5aa555542e
Use the upstream xdr, with a rename dance to avoid clash with libc.
...
The upstream xdr provides some functionality expected by the upstream
libnvpair which our libc's xdr does not provide.
When modifying, double-check that libnvpair.so defines no xdr_*
symbols, only _solaris_xdr_*. (XXX Put this note somewhere...)
2015-04-10 22:44:20 +00:00
cee8cbe23f
GCC doesn't speak #pragma init, try __attribute__((constructor)).
2015-04-10 22:38:30 +00:00
708e4d7282
One usually puts .c, not .o, files in SRCS.
2015-04-10 22:36:50 +00:00
495fdb2f81
Initialize variables. Don't bother supporting nv doubles.
2015-04-10 22:29:35 +00:00
0126bcdf7e
Use 0, not NULL, for zero integer constant.
2015-04-10 22:28:27 +00:00
f7e5f84747
Use dprintf, not printf, for debug messages.
2015-04-10 20:55:38 +00:00
7c5ca5f421
Use CPPFLAGS, not CFLAGS, so this actually takes effect.
2015-04-10 20:34:16 +00:00
d4677cf380
Disable mmap on zfs for now.
...
The implementation is not coherent with read/write, and apparently it
is is so broken that write(2) on a region mapped from a zfs file will
often fail with EINVAL because something passes an offset that is too
large to getpages.
2015-04-10 19:59:10 +00:00
057ee3c54c
Pass NetBSD VA_UTIMES_NULL through to ZFS as Solaris ATTR_UTIME.
...
Makes t_unpriv:zfs_times fail the wrong way instead of silently
allowing the change.
2015-04-09 20:18:28 +00:00
3d61be1880
Advisory locking for zfs.
2015-04-09 19:47:05 +00:00
5b72b2208e
Remove some vestigial #if 0 sections.
2015-04-09 19:35:55 +00:00
bc95daf1f6
Fix vattr_init_mask: mode is mode_t, not u_short.
...
Fixes fs/vfs/t_vnops:zfs_attrs test.
2015-04-09 05:32:53 +00:00
8aa34a7d8b
Fix return code of rmdir(nonempty) in zfs: ENOTEMPTY, not EEXIST.
2015-04-09 05:02:28 +00:00
6c33e40a21
These are expressions, not statements. No semicolon.
2015-03-19 22:57:45 +00:00
222dd8ff89
Move initialization of XUIO_XUZC_PRIV out of #ifdef PORT_SOLARIS.
...
This code is not currently used, so no functional change, but it may
be used when we hook zfs into the ubc.
2015-03-19 17:16:22 +00:00
71da105a96
fix reversed test.
2015-03-10 12:17:50 +00:00
bf803657e5
Rewrite dtrace provider prototype from scratch.
2015-03-09 03:43:02 +00:00
4d0ad52103
keep track of the number of enabled probes and set the lockstat_dtrace_enabled
...
flag appropriately.
2015-03-09 01:42:26 +00:00
07b5668221
Remove debug prints, simplify fini logic.
2015-03-09 01:07:27 +00:00
bf089fd6e2
- remove device stuff
...
- compress ops
- terminate array of probes
2015-03-09 00:53:39 +00:00
409555c5e2
Rewrite dtrace lockstat from the Solaris version and simplify.
...
- Fix mutex_adaptive probes to use the right lockstat flags.
- Install lockstat_probe_func on module load, not trace enable.
- Use a more sensible memory barrier arrangement.
2015-03-08 23:56:59 +00:00
007a77aa4b
preliminary support from riastradh (does not work yet, toxic)
2015-03-08 04:13:46 +00:00
aadbd1256a
we have space for 2 more arguments so use it.
2015-03-07 17:47:09 +00:00
6d7ab7eb57
constify
2015-03-07 15:14:29 +00:00
c0342a4ea1
- cleanup FreeBSD ifdefs
...
- remove unused code
- adjust to NetBSD struct emul and types.
2015-03-07 15:14:09 +00:00
56bb5331f6
XXX: Use only a single thread. We seem to have a bug in our threading code
...
that causes us to hang in the ksem code if we use more than one.
2015-03-06 11:49:30 +00:00
45c3185cfc
Fix inconsistency between module and internal names of sdt and fbt
...
PR 49697
OK wiz@
2015-02-26 10:31:52 +00:00
9ce3434ff9
Fix module unload of dtrace, sdt and fbt
...
- Don't unload when there are users of dtrace
- Forbid auto unloading (not supported for now)
PR 49695
2015-02-26 09:10:52 +00:00
2be0e59f43
Fix dtrace build for arm with gcc 4.8
2015-02-21 15:13:20 +00:00
3bcb5a6ccb
Fix dtrace build error with gcc 4.8 on i386 and arm
...
_ILP32 is required by dt_popc.
The fix is inspired by FreeBSD.
2015-02-21 15:00:30 +00:00
47d3523379
Add assert
2015-02-18 03:08:38 +00:00
99f05559ed
Don't return success when the target CPU is offline
...
dt_status originally expected the behavior.
This fixes a segfault happens in dt_aggregate_go (a user of dt_status)
that depends on the behavior.
2015-02-18 03:07:56 +00:00
bf418d8f39
PR/49667: Henning Petersen: Missing initialization of values in dt_parser.c
2015-02-14 15:55:05 +00:00
92111d5005
PR/49666: Henning Petersen: initialization of instr wrong in dt_cg.c NULL
...
instead of 0 for uint32_t.
2015-02-14 15:53:23 +00:00
92e3cdde4e
delete error(1) strings.
2015-02-07 20:33:22 +00:00
b8c591ba5d
check and fix printf formats, negative array indices, uninitialized variables,
...
shadowed variables.
2015-02-07 20:30:03 +00:00
659c05d388
Comment out currently unused dtrace_dof_char function (for clang build).
2015-01-31 13:36:29 +00:00
aaf5c384d9
Change zfs to vcache.
...
- zfs_loadvnode() is based on zfs_zget() plus zfs_znode_alloc().
2014-12-05 09:13:32 +00:00
f43b9b375a
Remove vnode creation from zfs_create_fs() and zfs_create_share_dir()
...
as a vnode is not needed here.
Inspired by FreeBSD r253603.
2014-12-05 09:12:29 +00:00
be13e1550a
don't fail for anonymous unions.
2014-11-01 23:21:31 +00:00
3cf1fc0878
Add libraries in order that they will link if static as well as dynamic
2014-10-30 21:15:13 +00:00
18fd8dcd8d
SPEC_MAXOFFSET_T doesn't exist anywhere else, so don't use it here
...
either.
2014-07-28 10:01:23 +00:00
4575a9c0a5
Fix dtrace kernel module build.
...
Add nodiscard to cdevsw.
They are not fixed in fallocate/fdiscard changes.
2014-07-26 04:54:20 +00:00
02cb0c6eaf
Introduce a selector function to the vfs vnode iterator so that we don't
...
need to vget() vnodes that we are not interested at, and optimize locking
a bit. Iterator changes reviewed by Hannken (thanks), the rest of the bugs
are mine.
2014-05-24 16:34:03 +00:00
7081cf9a59
need bsd.own.mk
2014-04-28 21:56:20 +00:00
0b24109d1b
make compilable from the kernel
2014-04-22 15:12:35 +00:00
c8ad22d607
Add fake atomic 64 ops for those who don't have them.
...
XXX: used to be arm, now it is also i386.
2014-04-22 15:02:38 +00:00
23f76b6d00
An (un)privileged user can easily make the kernel dereference a NULL
...
pointer.
The kernel allows 'data' to be NULL; it's the fs's responsibility to
ensure that it isn't NULL (if the fs actually needs data).
ok christos@
2014-04-16 18:55:17 +00:00
418e8d3033
Don't forget ZFS_EXIT after ZFS_ENTER.
2014-04-09 22:41:38 +00:00
c83470498f
Try vflush, which may fail, before we mark the file system mounted.
2014-04-09 22:40:32 +00:00
897d836e1d
Handle assembly code built with -g
2014-04-05 23:33:15 +00:00
b7c3845c47
Complete migrating to snprintf_blkptr
2014-03-28 03:46:56 +00:00
c5002aae34
Fix migrating to snprintf_blkptr
2014-03-28 03:18:24 +00:00
2821606826
Add missing 4th argument of ddt_object_name
2014-03-28 03:11:00 +00:00
4e2e658100
Fix some snprintf
2014-03-28 02:58:36 +00:00
bc3ce11eda
Add missing 3rd argument of nicenum
2014-03-28 02:50:18 +00:00
6713b766dd
format is a pointer, so sizeof(format) is not a good length
...
to pass to snprintf. Try to calculate the correct length,
taking into account the amount of space already used.
2014-03-27 19:34:39 +00:00
a2d437b386
kill sprintf
2014-03-27 15:50:48 +00:00
1339b2cc13
Add MKCTF
...
This variable separates CTF stuffs from MKDTRACE; we can build DTrace
solely without building and using them. This allows us to use DTrace
even if CTF stuffs have problems (actually they have now).
This variable would be merged into MKDTRACE eventually, once CTF stuffs
work correctly again.
2014-03-25 09:52:55 +00:00
4ae43925ec
Exclude mvsoctmr_* functions from FBT targets
...
Probing them doesn't work for DTrace/arm for some reason.
Anyway with this fix, DTrace/arm works on SHEEVAPLUG
as well as BEAGLEBONE.
2014-03-19 11:14:12 +00:00
33b1d7765e
Change zfs_sync() to use vfs_vnode_iterator.
2014-03-17 09:37:41 +00:00
e2fc5c59a0
Use MACHINE_CPU instead of MACHINE_ARCH with pattern matching
...
Include bsd.own.mk to use MACHINE_CPU.
Advised by matt@
2014-03-16 06:51:43 +00:00
f27e224fab
Look up arm directory on MACHINE_ARCH=*arm*
2014-03-16 05:11:19 +00:00
609ba481a5
Implement DTrace FBT provider for ARM
...
- Finding probable functions
- Replacing instructions
- Emulating instructions
It is tested only on ARMv7 CPUs yet, for example,
-m evbarm (-a earm) kernel=BEAGLEBONE.
2014-03-15 08:00:19 +00:00
48e9045322
Enable invop handler on module load
...
It sets a fbt handler to dtrace_invop_jump_addr on module load,
which is used to call the dtrace probing function from the trap
handler.
2014-03-15 07:49:15 +00:00
26e6962bfe
Make arm/dtrace_isa.c buildable
...
This is a band-aid for build, but it works anyway.
We have to do a real fix in the future.
2014-03-15 07:43:01 +00:00
18c3aa4966
Fix header inclusions
2014-03-10 06:38:22 +00:00
7ea504b141
Make arm/dtrace_subr.c buildable
...
The change just imports the code from
http://people.freebsd.org/~gonzo/arm/dtrace-arm.diff
and tweaks for NetBSD.
2014-03-10 03:04:57 +00:00
60a5bae159
Tidy up paths.
2014-03-09 22:01:31 +00:00
0afc3eb7f2
put back our local fixes:
...
- don't die on unresolved types.
- it is ok to have 0 sized arrays in structs
- forward enum decls.
2014-03-09 20:48:01 +00:00
cf96d463d6
Temporarily add a -S flag to limit the number of threads.
2014-03-09 19:14:15 +00:00
e3afac50d3
fix constant name
2014-03-09 17:07:46 +00:00
7016e57516
use merged ctf rules
2014-03-09 17:05:23 +00:00
abe7c4c08a
Merge ctf build rules.
2014-03-09 17:05:01 +00:00
381ca28d6d
sync with freebsd
2014-03-09 17:04:00 +00:00
959d62341c
Prepare dtrace_isa.c for build
...
- Tidy up header file inclusions
- Replace curcpu with cpu_number()
- Fix argument of dtrace_getreg
- Tweak for gcc 4.8
Note that the file is still not buildable.
2014-03-08 11:46:01 +00:00
e9ddbe399c
Fix build of DTrace with GCC 4.8
...
Use __diagused for a variable which is only referred by KASSERT.
Advised by matt@
2014-03-07 02:37:16 +00:00
f7ff511838
Fix build of DTrace with GCC 4.8.
2014-03-05 20:14:15 +00:00
f6c17acba3
Add cyclic_machdep.c for ARM
...
The code is still almost same as i386's one.
2014-03-05 06:35:44 +00:00
bee7b32afb
Add defined(__arm__)
2014-03-05 06:12:00 +00:00
9d2dfffd73
Use native atomic function instead of DTrace's own one
...
So we don't need to implement it for each architecture.
This change is from FreeBSD (not as-is though).
2014-03-05 06:06:42 +00:00
7ee4c5ecd0
Fix dtrace_cas32 logic
...
- compare with cmp, not new
- return old value
2014-03-05 03:09:17 +00:00
1bf04388c3
Use uint32_t instead of u_int32_t
2014-03-05 02:00:24 +00:00
07efccf499
Add missing END for each function
2014-03-04 11:07:40 +00:00
bc132bbc2a
Get rid of unused variable definition za from mount.h
...
There is no user for it. Even worse, it breaks dtrace build for arm.
2014-03-04 09:24:42 +00:00
ad5eab92b9
Fix wrong conditional statement in dtrace_debug_output
...
Reported in PR kern/48547 by ozaki-r.
2014-03-04 03:20:49 +00:00
0d8247f549
Don't core-dump if a type cannot be resolved. Still gives an error.
2014-03-03 00:09:52 +00:00
97834f7ba0
Change vnode operation lookup to return the resulting vnode *vpp unlocked.
...
Change cache_lookup() to return an unlocked vnode.
Discussed on tech-kern@
Welcome to 6.99.31
2014-02-07 15:29:20 +00:00
04c776e5c8
Change vnode operations create, mknod, mkdir and symlink to return
...
the resulting vnode *vpp unlocked.
Discussed on tech-kern@
Welcome to 6.99.30
2014-01-23 10:13:55 +00:00
1139274440
Change vnode operations create, mknod, mkdir and symlink to keep the
...
directory node dvp locked on return.
Discussed on tech-kern@
Welcome to 6.99.29
2014-01-17 10:55:01 +00:00
11beb836b0
Replace llx with PRIx64 and some llu with PRIu64 in format strings.
2014-01-12 17:49:30 +00:00
f664413dc2
Field width is an int, so cast the result of strlen() appropriately.
2014-01-12 17:48:59 +00:00
b3fba11357
Disable the extra-args warning for printf formats in dtrace.c
2014-01-12 17:48:25 +00:00
99de37ccf7
Don't define ptrdiff_t since that's done in the NetBSD <sys/types.h>
2013-12-12 18:20:49 +00:00
b90de0d04a
fix one more mountlist instant
2013-11-25 22:48:05 +00:00
65b1f85ab6
Vnode API cleanup pass 1.
...
- Make these defines and functions private to vfs_vnode.c:
VC_MASK, VC_LOCK, DOCLOSE, VI_IANCTREDO and VI_INACTNOW
vclean() and vrelel()
- Remove the long time unused lwp argument from vrecycle().
- Remove vtryget(), it is responsible for ugly hacks and doesn't
look that effective.
Presented on tech-kern.
Welcome to 6.99.25
2013-10-29 09:53:51 +00:00
419fd4e177
NetBSD does not have vn_in_dnlc
2013-10-21 15:31:26 +00:00
e1fe9cc144
use proper format strings.
2013-10-19 23:07:39 +00:00
50a36d05ad
don't use IO_APPEND. In the vdev_file.c we end up appending to our virtual
...
file as a filesystem! With this change:
$ dd if=/dev/zero of=disk1 count=266144
$ zpool create crap $PWD/disk1
mounts /crap and it seems to work ok, until I try unmounting
2013-09-23 20:44:24 +00:00
4f1b5d5104
not nice to play with resid before checking for error, but this is how they
...
fixed it, and I am following suit. I would have preferred if (error && resid)
instead.
2013-09-23 20:41:19 +00:00
618c231990
Hack: Instead of dying in vrele with a kernel assertion, detect that the
...
usecount of the vnode is 0 and treat it as a noop. This is wrong, buf it
get us through the VN_RELE() after the close in vdev_file.c
2013-09-23 19:44:21 +00:00
54639fcdc3
Don't forget to destroy the mutex we created to the zfsvfs transient object,
...
otherwise if we load and unload the module the lockdebug code will blow chunks.
2013-09-23 19:42:42 +00:00
975a152cfc
If a library needs a symbol from another library, pull that library in
...
explicitly, even if the DT_NEEDED closure would normally already ensure
the presence.
2013-09-11 23:04:09 +00:00
b964869221
-D and -I are CPPFLAGS
2013-09-10 11:23:20 +00:00
22898e3a43
Make dtrace/zfs compile on arm based on the diff from:
...
http://people.freebsd.org/~gonzo/arm/dtrace-arm.diff
Completely untested, kernel support not added.
2013-06-21 19:16:00 +00:00
7733018eca
rename delay to xdelay to avoid conflicts.
2013-06-21 16:22:45 +00:00
ea65093059
stub out for now
2013-04-14 16:52:47 +00:00
1dd1e91a24
add a stub for arm
2013-04-14 16:39:59 +00:00
213aef742a
let zfs autoload too.
2013-03-04 01:14:35 +00:00
db9ff9a846
print the function name that was not found.
2013-03-03 18:18:13 +00:00
c43c276ced
autoload tracing modules.
2013-03-03 18:16:35 +00:00
9776ad40b3
need to be a DRIVER class module to be able to autoload when /dev/dtrace/dtrace
...
is opened.
2013-03-03 16:56:15 +00:00
53529c743e
Fix dodgy arguments to snprintf(3) which cause a warning that breaks
...
the build (at least with "USE_SSP" set to "yes").
2013-02-20 22:45:12 +00:00
55ad42a55f
Avoid comparison between pointer and integer.
2013-01-26 20:15:50 +00:00
3c7ce2aac6
Put optional args before non-optional ones.
2013-01-20 23:51:40 +00:00
519dc5efd2
Sort options in SYNOPSIS and text.
...
XXX: please check if SYNOPSIS are correct, I think they missing some .Op.
2013-01-19 01:22:12 +00:00
d3d2c27409
add man pages
2013-01-18 19:58:43 +00:00
a83c8cf9a9
Handle enum forward declarations.
2013-01-18 16:23:48 +00:00
a34e002f26
Spell cprng_fast() properly
2013-01-03 17:12:05 +00:00
278835249a
Use cprgn_fast() not rnd_extract_data().
...
The latter isn't in any header files any more.
2013-01-03 16:42:55 +00:00
51f5533426
Compilation fixes:
...
- include sys/ioctl.h
- NULL isn't an int constant
- %llf isn't a printf format
2013-01-03 16:41:51 +00:00
888f2d7310
Include "sys/ioctl.h" for the prototype of ioctl(2) which fixes the build.
2013-01-02 21:19:13 +00:00
6c6b743ee5
I've NFI why this code has userspace stuff playing with 'struct proc *'
...
nor why it uses &p0 as a magic constant (rather than NULL).
Re-instate the definition of p0, but enable the 'fake' definition
of 'struct proc' if the relevant part of sys/proc.h seems to have
been #if'ed away.
Should fix the build.
2013-01-02 16:31:24 +00:00
be6d96d7d9
Explicitly include sys/ioctl.h
2013-01-02 10:33:49 +00:00
07465819ee
Replace alloca(256) with an array.
...
The code ended up using alloca() from libc - which 'just can't work' (tm).
It is also pointless!
2013-01-02 10:33:19 +00:00
b4041ccef0
Comment out the unused 'struct proc p0;'
...
This is userspace, it isn't supposed to know what a 'struct proc' is.
2013-01-01 19:03:26 +00:00
f337b6c74a
This needs an explicit include of sys/ioctl.h.
...
XXX: There is some crappy code in here :-)
2013-01-01 19:02:19 +00:00
8f4ea3c47b
adapt the profile dtrace provider to netbsd.
2012-12-07 03:11:17 +00:00
11c69f2d20
adapt the cyclic module and profile dtrace provider to netbsd.
...
for now, just hook the cyclic callback into hardclock().
2012-12-02 01:05:16 +00:00
a32db86a29
update cyclic module to the freebsd 8-stable version as of svn r219520.
2012-12-02 00:05:38 +00:00
7c1cb982d4
we have _SC_PHYS_PAGES now
2012-11-04 19:12:41 +00:00
d9310ec72a
Use 0 for MS_NOMNTTAB since we have no mnttab anyway.
2012-10-20 22:11:38 +00:00
fdd8614af9
Hack to avoid missing macro to unbuild the tree.
2012-10-20 11:37:37 +00:00
0041a2c6fe
Let's try that again without the static...
2012-10-19 22:19:15 +00:00
e75d983bfe
Add omitted secpolicy_vnode_utime_modify, for zfs.
2012-10-19 19:58:33 +00:00
4797828821
Back out accidental commit of errno kludge for rmdir(".") &c.
...
Solaris returns EEXIST, whereas we want to return ENOTEMPTY (POSIX
allows both), but this got included in an unrelated commit and should
be separated into a common commit for other related error code fixes.
2012-10-18 14:29:44 +00:00
2111fc7126
Take a first whack at making zfs permissions work.
...
zfs_access uses secpolicy_vnode_access, so it makes no sense for the
latter to call VOP_ACCESS!
Everything seems to return EACCES instead of EPERM, probably because
that's what kauth returns. This should be fixed, but that may
require some nontrivial surgery to zfs's calls to secpolicy_*, which
is where kauth gets involved.
This commit imports some code from illumos to implement the routine
secpolicy_vnode_setattr. This shouldn't be outside dist/, but for
now it is expedient to do so. We ought to fix that, along with all
the other CDDL code outside dist/, when we next import a newer
version of zfs.
2012-10-18 14:22:57 +00:00
8e972045d7
Restore _PC_NO_TRUNC in zfs_netbsd_pathconf.
...
Accidentally clobbered it in a previous commit while moving patches
around between source trees -- oops.
2012-10-16 00:04:15 +00:00
6d6cb07bff
Fail lookup of .. in a deleted directory with ENOENT.
...
Don't try to zget the parent, whose znode id may have been recycled
by now.
It's not clear to me how Solaris avoids this, but maybe I'm just
missing something obvious.
2012-10-15 23:51:11 +00:00
1baf22d09c
Fix various issues in zfs life cycle, locking, and vop protocol.
...
- Restore some zfs locking and unlocking that got lost randomly.
- Enable use of the BSD vnode lock. Lock order: all BSD vnode locks
are taken before all zfs internal locks. There remains an issue with
O_EXCL, to be solved later (famous last words). KASSERT the locking
scheme up the wazoo.
- Take our cruft out of zfs_lookup and move it to zfs_netbsd_lookup.
Restore much of the way zfs_lookup looked to make merging future
versions easier. Disable use of the namecache for now because its
locking dance is too scary to contemplate.
- Implement BSD semantics for rename, to appease our tests. This is
a provisional kludge; eventually we need VOP_RENAME to take a flag
specifying whether to use BSD semantics or POSIX semantics.
- Simplify zfs_netbsd_reclaim and make it work. Now that getnewvnode
never tries to vclean anything itself, we need not worry about
recursion of ZFS_OBJ_MUTEX locks.
- Clarify and fix genfs node initialization and destruction.
zfs passes most of our atf vfs tests now, including the rename races.
Still to do:
- fix the impedance mismatch between our permissions model and zfs's;
- fix O_EXCL (nontrivial);
- throw dirconc at it and see how badly it explodes;
- find why zpool sometimes wedges itself during mkfs; and
- find why pool caches sometimes seem to get corrupted.
2012-10-15 23:08:19 +00:00
1b44cf7c91
secpolicy_vnode_access must be called with the vnode lock held.
...
Don't unlock the vnode and then tell the caller about what the world
was like while the vnode was locked. The world changes fast.
2012-10-15 22:50:25 +00:00
2ea30100de
Simplify zfs dirlock reference counting.
...
No need to pass dzp around; dl has a pointer to it.
2012-10-15 22:43:50 +00:00
b2ba8fafaa
Do reference counting for zfs range lock waiters.
...
Avoid cv_broadcast(&cv); cv_destroy(&cv); which works in Solaris only
by abuse of the condvar abstraction.
There are parts of this code that should be factored into smaller
subroutines, mainly range lock allocation and initialization, but
that would make it harder to merge newer versions of zfs, so for now
I've just expanded those parts further in-line.
2012-10-15 14:15:59 +00:00
f1e3330bc7
Do reference counting for zfs dirlock waiters.
...
Solaris relies on cv_broadcast(&cv); cv_destroy(&cv) working, but
that hoses our cv_wait, which needs to continue using cv after it is
woken. Solaris's idiom is an abuse of the condvar abstraction, but
we can get the same effect with reference counting.
2012-10-15 14:03:06 +00:00
172b16130f
No need to include files more than once.
2012-10-11 08:01:23 +00:00
6b1609c666
Add _PC_NO_TRUNC support in zfs_netbsd_pathconf().
2012-10-01 18:19:18 +00:00
91aa9460ce
Don't build with debugging (-g) by default.
2012-09-05 23:08:42 +00:00
fa4bf43b79
For now, make vop_islocked more like vop_lock and vop_unlock (i.e. don't
...
even attempt to do anything).
2012-09-01 16:19:00 +00:00
ffd8412325
PR/46660: Henning Petersen: Fix typo in deviceid.c
2012-07-05 13:55:58 +00:00
b813bd15cb
Use CTLTYPE_QUAD for uint64_t type in sysctl_createv.
2012-06-19 21:25:26 +00:00
22738551d6
reduce diff to freebsd:
...
rename "xcpu" back to "cpu" now that the conflicting global variable
has been renamed out of the way. this also fixes some cases
where references to the local variable "cpu" had not been renamed
and thus were accidentally referring to the former global "cpu".
2012-06-16 17:31:47 +00:00
6e9ead4606
make dtrace work on amd64.
2012-06-11 15:18:05 +00:00
57d7988f76
Now that pool_cache_invalidate() is synchronous and can handle per-CPU
...
caches, merge together pool_drain_start() and pool_drain_end() into
bool pool_drain(struct pool **ppp);
"bool" value indicates whether reclaiming was fully done (true) or not (false)
"ppp" will contain a pointer to the pool that was drained (optional).
See http://mail-index.netbsd.org/tech-kern/2012/06/04/msg013287.html
2012-06-05 22:51:47 +00:00
f7622c11cc
Not sure why this has just started to fail!
...
But it really isn't a good idea to compare a uint64_t to NULL.
2012-06-03 10:53:51 +00:00
27e3db30d5
Add FREAD for the benefit of kauth!
2012-04-26 23:34:09 +00:00
2b8cc88a65
Use the correct path in dt_module.c.
...
From clang via Henning Petersen in PR 46009.
2012-02-13 13:05:36 +00:00
6ef3ca5393
Use the correct MACHINE_ARCH for amd64.
...
Remove temporary hack while here.
From Henning Petersen in PR 46001.
2012-02-13 12:52:54 +00:00
e62ee4d475
extending vmem(9) to be able to allocated resources for it's own needs.
...
simplifying uvm_map handling (no special kernel entries anymore no relocking)
make malloc(9) a thin wrapper around kmem(9)
(with private interface for interrupt safety reasons)
releng@ acknowledged
2012-01-27 19:48:38 +00:00
9a00affba3
Fix a segfault in ctfmerge.
...
GCC can generate bogus dwarf attributes with DW_AT_byte_size set to 0xFFFFFFFF.
See http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35998 .
GCC is currently doing this for external/bsd/tmux/dist/compat/imsg-buffer.c:
readelf -a --debug-dump imsg-buffer.o
...
<2><6e3>: Abbrev Number: 32 (DW_TAG_union_type)
<6e4> DW_AT_byte_size : 0xffffffff
<6e8> DW_AT_decl_file : 1
<6e9> DW_AT_decl_line : 229
<6ea> DW_AT_sibling : <0x705>
This resulted in ctfconvert generating a faulty CTF entry which then caused the
segfault in ctfmerge.
The fix has ctfconvert check for the bogus 0xFFFFFFFF value and works around it.
It also adds some protection to ctfmerge to avoid the segfault and fail
more gracefully if the error should occur in the future.
2012-01-10 08:42:22 +00:00
dd89f66f56
make this compile again. VOP_CLOSE is really vn_close() through macros.
2011-11-21 17:51:03 +00:00
bde7230620
correct declaration of vn_close()
...
XXX: Should all those dup declarations be here?
2011-11-21 05:50:11 +00:00
6a2b307489
close file in a way that it does not trigger an assertion :-)
...
now zpool create works again.
2011-11-21 00:24:24 +00:00
eb50a1b80e
enough with the compile time warnings.
2011-11-20 02:54:25 +00:00
fb69e65229
make aggregate work for > 1 cpus
2011-10-19 11:03:52 +00:00
f34e2253c6
because our xcall stuff can't be used in softint, use kernel threads
...
for cleaner/deadman low-level cyclic replacement. PR/44905
2011-10-19 11:02:08 +00:00
ad773a12d3
fbt_provide_module_cb: exclude some functions known unsafe to probe
2011-10-19 10:55:50 +00:00
34f54c83be
As vnalloc() always allocates with PR_WAITOK there is no longer the need
...
to test its result for NULL.
2011-10-07 09:35:04 +00:00
b6f6e1bfbd
fix a shadow variable warning.
2011-09-09 17:48:39 +00:00
a222eb155b
catch up with vmem_alloc signature change.
2011-09-07 21:32:07 +00:00
4aa752a401
fix shadow warnings.
...
XXX: might be a good idea to kill global 'cpu' and 'version' variables.
2011-08-31 21:57:16 +00:00
eab2487254
Fix a local var name.
2011-07-30 10:12:14 +00:00
6ed9c23d3b
KNF.
2011-07-29 13:10:23 +00:00
e3bb2d6eba
mark this driver D_MPSAFE
2011-07-26 12:54:28 +00:00
afe833af51
fix a merge botch in rev.1.7
2011-07-26 12:33:23 +00:00
696f2a9301
remove debug printfs
2011-07-26 12:28:39 +00:00
7695ef3cd1
remove a debug printf
2011-07-26 12:28:01 +00:00
223b285dcf
fix a kmem_alloc/free size mismatch
2011-07-26 12:26:33 +00:00
71477abcdc
Fix uninitialized variables reported by GCC 4.5.
2011-07-20 19:51:57 +00:00
9b1cdf6a14
better fix for gcc uninitialized.
2011-07-18 02:47:52 +00:00
6e03a05a94
fix gcc uninitialized warnings
2011-07-18 00:43:35 +00:00
6c10492aaa
fix uninitialized gcc variables
2011-07-18 00:42:40 +00:00
3eb244d801
Retire varargs.h support. Move machine/stdarg.h logic into MI
...
sys/stdarg.h and expect compiler to provide proper builtins, defaulting
to the GCC interface. lint still has a special fallback.
Reduce abuse of _BSD_VA_LIST_ by defining __va_list by default and
derive va_list as required by standards.
2011-07-17 20:54:30 +00:00
1ab522108e
Fix ZFS build.
2011-06-12 14:47:20 +00:00
b156c02e66
v_interlock is now a pointer to kmutex_t, use it as one.
...
fixes many of the 5.99.53 build issues in zfs, but not all of them.
2011-06-12 04:43:11 +00:00
4e14e91053
fix missing prototypes.
2011-05-15 18:33:33 +00:00
756122ce22
Remove unused M_MOUNT.
2011-04-02 04:57:35 +00:00
e2e40b270e
Don't define acl_alloc and acl_free during kernel builds.
...
Fixes module build with clang.
2011-03-15 00:48:04 +00:00
eb63257707
adjust some pretty likely broken assertion (stksize is hardcoded
...
from caller, but not used here). now zpool doesn't cause a
kernel crash anymore.
2011-03-10 19:35:24 +00:00
3140523bc0
pass in correct variable, the old code can't possibly be correct.
2011-03-10 19:32:39 +00:00
3406e189a5
#include <sys/debug.h> (from the ZFS sources) so that ASSERT() is
...
defined, and we don't get problems like kern/44430. This is
more a workaround than a fix, but at least I can load zfs now.
ok haad@
2011-03-09 23:55:46 +00:00
c29d7160e5
If the config is zero size, don't try to read it.
...
OK haad@
2011-03-09 23:49:06 +00:00
1c44cbfd96
Don't allow module to init if root is not mounted since the modcmd
...
wants to traverse the file system. This *might* fix statically
linking zfs support into the kernel (at least it fixes rump kernel
bootstrap).
2011-01-21 13:08:42 +00:00
51f08b56c8
fix kobj_open_file_vnode()
...
patch from haad
2011-01-21 12:24:24 +00:00
c7287ec04b
Fix zfs Makefiles + debug, don't add -g by default only when MKDEBUG is
...
enabled. Fix header assert conflicts + bump WARNS for some libs.
2011-01-13 23:18:23 +00:00
42bae07fce
Remove references to SAVESTART in zfs.
2011-01-02 05:12:03 +00:00
bec5c61e50
Define vn_getattr in zfs_context.h.
2010-12-28 16:44:45 +00:00
bca088ab45
Fix problem where vnode vp_size was not updated when vndoe was already allocated,
...
for vdev. This makes ztest survive ztest_vdev_LUN_growth test. Replace dummy
VOP_GETATTR with vn_getattr routine which reset vp_size and vattr_size accordingly
to reality.
2010-12-28 13:36:09 +00:00
8f7b707f59
libzfs_fru_clear was not ported to NetBSD disable it foor now.
2010-12-24 16:13:58 +00:00
363aa1ba89
Whitespace police.
2010-12-15 22:49:42 +00:00
59385470cf
Only KM_SLEEP | KM_NOSLEEP can be passed down to kmem layer, clean up
...
tqflags from internal taskq flags (TQ_FRONT) before calling kmem_alloc.
2010-12-15 22:09:14 +00:00
riastradh