d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
108,187 Commits 4 Branches 49 Tags
Commit Graph

834 Commits

Author SHA1 Message Date
le
ada7c51edc Make it possible to rebuild degraded RAID5 plexes. Note that it is 
currently not possible to do this while the volume is mounted.

MFC in:  1 week
 2004-09-30 12:57:35 +00:00
phk
dd7446483c Protect the start/end counts on consumers and providers with the up/down 
mutexes.

Make it possible to also protect the disk statistics (at a minor cost in
performance) by setting bit 2 of kern.geom.collectstats.
 2004-09-28 11:56:37 +00:00
pjd
5b9c332a3c - Set maximum request size to MAXPHYS (128kB), instead of DFLPHYS (64kB). 
- Set minimum request size to sectorsize, instead of 512 bytes.

Approved by:	phk (some time ago)
 2004-09-28 08:34:27 +00:00
pjd
63dd0f756b Just use MAXPHYS as maximum I/O request size, instead of using my own 
#define for this purpose.
No functional change.
 2004-09-28 07:33:37 +00:00
pjd
ef6747fa18 Decrease kern.geom.raid3.timeout to 4, so it is smaller than 
vfs.root.mountdelay by default.
 2004-09-27 22:12:14 +00:00
pjd
25b7bcde0d Deny invalid I/O requests which comes from userland here, because later 
we'll get a panic.
MT5 candidate.

Reviewed by:	phk
 2004-09-27 22:10:01 +00:00
pjd
34196975cf Avoid race while synchronizing components. It is very hard to bump into, 
but it is possible:
1. Read data from good component for synchronization.
2. Write data to the same area.
3. Write synchronization data, which are now stale.

Found by:	tegge (for gmirror)
 2004-09-27 20:32:35 +00:00
pjd
cad6af1c8f Minor, but very important condition fix. The current one can never be true. 2004-09-27 19:32:26 +00:00
pjd
4081d9d5d1 Decrease kern.geom.mirror.timeout to 4, so it is smaller than 
vfs.root.mountdelay by default.
 2004-09-27 13:47:37 +00:00
pjd
9b6a1c588a Forgot to commit addition of ds_resync field. 2004-09-26 20:42:35 +00:00
pjd
9871848b34 Avoid race while synchronizing components. It is very hard to bump into, 
but it is possible:
1. Read data from good component for synchronization.
2. Write data to the same area.
3. Write synchronization data, which are now stale.

Found by:	tegge
 2004-09-26 20:41:07 +00:00
pjd
48183ebbc3 Simplify code a bit. 2004-09-26 20:30:15 +00:00
phk
54ba36ef3e Assert topology is held in g_dev_getprovider(). 
Don't call devsw().  It is not necessary, and we do not need to hold dev_lock
to compare the devsw pointer to our own since we do not dereference it.
 2004-09-24 06:43:20 +00:00
pjd
d7954bf77f This is not needed anymore, it is forced in GEOM now. 
Actually, it can even cause some problems, because GEOM requires sectorsize
to be more than 0 on first access, not on provider creation, so we can skip
valid providers by doing this check here.

Reported by:	Divacky Roman <xdivac02@stud.fit.vutbr.cz>
		Sven Willenberger <sven@dmv.com>
 2004-09-20 17:26:25 +00:00
fjoe
2a220bb5b0 Use correct malloc type when freeing memory allocated by g_read_data. 
PR:		71431
Submitted by:	daichi
 2004-09-19 10:27:46 +00:00
le
5dae52591b Single concat or striped plexes don't need no special initialization 
if their subdisks are all available, so let them be brought up.
 2004-09-18 18:03:20 +00:00
le
18ba8315a7 Re-vamp how I/O is handled in volumes and plexes. 
Analogous to the drive level, give each volume and plex a worker thread
that picks up and processes incoming and completed BIOs.

This should fix the data corruption issues that have come up a few
weeks ago and improve performance, especially of RAID5 plexes.

The volume level needs a little work, though.
 2004-09-18 13:44:43 +00:00
fjoe
76d2723d97 g_nop_create: destroy newly created provider in case of errors. 2004-09-16 15:28:48 +00:00
le
eb087a4ece Give the DRIVE geom a worker thread that picks up incoming bios, 
sends them down, and takes care of the finished bios.  This makes it
easier to handle I/O errors at drive level.
 2004-09-13 21:01:36 +00:00
le
a4de4268bd Rename gv_kill_thread() to gv_kill_plex_thread(), since there are more 
threads to come.
 2004-09-13 17:44:47 +00:00
le
876ffa30db Save the config back to disk when a drive goes down. 2004-09-13 17:33:52 +00:00
le
b19a8d0082 Read a whole sector instead of GV_HDR_LEN, since a sector might be 
bigger (i.e. on CD-ROMs).
 2004-09-13 17:27:58 +00:00
pjd
a5693287e1 Make kern.geom.debugflags sysctl tunable from /boot/loader.conf. 
It will help to debug problems when booting.

Approved by:	phk
 2004-09-13 14:58:27 +00:00
phk
3e1bf8d8b5 Fix a problem that shows up if less than the full complement of 
lock sectors are defined ("number_of_keys" argument to gbde init being
less than 4 in the default compile).
 2004-09-11 17:58:53 +00:00
phk
9b11b1fb2c Respect that G_BDE_MAXKEYS is a compile time variable. 2004-09-11 17:57:51 +00:00
fjoe
7531fcdf4d Do not compile in zlib.c. Add a dependency on module instead. 2004-09-08 17:27:31 +00:00
pjd
555e9e698d Show current status of mirror device directly. 
Suggested by:	Krzysztof Ciep³ucha <kris@home.pl>
 2004-09-08 16:37:22 +00:00
phk
38ae102f61 For removable devices without media we set a zero mediasize but a non-zero 
sectorsize in order to avoid a lot of checks around various divisions etc.

Enforce the sectorsize being > 0 with a KASSERT on successful open.

Fix scsi_cd.c to return 2k sectors when no media inserted.
 2004-09-05 21:15:58 +00:00
pjd
4689077c9e Allow to configure debug level from /boot/loader.conf. 2004-08-30 18:50:06 +00:00
phk
02b88adf06 Add more KASSERTS and checks. 2004-08-30 09:33:06 +00:00
pjd
3bedfb04b2 GCC, ehh. 2004-08-29 14:29:30 +00:00
pjd
354291a750 Use sc->sc_mediasize instead of sc->sc_provider->mediasize which contains 
exactly the same value, but is shorter.
 2004-08-28 02:35:43 +00:00
pjd
2d011d4672 Warn the user if we are not going to use whole provider space. 
Requested by:	Michael Handler <handler@grendel.net>
 2004-08-28 02:34:10 +00:00
pjd
f7c5bbc8f2 Don't allow to insert providers, which are too small. 
Reported by:	Michael Handler <handler@grendel.net>
 2004-08-28 02:02:48 +00:00
le
440914a8bb Move config_new_drive() to the correct place and rename it to 
gv_config_new_drive().
 2004-08-27 21:32:18 +00:00
phk
4759b478a2 Introduce g_alloc_bio() as a waiting variant of g_new_bio(). 
Use in places where we can sleep and where we previously failed to check
for a NULL pointer.

MT5 candidate.
 2004-08-27 14:43:11 +00:00
le
0fd3e9eb56 When attaching a consumer from a volume to a plex, check if the 
volume already has a plex attached and adjust the access counts
of the new consumer accordingly.
 2004-08-26 21:04:41 +00:00
pjd
7f46afc9bf Skip providers with not defined sector size. 
Reported by:	kuriyama
 2004-08-26 12:42:47 +00:00
pjd
18cd875885 Log verification errors at level 1. 2004-08-25 19:18:07 +00:00
pjd
a2512dc38e Dump disk number. 2004-08-25 12:14:44 +00:00
pjd
02488a6b3c Allow to set kern.geom.mirror.timeout from /boot/loader.conf. 2004-08-23 20:42:34 +00:00
le
b83d5cd619 Compare the addresses of two RAID5 work packets directly instead 
of the addresses of their related bios when locking one out, since
they could share a bio and this could lead to parity corruption.
 2004-08-23 17:50:18 +00:00
le
a52623fe58 Implement the possibility to remove drives. 2004-08-22 17:07:55 +00:00
pjd
7e2ef21ad9 Implementation of 'verify reading' algorithm, which uses parity data for 
verification of regular data when device is in complete state.
On verification error, EIO error is returned for the bio and sysctl
kern.geom.raid3.stat.parity_mismatch is increased.

Suggested by:	phk
 2004-08-22 16:21:12 +00:00
le
9c58beacac Add forgotten format specifier in a KASSERT and shut up the compiler. 
Submitted by: Gavin Atkinson <gavin.atkinson@ury.york.ac.uk>
 2004-08-22 13:34:24 +00:00
pjd
2acdd880f4 Add version history. 2004-08-21 21:15:03 +00:00
pjd
c3c6740d1a Implement new reading algorithm, which will use parity component for reading 
as well, even if device is in complete state.
I observe 40% of speed-up with this option for random read operations,
but slowdown for sequential reads.
Basically, without this option reading from a RAID3 device built from 5
components (c0-c4) looks like this:

	Request no.	Used components
	1		c0+c1+c2+c3
	2		c0+c1+c2+c3
	3		c0+c1+c2+c3

With the new feature:

	Request no.	Used components
	1		c0+c1+c2+c3
	2		(c1^c2^c3^c4)+c1+c2+c3
	3		c0+(c0^c2^c3^c4)+c2+c3
	4		c0+c1+(c0^c1^c3^c4)+c3
	5		c0+c1+c2+(c0^c1^c2^c4)
	6		c0+c1+c2+c3
	[...]
 2004-08-21 18:11:46 +00:00
le
7f135ded9a A volume can be up if it has a degraded RAID5 plex. 2004-08-19 12:03:27 +00:00
pjd
ad8a5e508d We really don't want to receive spoil event for synchroniztion consumers. 2004-08-18 23:33:37 +00:00
phk
90dd7ea538 Do not override the class provided dumpconf function. 2004-08-18 21:42:08 +00:00