25 +00:00
mrg
d582952b83 import mesa-drm git from version 5d27fd94afaaf434c3a92af0075420b550055bfb,
which is approximately june 25th or so.
2008-07-19 05:30:36 +00:00
christos
dac1180a87 fix debugging printf 2008-07-18 21:29:48 +00:00
apb
a41d4962be s/msgs/msgs(1)/ in a comment. From PR 39167. 2008-07-18 21:12:03 +00:00
apb
4ce9a85c41 man page fixes from PR 39164. 2008-07-18 21:03:03 +00:00
christos
19152e60e3 document -j from Dennis den Brok 2008-07-18 20:22:51 +00:00
reinoud
a48555c3a3 Hopefully fix gcc type puned warning. 2008-07-18 16:21:12 +00:00
pooka
1d89d2370c enable DIAGNOSTIC 2008-07-18 16:20:03 +00:00
pooka
28fdf2df68 support cv_has_waiters() 2008-07-18 16:19:12 +00:00
pooka
00e37d073d emulate vlog() 2008-07-18 16:18:04 +00:00
pooka
6ebb8696c9 add pthread_cond_has_waiters_np() 2008-07-18 16:17:11 +00:00
pooka
6fed9eb99d Don't biohazardwait B_ASYNC buffers. 2008-07-18 16:15:56 +00:00
reinoud
e7c5eefcfb fix comment 2008-07-18 15:55:32 +00:00
drochner
93e3fc3cf6 thanks David, but the initialisation might not work, dependent on
the %union used -- remove the "const" instead (and add a "static"
which desn't seem to hurt)
2008-07-18 14:25:37 +00:00
reinoud
5f233d9346 On dirhash lookup, don't retrieve info you already have. 2008-07-18 14:18:51 +00:00
reinoud
f0040136a6 Remove double check 2008-07-18 13:39:41 +00:00
reinoud
43188548d8 First retrieve length and hash values *before* advancing to the next entry.
On hashline collision it would check a random one next.
2008-07-18 13:34:05 +00:00
reinoud
aa9cb3deaa Instead of bombing out when trying to mount a disc RW when UDF_READWRITE is
not defined, downgrade it to a read-only mount.
2008-07-18 12:45:08 +00:00
dsl
6889d181fd groff puts this code through g++ which is rather happier if we give
yyvalzero an initialiser.
2008-07-18 09:57:15 +00:00
rmind
a7fc471680 PR/39155: Wolfgang Solfrank: Invalid parameter to turnstile_exit in rw_downgrade. 2008-07-17 23:00:01 +00:00
christos
c709bea66c Use htonl like ping/traceroute6 do, from Markus Mayer 2008-07-17 19:23:17 +00:00
reinoud
2ac28d554b Implement directory hashing to speed up directory traversals. Speed
improvements of at least 4 times in untarring and roughly 100 to 500 times
on file creation in big directories. Lookup of files was O(n*n) and is now
O(1) even for file creation. Free spaces in the directory are kept in a
seperate list for fast file creation.

The postmark benchmark gives:

UDF old:
pm>set transactions 2000
pm>set number 3000
pm>run
Creating files...Done
Performing transactions..........Done
Deleting files...Done
Time:
        1593 seconds total
        681 seconds of transactions (2 per second)

Files:
        3956 created (2 per second)
                Creation alone: 3000 files (4 per second)
                Mixed with transactions: 956 files (1 per second)
        990 read (1 per second)
        1010 appended (1 per second)
        3956 deleted (2 per second)
                Deletion alone: 2912 files (9 per second)
                Mixed with transactions: 1044 files (1 per second)

Data:
        5.26 megabytes read (3.38 kilobytes per second)
        21.93 megabytes written (14.10 kilobytes per second)
pm>



UDF new:
pm>set transactions 2000
pm>set number 3000
pm>run
Creating files...Done
Performing transactions..........Done
Deleting files...Done
Time:
        19 seconds total
        3 seconds of transactions (666 per second)

Files:
        3956 created (208 per second)
                Creation alone: 3000 files (230 per second)
                Mixed with transactions: 956 files (318 per second)
        990 read (330 per second)
        1010 appended (336 per second)
        3956 deleted (208 per second)
                Deletion alone: 2912 files (970 per second)
                Mixed with transactions: 1044 files (348 per second)

Data:
        5.26 megabytes read (283.66 kilobytes per second)
        21.93 megabytes written (1.15 megabytes per second)
2008-07-17 19:10:22 +00:00
drochner
f2e25e0cba fix a longstanding bug, from Otto Moerbeck:
if we're reducing a rule that has an empty
right hand side and the yacc stackpointer is pointing at the very
end of the allocated stack, we end up accessing the stack out of
bounds by the implicit $$ = $1 action

do it a bit differently than in Openbsd: avoid pulling in <string.h>
because this is not compatible with wacky namespace games we do in libc
2008-07-17 16:28:14 +00:00