Commit Graph

529 Commits

Author SHA1 Message Date
Marcel Moolenaar
e2fb7a7be0 Update for the KDB framework:
o  Call kdb_enter() instead of Debugger().
2004-07-10 21:21:13 +00:00
Lukas Ertl
3202f019e7 Catch up with g_dev_open() changes and pass over the current thread
pointer.

Submitted by:    cperciva
2004-06-22 06:38:01 +00:00
Poul-Henning Kamp
f3732fd15b Second half of the dev_t cleanup.
The big lines are:
	NODEV -> NULL
	NOUDEV -> NODEV
	udev_t -> dev_t
	udev2dev() -> findcdev()

Various minor adjustments including handling of userland access to kernel
space struct cdev etc.
2004-06-17 17:16:53 +00:00
Poul-Henning Kamp
89c9c53da0 Do the dreaded s/dev_t/struct cdev */
Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
Lukas Ertl
4e72cdbdc6 Plug a mem leak in vinum_scandisk().
Approved by:    grog (mentor)
2004-04-24 23:41:21 +00:00
Lukas Ertl
2c1305420f When doing round-robin reads from a multi-plex volume, only switch to the
next plex if the sector to be read isn't nearby the last read sector.

Submitted by:  Vsevolod Lobko <seva@ip.net.ua> via ru@
Approved by:   grog (mentor)
2004-03-19 10:28:34 +00:00
Lukas Ertl
5faeb9c682 Properly count references of our dev_t to avoid triggering a KASSERT in
dev_strategy().

Submitted by:   dwmalone
Approved by:    grog (mentor)
2004-03-11 14:11:08 +00:00
Lukas Ertl
6c8740f899 Fix an integer overflow when dealing with very large volumes. This bug
prevented newfs to work on volumes that are larger than 1TB.

PR:             63577
Submitted by:   Masaki Takakashi <mtakahashi@se.gtd.cosmo.co.jp>
Approved by:    grog (mentor), bde
2004-03-09 12:45:43 +00:00
Lukas Ertl
664e22ad1e Since vinum doesn't fake disklabels anymore, remove get_volume_label().
Also, remove stale write_volume_label() declaration; the write_volume_label()
function was deleted 8 months ago.

Approved by:    grog (mentor)
2004-03-09 09:50:15 +00:00
Poul-Henning Kamp
6fe9c1609f Fix vinums cdevsw{} to initialize d_version.
The nonstandard formatting made my mega-patch scripts miss it.

Retire the static major number while we're here anyway.

Reported by:	Niels Chr. Bank-Pedersen <ncbp@bank-pedersen.dk>
2004-02-23 08:55:13 +00:00
Lukas Ertl
ce7565b430 Remove some more 'makedev' related macros.
Approved by:     grog (mentor)
2004-02-16 23:14:13 +00:00
Lukas Ertl
60b42b1c00 Don't free ressources that haven't been allocated. This should fix
the "disappearing subdisks" problem when new subdisks can't be created
due to some errors.

This is in fact an ugly hack, but a more elegant solution would probably
require a redesign of vinum in several places.

Approved by: joerg (mentor)
2004-02-08 15:27:58 +00:00
Lukas Ertl
3a1ab63eb7 When creating raid5 or striped plexes, avoid falling out of bounds
when checking the given stripe size.

Also move the code a bit around to avoid duplication.

Approved by: joerg (mentor)
2004-02-04 22:29:52 +00:00
Greg Lehey
458252a742 initsd: For striped and RAID-[45] plexes, don't restrict the I/O
transfer size to the stripe size.  This is a different
        situation from reviving, where this limitation is necessary.
        In initsd we're simply writing binary zeroes to the entire
        disk, so the only effect of limiting the transfer is to slow
        things down.
2003-12-15 00:45:53 +00:00
Greg Lehey
80d23b9669 free_plex, free_volume: Don't try to destroy the underlying device if
it doesn't exist.  This can occur under
                        certain failure situations.
2003-12-15 00:44:05 +00:00
Greg Lehey
684873a192 vinumclose: Fix day-one bug. Subdisks and plexes were not being
marked closed.  This made it impossible to stop Vinum
	    after explicitly opening a plex or subdisk.
2003-12-15 00:42:03 +00:00
Greg Lehey
76ffeff6bd Initalize buffer headers correctly for current version of bio subsystem.
Not mentioned by: phk
Approved by:	  re (scottl)

This makes Vinum work again, at least in my testing.
2003-11-24 04:06:56 +00:00
Scott Long
8427f795da Fix vinum by correctly setting the b_offset bio filed when doing I/O.
Submitted by:	grog
2003-11-23 17:10:58 +00:00
Poul-Henning Kamp
06a259faf6 Do not initialize bp->b_pblkno, it is going away. 2003-10-18 17:57:48 +00:00
Greg Lehey
be0dabc1bc If we fail to complete configuring a subdisk because of some error,
make sure we return any allocated space to the drive.  This should get
rid of a number of inconsistencies (hopefully all) that have been seen
after configuration errors.
2003-09-29 08:50:03 +00:00
Greg Lehey
692db2e5c9 Revert last commit. It built the kernel module just fine, but broke
world in /sbin.

Reported by:	erwin
Pointy hat to:	grog
2003-09-29 08:19:06 +00:00
Greg Lehey
2b0761b360 Remove calls to makedev. They were never needed.
Requested by:	phk
Explained by:	rwatson
2003-09-29 03:16:20 +00:00
David E. O'Brien
aad970f1fe Use __FBSDID().
Also some minor style cleanups.
2003-08-24 17:55:58 +00:00
Robert Watson
4ac9fbcf79 #ifdef __i386__ the "int slice" definition that became i386-only with
recent Vinum changes.  Fixes non-i386 build on vinumio.c.

Accidental oversight of:	grog
Fix hopefully not minded by:	grog
2003-08-15 17:56:44 +00:00
Greg Lehey
d4bd229fa5 vinum_scandisk:
Correctly handle additional disks without BIOS partition tables.
  Previously, vinum_scandisk stopped scanning additional disks for
  native partitions after any good partition was found.  This applies
  to all platforms, but was a particular problem on systems without
  BIOS partition tables.

Submitted by: harti
2003-08-15 04:36:53 +00:00
Greg Lehey
5466a9cb4c Drop Giant when calling the disk drivers directly, and reacquire
afterwards.  This fixes the Vinum breakage in -CURRENT.

Breakage explained by: phk
2003-08-15 03:57:15 +00:00
Greg Lehey
8517ec8be1 free_drive: Free the drive even if it's referenced. I don't know what
I was smoking when I wrote this stuff, but another fix resulted in
every partition in the system being entered as a "referenced" drive.
2003-07-01 07:53:54 +00:00
Hartmut Brandt
1232acdba3 Check compatibility partitions on all archs. This makes vinum
work on sparc64 again, because on sparc i386-compatibility
partitions just happen to have the same name as sparc native partitions.

Okay-ed by: grog@
2003-06-23 14:49:57 +00:00
Greg Lehey
be05126e24 check_drive: If the partition isn't a Vinum drive, release it again
and return NULL.

vinum_scandisk: Don't handle NULL device pointers.
	        Only look at compatibility partition for i386.  This
	        is a kludge which should go away once I have adequate
	        documentation for the New World Order.

Together, these fixes remove occasional error messages about
non-existent drives.  They may also fix a number of problems that have
been reported without a PR.

PRs: None
2003-06-15 01:42:01 +00:00
Greg Lehey
ed0a525e09 Silence a warning. 2003-06-12 06:17:12 +00:00
Greg Lehey
de52a6f7a6 Remove "raw" and "label" keywords. 2003-06-12 05:45:45 +00:00
Greg Lehey
bfe670eec4 vinum_bounds_check: Don't check the label sector if the architecture
doesn't have one.  The test was bogus on these architectures, but
recent changes broke it altogether.

Prompted by:  phk

This should fix the recent SPARC 64 build problems.
2003-06-12 05:34:42 +00:00
Greg Lehey
98142e711a Remove VINUM_LABEL ioctl. It has been deprecated since December 1998,
and recent changes broke it altogether.

Prompted by:  phk
2003-06-12 05:32:31 +00:00
Greg Lehey
00ee3bf535 Remove write_volume_label, the implementation of VINUM_LABEL ioctl.
It has been deprecated since December 1998, and recent changes broke
it altogether.

Prompted by:  phk
2003-06-12 05:32:00 +00:00
Greg Lehey
295cb8611f Remove kw_raw. I've forgotten what it was intended for, and it was
probably obsoleted by some changes a couple of months ago.
2003-06-12 05:28:18 +00:00
Greg Lehey
14b1df077e Don't deregister the kld on resetconfig, only when we're really ready
to unload.  This would cause a panic on the second resetconfig.

Start Vinum at boot time at SI_SUB_RAID, not SI_SUB_VINUM.
SI_SUB_VINUM was there first, but there's no real distinction, and
SI_SUB_RAID is a more neutral name.

Submitted by:  hmp
2003-06-03 04:51:31 +00:00
Greg Lehey
74f2cc2c9c Change the way the plex lock mutexes work. Previously they were part
of the struct plex, which tore apart the mutex linked lists when the
plex table was expanded.  Now we maintain a pool of mutexes (currently
32) to be shared by all plexes.  This is still a lot better than the
splhigh() method used in other architectures.

expand_table: Add parameters file and line if we're debugging.

Approved by: re (jhb)
2003-05-23 01:15:55 +00:00
Greg Lehey
93573e2e76 Change the way the plex lock mutexes work. Previously they were part
of the struct plex, which tore apart the mutex linked lists when the
plex table was expanded.  Now we maintain a pool of mutexes (currently
32) to be shared by all plexes.  This is still a lot better than the
splhigh() method used in other architectures.

Add and clarify comments.

Approved by: re (jhb)
2003-05-23 01:15:30 +00:00
Greg Lehey
7db14b2ff2 expand_table: Add parameters file and line if we're debugging.
MMalloc, vinum_meminfo: Use strlcpy to copy file name.

Approved by: re (jhb)
2003-05-23 01:15:01 +00:00
Greg Lehey
d026346c86 Change the way the plex lock mutexes work. Previously they were part
of the struct plex, which tore apart the mutex linked lists when the
plex table was expanded.  Now we maintain a pool of mutexes (currently
32) to be shared by all plexes.  This is still a lot better than the
splhigh() method used in other architectures.

Approved by: re (jhb)
2003-05-23 01:14:35 +00:00
Greg Lehey
8a697ff435 detachobject: Update volume config after detaching a plex.
update_volume_config: Remove redundant diskconfig parameter.

Approved by: re (jhb)
2003-05-23 01:14:13 +00:00
Greg Lehey
cb5eba5e09 Change the way the plex lock mutexes work. Previously they were part
of the struct plex, which tore apart the mutex linked lists when the
plex table was expanded.  Now we maintain a pool of mutexes (currently
32) to be shared by all plexes.  This is still a lot better than the
splhigh() method used in other architectures.

update_volume_config: Remove redundant diskconfig parameter.

expand_table: Add parameters file and line if we're debugging.

Approved by: re (jhb)
2003-05-23 01:13:43 +00:00
Greg Lehey
f7b76dc815 Change many strcpys to strlcpys, etc.
Submitted by:	   Ted Unangst <tedu@stanford.edu>

Correct some inaccurate and badly formatted comments.

config_subdisk: If our drive is down, ensure that the subdisk is
		crashed.  Previously it was possible for the subdisk
		to be up when the drive was down.

Change the way the plex lock mutexes work.  Previously they were part
of the struct plex, which tore apart the mutex linked lists when the
plex table was expanded.  Now we maintain a pool of mutexes (currently
32) to be shared by all plexes.  This is still a lot better than the
splhigh() method used in other architectures.

update_volume_config: Remove redundant diskconfig parameter.

Approved by: re (jhb)
2003-05-23 01:13:10 +00:00
Greg Lehey
4555a3de62 print_config:
Change config format slightly to save plex preferences correctly.

vinum_scandisk: reinitialise volatile pointer after function call.
This is the "deafc0de" bug.

Approved by: re (scottl)
2003-05-19 02:21:31 +00:00
Greg Lehey
108b696afe ioctl VINUM_READCONFIG: Don't lock configuration here. vinum_scandisk
needs to do it anyway to handle the startup case.  This is
            part of a fix for the recently reported hangs.

Approved by:  re (scottl)
2003-05-08 00:36:20 +00:00
David E. O'Brien
2875867356 Fix usages of %ll[dx] with typedef'ed created types.
In the kernel it is wrong 99.9 times out of 100 to use %ll rather than cast
to intmax_t and use %j.
2003-05-05 16:56:44 +00:00
Greg Lehey
3cff2a7306 Add sector size member to subdisk, plex and volume objects. 2003-05-05 08:43:15 +00:00
Greg Lehey
7f55835e9c Remove an accidental #include.
Maintain sector sizes for all objects, not just for drives.  Some of
this could do with improvement: in particular, we get an error if the
components of an object have different sector sizes.

Clean up some comments.
2003-05-05 08:42:42 +00:00
Greg Lehey
63deaeffac Implement DIOCGSECTORSIZE and DIOCGMEDIASIZE ioctls. It appears that
they are now necessary to use newfs(8).
2003-05-05 08:41:53 +00:00
Greg Lehey
2154a1eed3 Correct failed checkin. Only one of the hunks in the last checkin
seems to have made it.
2003-05-05 07:50:15 +00:00