e38b4355e2
Abort any parity re-writes before unconfiguring a RAID set.
...
Partially addresses PR#10182.
2000-05-28 03:00:31 +00:00
228570e110
When setting auto-config/rootable bits, only try to update the
...
component label if the component is actually alive.
2000-05-28 01:03:22 +00:00
1aaf6772f1
- Add a mechanism for obtaining finer-grained 'progress' information
...
regarding reconstructs, copybacks, etc.
- RAID 0 doesn't do copybacks, but don't make raidctl sweat about it.
2000-05-28 00:48:30 +00:00
26b4598736
Components which match on everything but the mod_counter are
...
renegades, and must be handled correctly. In particular, they should
be added to their old auto-config set, but then immediately released.
Failing to do otherwise means that they potentialy end up in a
different (and competing!) RAID set which may auto-configure in place
of the correct one, and cause all sorts of chaos at auto-configure
time.
2000-05-27 20:29:14 +00:00
e9b5b7aa5e
Add a comment to clarify why we actually want the component label to
...
be considered a 'match' even if the mod_counter is wrong.
2000-05-27 19:12:03 +00:00
d2b2ed0970
Checking parity re-write status makes no sense on RAID 0, but let's
...
return something reasonable anyway.
2000-05-27 18:23:27 +00:00
b8322d5c23
Const poison the layout name.
2000-05-23 00:44:38 +00:00
071aed40ac
A foolish consistency; most parts of the kernel use bp->b_data, so
...
change these from bp->b_un.b_addr to bp->b_data, as well. This also
allows us more flexibility to experiment with other data buffer types
hung off of struct buf.
2000-05-19 18:54:22 +00:00
c2e8f67477
Add some missing casts of ioctl arg.
...
Obviously autoconfiguration wouldn't work on big-endian machines.
2000-05-19 04:53:25 +00:00
db11351ced
Make sure the config structure allocated during autoconfiguration actually
...
gets initialized properly. Should fix PR#9816.
2000-04-27 00:57:48 +00:00
5845348e54
Revert last change (the last change was not necessarily desirable, and
...
had some other minor side-affects that were troublesome).
2000-04-17 19:35:12 +00:00
c71398f7a9
Don't allow configuration where the specified SectPerSU may result in IO larger
...
than MAXPHYS being sent to underlying devices. Addresses PR#9868.
Will change if/when MAXPHYS goes bye-bye.
2000-04-16 03:24:26 +00:00
cf0b47bd0e
Minor cleanup of unneeded verbosity.
2000-03-31 02:05:24 +00:00
cd61979606
Print out the 'root' status of a RAID set during autoconfiguration.
2000-03-31 01:20:43 +00:00
8d702c4b54
Nuke 'register' from some function prototypes.
2000-03-30 13:39:07 +00:00
169ac5b3c1
Remove register declarations.
2000-03-30 12:41:09 +00:00
d0c76be757
cleanup function prototypes.
2000-03-27 03:25:17 +00:00
b97e06092d
Shuffle some prototypes to a more appropriate location.
2000-03-27 03:01:33 +00:00
2ee63332b3
Add bits for eventual support of deleteing components and moving
...
hot-spares into the main set.
2000-03-26 22:38:28 +00:00
d9da26f998
Nuke a function prototype that shouldn't be here.
2000-03-25 16:24:57 +00:00
e44151e8d6
Fix up some autoconfigure messages. In particular, don't try printing
...
the value of an uninitialized variable. Noted by Luke Mewburn.
2000-03-22 01:41:41 +00:00
7f288b4e0d
Just return "100% done" when asked about reconstruction progress on
...
a RAID0 set.
2000-03-20 22:59:26 +00:00
95054da1a1
Fix doubled 'the's in comments.
2000-03-13 23:52:25 +00:00
7c57ac56d7
Cleanup a couple of function declarations. No functional changes.
2000-03-07 03:09:47 +00:00
51ffb0924c
Create a new rf_close_component() to handle vnode operations for closing
...
components. Teach rf_UnconfigureVnodes() how to use it, and tell
the copyback and reconstruction code about it too.
2000-03-07 02:59:50 +00:00
ddebc1cad0
Move rf_UnconfigureVnodes to rf_netbsdkintf.c where it will be more at home.
...
Add prototype to rf_kintf.h, and nuke old prototypes from rf_disks.c and
rf_driver.c .
2000-03-07 02:28:05 +00:00
33a55d022b
rf_ protect the component label print routine, and make it more easily
...
available to the rest of RAIDframe.
2000-03-07 02:12:13 +00:00
dbf5d1f7a6
Nuke unused code.
2000-03-07 01:54:29 +00:00
1231bd6d14
Umm... don't allow attempts to configure the same RAID device multiple times.
...
The check was there, but the return() was missing :(
2000-03-04 06:03:21 +00:00
0049933f0f
Reorganize some comments.
2000-03-04 04:22:34 +00:00
194aba4e41
Garbage collect the (already disabled) "random" queue type.
2000-03-04 03:27:13 +00:00
612e2e8f92
Nuke an unneeded #define.
2000-03-04 03:22:17 +00:00
88a8d06a0b
Use the common disk allocation code, rather doing things the old way.
2000-03-03 03:47:17 +00:00
142175b1a1
Minor formatting cleanup.
2000-03-03 03:10:03 +00:00
7b1139520e
More minor cleanup.
2000-03-03 02:04:48 +00:00
87f46374bb
Minor cleanup. No functional changes.
2000-03-03 01:46:36 +00:00
8cde772948
Oops.. Need to note the parity status when we unconfigure.
...
Noted by Aaron J. Grier. (Thanks!)
2000-02-29 16:54:29 +00:00
8c46e74418
Nuke some debugging output that got left in by mistake.
2000-02-27 02:35:33 +00:00
82419f5ae9
Only mark the components as CLEAN if it's the last thing we're doing.
...
Marking them CLEAN earlier is (currently) asking for trouble.
(XXX This could have been done more elegantly.)
2000-02-27 01:50:22 +00:00
d9c47013b2
Defopt 'RAID_AUTOCONFIG'. Adding 'options RAID_AUTOCONFIG' turns on
...
the component auto-detection and auto-configuration of RAID sets.
Also, add "#options RAID_AUTOCONFIG" to the GENERIC config files.
2000-02-26 17:35:33 +00:00
caa3932441
Minor cleanups. Remove extra debugging printouts.
2000-02-25 20:11:00 +00:00
21243f6205
- ignore the mod_counter when checking to see if a component fits with the
...
rest of an autoconfig set.
- minor cleanups
2000-02-25 19:56:56 +00:00
503220a18f
- ignore components whose mod_counter values are too low.
...
- noting the mod_counter for autoconfigured sets would be a win too.
2000-02-25 19:56:32 +00:00
8d63540a8c
When we close autoconfigured components, we need to note that they
...
are no longer in 'autoconfigured' status.
2000-02-25 17:14:18 +00:00
a4f7f062fd
- cleanup raidinit()
...
- cleanup rf_mountroot_hooks()
- various other cleanups.
2000-02-25 02:42:30 +00:00
5716b5d6d1
Be *slightly* more consistent about the use of num and numraid in raidattach().
2000-02-25 02:21:12 +00:00
e4e4483cb7
Fix a (slightly) bogus status message.
2000-02-25 02:06:34 +00:00
bdd659fc7f
Deal with unconfiguring components in the autoconfigure case.
2000-02-24 23:39:21 +00:00
aea97a5a17
Oops... Missed commmitting this last night. Thanks to Matthias Scheler
...
for letting me know.
(Add the DiskQueueSW pointer)
2000-02-24 17:12:10 +00:00
e644abfe94
Make sure we close auto-configured components appropriately when
...
attempting a rebuild-in-place.
2000-02-24 04:39:41 +00:00
c5961cfba3
Garbage collect some unused stuff. Include rf_kintf.h for a
...
function prototype.
2000-02-24 03:52:15 +00:00
851a2cdcd8
- record the pointer to the DiskQueueSW so we can use that when
...
we hot-add a spare.
- use the above info when adding a spare.
2000-02-24 03:48:41 +00:00
9e81bc058e
Now that we have the information available, use the general maxOutstanding
...
value for the RAID set, rather than looking at Queue[0][0] (which may not
exist).
2000-02-24 02:55:05 +00:00
99f070bab8
- mod_counter needs to get incremented more often.
...
- use raid_init_component_labels instead of hand-rolling a component_label.
(missed doing this for the rf_used_spare case when updating component labels).
- fix serial number initialization in raid_init_component_labels
2000-02-24 01:23:05 +00:00
fb13ce76fc
- make sure we note the parity status in the autoconfig case.
...
- also, make sure we grab the serial number from the component labels
when we autoconfig.
2000-02-24 01:22:32 +00:00
d91ecfbcfd
Be more aggressive about updating component labels in the event
...
of a real component failure (or a simulated failure):
- add 'numNewFailures' to keep track of the number of disk failures
since mod_counter was last updated for each component label.
- make sure we call rf_update_component_labels() upon any component failure,
real or simulated.
2000-02-23 03:44:02 +00:00
e7f3dc2ec9
blockSize for the component label is already set correctly here.
2000-02-23 02:11:05 +00:00
0c43bf18f0
- cleanup of more component label code
...
- make current default label values available everywhere
- make sure numBlocks and blockSize in component labels get initialized
for all component labels
- check for component size to be smaller than or equal to the partition size
when autoconfiguring
2000-02-23 02:04:21 +00:00
6aaac32240
Do a better job of (re)initializing the component labels after
...
a reconstruct or a copyback.
2000-02-23 02:03:03 +00:00
d84e4ed019
Make a note of autoconfigured components.
2000-02-23 02:01:55 +00:00
e0ab2f3d0f
Correct a comment.
2000-02-23 00:37:11 +00:00
214822e08c
Note the fact that the first arg of raidinit() may be bogus in
...
the autoconfig case (the value of the first arg never gets used, but...).
2000-02-22 23:13:15 +00:00
e8ff7f6155
Bounds-check the 'last unit' a little more closely.
2000-02-22 16:07:02 +00:00
fb56415023
Add a few comments, and an indicator of whether or not an autoconfig set
...
is 'rootable'.
2000-02-22 03:39:47 +00:00
bd2230fd09
- implement additional component checking before autoconfiguration
...
- make component label setting work again(!!!!)
2000-02-22 03:38:42 +00:00
afccbaa6d7
Umm.. let's set 'numraid' to something useful before we try to use it.
2000-02-21 23:33:45 +00:00
7deba340d7
Some of these function names are bad enough that no one else would probably
...
pick them, but preface them with rf_ anyway.
2000-02-16 01:10:44 +00:00
445591e874
Get recent changes into the tree:
...
- make component_label variables more consistent (==> clabel)
- re-work incorrect component configuration code
- re-work disk configuration code
- cleanup initial configuration of raidPtr info
- add auto-detection of components and RAID sets (Disabled, for now)
- allow / on RAID sets (Disabled, for now)
- rename "config_disk_queue" to "rf_ConfigureDiskQueue" and properly prototype
in rf_diskqueue.h
- protect some headers with #if _KERNEL (XXX this needs to be fixed properly)
and cleanup header formatting.
- expand the component labels (yes, they should be backward/forward compatible)
- other bits and pieces (some function names are still bogus, and will get
changed soon)
2000-02-13 04:53:57 +00:00
edc931efea
Make this compile without warnings on alphas.
...
(Problem noted by Aaron J. Grier)
2000-02-12 16:06:27 +00:00
dc59bc1db3
Update for sys/buf.h/disksort_*() changes.
2000-01-21 23:39:55 +00:00
138d8fef04
Fix typo + cleanup a bit.
2000-01-15 01:57:57 +00:00
ef035f5a7b
Make more 80-column friendly. No functionality changes.
2000-01-14 04:03:52 +00:00
11e019370e
"fix" a serious problem with RF_AccTraceEntry_t tracerec stuff.
...
XXX the "fix" is incorrect, but that makes it consistent with
some of the other code here. A this point you should *not* be
thinking about using this code. (At all.) (And I mean it.)
2000-01-14 01:00:26 +00:00
ee867c7fdf
Add some debugging printfs.
2000-01-14 00:57:57 +00:00
ec84f30ddd
If we can't malloc a RF_ShutdownList then we should return ENOMEM right away.
...
Waiting for more kernel memory to come along can cause more problems than
we're willing to deal with right now.
2000-01-13 23:41:18 +00:00
2c4bbb1b26
At (closer to) 80-columns, this may actually be readable code.
2000-01-09 04:35:13 +00:00
f2a8aaee02
Garbage collect unused code.
2000-01-09 03:44:33 +00:00
7ce0db89f9
Nuke more debugging stuff and cleanup a tad.
2000-01-09 03:39:13 +00:00
b0caf96f87
It's got to help to check the *right* value if you care what
...
RF_Malloc() is returning.
2000-01-09 03:28:11 +00:00
fbd3b6b396
Another one bites the dust.
2000-01-09 03:17:12 +00:00
87ec1ad21b
Nuke dependency on rf_diskthreads.h
2000-01-09 03:15:59 +00:00
4a54cff7d7
Nuke dependencies on rf_cpuutils.h.
2000-01-09 03:14:32 +00:00
23c4c71c8f
By popular demand (Hi Thor!) make the cylinder size in a default
...
disklabel be the size of the (data part) of the stripe.
2000-01-09 02:56:13 +00:00
754613774a
rf_cpuutil.c isn't useful for the kernel driver.
...
*poof* *poof*
2000-01-09 02:23:21 +00:00
c8723703d2
Not sure why this file was still around since it's really not required.
...
*poof*
2000-01-09 02:05:12 +00:00
6e92bd4525
Nuke unused debugging stuff. Clean up a whole bunch of comments.
2000-01-09 01:45:58 +00:00
a5dcdf8572
Nuke a couple of obsolete comments.
2000-01-09 01:34:29 +00:00
c8730b35e6
- move a bunch of function prototypes to rf_kintf.h
...
- general cleanup of a number of prototypes that were scattered around.
2000-01-09 01:29:27 +00:00
08817bcd50
Nuke #if 0'ed code.
2000-01-09 01:12:05 +00:00
cdb6add0f7
Nuke desc->tid.
2000-01-09 00:00:18 +00:00
37c75452f4
Time's up. This file is no longer needed. "toodles." *punt*
2000-01-08 23:54:26 +00:00
418d4b8471
- more cleanup of rf_get_threadid() and debugging printfs
2000-01-08 23:45:05 +00:00
524894d6f2
Make raidPtr a 'RF_Raid_t *' like it's supposed to be.
2000-01-08 23:02:16 +00:00
04719e3adc
- nuke calls to rf_get_threadid() and associated #include
...
- change a bunch of debugging printfs from
"[%d] ...", tid (where tid is the "thread id")
to
"raid%d: ...", raidPtr->raidid
- other minor rototillage
2000-01-08 22:57:30 +00:00
23864e9422
Bail if rf_numParityRegions is not > 0. Cleanup a comment.
2000-01-08 05:13:26 +00:00
0d2f49b054
Nuke stale comment.
2000-01-08 03:49:37 +00:00
e7fe279817
Minor cleanup.
2000-01-08 03:34:31 +00:00
db96c5344b
Nuke some (now) incorrect comments. More minor cleanup.
2000-01-08 02:04:06 +00:00
ef979643a7
Replace raidPtrs[unit] with just raidPtr, and assign the former to the
...
later early on in raidioctl(). Code is easier to read now too.
2000-01-08 01:52:42 +00:00
b9a1fe405c
Minor tillage on the raidioctl() stuff. More cleanup to follow.
...
No functional changes.
2000-01-08 01:37:37 +00:00
a4ea38dbe8
Nuke an unneeded #ifdef (__NetBSD__)
2000-01-08 01:18:36 +00:00
91273673e7
In RAIDFRAME_CONFIGURE, don't set values in raidPtrs[unit] unless
...
the configuration was successful. Noted by Peter Galbavy of OpenBSD.
2000-01-07 13:57:20 +00:00
bf2d4a40db
rf_GetNextReconEvent() is already prototyped in rf_revent.h.
...
minor cleanup
2000-01-07 03:56:14 +00:00
12dacda256
nuke one call to rf_get_threadid() and cleanup rf_State_Cleanup a bit.
2000-01-07 03:47:15 +00:00
bf081bf6eb
Nuke: #include "rf_threadid.h" and call to rf_get_threadid().
2000-01-07 03:43:39 +00:00
48301a8dae
Nuke: #include "rf_threadid.h".
2000-01-07 03:40:56 +00:00
e5ce17d0eb
Nuke RF_DECLARE_GLOBAL_THREADID, rf_setup_threadid(), rf_shutdown_threadid(),
...
and rf_assign_threadid(). rf_threadid.h, your days are numbered.
2000-01-07 03:25:34 +00:00
86f362fb97
Nuke unused #define.
2000-01-07 03:04:41 +00:00
5f144eaa4b
Shorten IO_BUF_ERR macro by one (unused) arg.
2000-01-07 03:03:44 +00:00
51380e0e8c
- bail if we can't allocate memory for raid_softc's. Set 'numraid' as
...
appropriate for that condition.
- nuke some debugging stuff that we hope we'll never need again
- minor cleanups
2000-01-06 02:06:41 +00:00
d4dfa2a664
Make sure we really get something back from RF_Malloc before we use it.
2000-01-05 04:15:30 +00:00
77ab4d1598
- update RF_CREATE_THREAD to handle a 'process name' argument.
...
- fire up a new thread for parity re-writes, copybacks, and reconstructs.
The ioctl's which trigger these actions now return immediately.
- add progress accounting for the above actions.
- minor rototillage of rf_netbsdkintf.c to deal with all of the above.
2000-01-05 02:57:28 +00:00
28c2380f21
Nuke a bunch of debugging cruft. No functional changes.
1999-12-15 02:02:16 +00:00
fe0726b921
Whoops! Neglected to initialize the buffer queues for anything other
...
than raid0. Fix that...
1999-12-14 15:27:00 +00:00
c8753ea619
Rework how we do the 'wakeup' when an IO completes.
1999-12-12 20:52:37 +00:00
706e3c13c0
Rename rf_DoAccessKernel() to raidstart() and cleanup arguments.
1999-12-12 20:52:16 +00:00
e8a639b1df
Rototill the IO throttle code. Rename rf_DoAccessKernel() to raidstart()
...
and cleanup arguments. While we're here, cleanup raidstrategy(), and nuke
a bunch of unused debugging stuff.
RAIDframe + softdeps now play very nicely together.
1999-12-12 20:51:41 +00:00
a0d72cb555
Trust only the data disk if the mirror is not known to be up-to-date.
...
(this should have been committed with a previous fix for the same
problem in another function in this file :( )
1999-12-09 02:26:09 +00:00
bd19af5aad
Nuke what's left of the DFSTRACE stuff.
1999-12-07 02:54:08 +00:00
e65f9b7c87
More cleanup. DKUSAGE (what little was left of it) goes bye-bye.
1999-12-07 02:40:27 +00:00
a38b721e5f
Nuke the last few bits of RF_DEBUG_ATOMIC that should have
...
disappeared a while back.
1999-12-07 02:13:28 +00:00
bed152a4e7
We don't support RF_DAG_TEST_ACCESS.
1999-12-03 03:35:30 +00:00
128e2e90c5
Nuke some duplicate code.
1999-12-03 03:18:45 +00:00
a038f90588
Kernel memory is not to be wasted. If we're bailing out because of an error,
...
give any allocated memory back to the kernel so someone else can use it.
1999-12-03 03:06:44 +00:00
3888a71ad1
Move LIST_INIT() to a more appropriate place.
1999-12-03 02:43:22 +00:00
5fd68b5d02
We must initialize bp->b_dep in InitBP().
1999-11-18 13:28:06 +00:00
e589155692
Remove some minor stuff that wasn't right, and wasn't even really used.
1999-11-17 01:16:37 +00:00
3949e74f78
If we have a choice: do not trust the parity disk for read
...
balancing in a RAID 1 set if we know that the parity might not
be up-to-date. Thanks to Thor for bringing this to my attention.
1999-11-09 03:07:20 +00:00
fe5b29ca42
Nuke unused prototype.
1999-09-08 02:01:55 +00:00
83abb87d9a
Garbage collect more unused stuff.
1999-09-05 03:05:55 +00:00
9c01870db1
Remove more unused stuff.
1999-09-05 02:54:09 +00:00
cfbdda8843
Garbage collect some more stuff.
1999-09-05 02:24:58 +00:00
6158faa69d
Cleanup a bit more.
1999-09-05 01:58:11 +00:00
e57673d0d4
Remove one of the two bzero calls in RF_Calloc -- the 2nd one doesn't
...
buy us anything at all.
1999-09-05 01:46:35 +00:00
d8144e5848
More cleanup of unused stuff. This time we nuke a bunch of memory
...
allocation stuff.
1999-09-04 21:11:21 +00:00
cfd861fcd8
Minor cleanup of (now) unneeded RF_UTILITY checks.
1999-09-04 16:26:30 +00:00
b1615831d2
Nuke some unused code which, were it enabled, might be useful for
...
performance measurement, but which would cause data corruption.
1999-08-26 02:40:27 +00:00
770de7bb5a
Print a error message instead of panicing in rf_RecoveryPQFunc().
1999-08-15 03:44:46 +00:00
d516e7b571
Make these compile with RF_INCLUDE_RAID6.
1999-08-15 02:36:40 +00:00
ff6d7dd494
Cleanup a bit.
1999-08-15 02:36:03 +00:00
07cc4855f1
Minor cleanup. No functional changes.
1999-08-14 23:34:18 +00:00
4ba14d22b0
Fix an 'off-by-one' error in a RF_ASSERT which meant that a reconstruct
...
would panic in the case where the stripe width was 1 block.
1999-08-14 21:44:24 +00:00
740b9848cc
Decrease the number of outstanding IO's threashold slightly.
...
The previous value may have been too high in some instances.
1999-08-14 03:47:07 +00:00
6409e641a6
Remove a 'struct proc *'-passing abomination that's been bugging me
...
for quite some time.
1999-08-14 03:10:03 +00:00
4a058e82c5
Correct and/or improve detection of bogus row/column values
...
passed in via ioctls.
1999-08-14 02:41:36 +00:00
db2eaf18cf
rf_sys.h does not need to be #included in any of these files, and, actually,
...
is no longer needed at all.
1999-08-13 03:41:52 +00:00
1ec5208866
This file is no longer needed.
1999-08-13 03:31:20 +00:00
1aa6ce1041
Garbage collect rpcc().
1999-08-13 03:27:46 +00:00
02acd1910a
First kick at cleaning up the RF_ETIMER code. New timer handling code
...
provided by Peter Galbavy via OpenBSD.
1999-08-13 03:26:55 +00:00
eca431e87c
Nuke configuration of ETimer stuff during main configuration.
1999-08-13 03:25:01 +00:00
5cab826918
Clean up reconstruction accounting a bit. While it worked before, it was
...
slightly broken in the case where the RAID set did not support reconstruction.
1999-08-13 01:43:33 +00:00
33e458640f
Garbage collect unused struct rf_test_acc.
1999-08-11 02:20:35 +00:00
31b98d3572
When reporting the status of a component during configuration,
...
provide more intuitive information about whether a component is clean
or not.
1999-08-10 22:56:20 +00:00
620eda50c7
Don't allow reconstructs, copybacks, or forced failures of components if
...
using a RAID 0, since RAID 0 can't handle those sorts of things.
1999-08-10 22:46:59 +00:00
0b08c839eb
Don't go ape if there is a problem verifying the parity. Note the condition
...
and inform the user (as opposed to panicing the system!).
1999-08-10 22:19:27 +00:00
49efd10006
Oops. Make minor change that I forgot when adding the parity check ioctl.
1999-08-10 21:41:37 +00:00
a3aefdaee3
Add support for easily checking the parity status from userland.
...
Also: spl protect rf_RewriteParity().
1999-08-10 18:18:30 +00:00
a531d8c31e
Trying to rebuild parity on a degraded set is bogus, so don't allow that.
...
Also, check sooner for a RAID 0 set, since rewriting parity makes no
sense there either.
1999-08-10 01:53:26 +00:00
e0c8203cb3
Nuke unused #define.
1999-08-03 22:07:39 +00:00
5cef006d40
Need to splbio()/splx() protect some of the reconstruction stuff.
...
Without this, we have potentially bad interatctions with the pool
code.
1999-07-21 03:15:26 +00:00
ce7cb7114d
rf_UnconfigureVnodes() needs to be called in the DO_RAID_FAIL macro.
1999-07-19 01:36:07 +00:00
0c7600ca4e
sectorsPerStripeUnit should be larger than 0. If it isn't, simply fail.
...
(Thanks to Thor Lancelot Simon for noting the problem).
1999-07-19 01:35:19 +00:00
324c76b3d9
Once upon a time, long long ago, there was a "fix" added to the
...
RAIDframe driver to stop it from eating too much kernel memory when
writing data. But that fix had a nasty side-affect of hurting write
performance (*much* more than I thought it would). These changes nuke
that "fix", and instead put in a more reasonable mechanism for limiting
the number of simultaneous IO's which can be happening for each RAID device.
The result is a noticeable improvement in write throughput. The End.
1999-07-08 00:45:23 +00:00
5f1ede624f
Make the kthread API a bit friendlier to loadable kernel modules.
1999-07-06 21:51:22 +00:00
86abd11d42
Remember to pass down the proc pointer when adding a hot spare.
...
(Thanks to Martin J. Laubach for helping track this one down.)
1999-06-13 20:36:17 +00:00
8c39e1bf02
When adding a hot spare, remember to lock down the main raidPtr
...
structure before mucking with its contents.
1999-06-04 02:02:39 +00:00
60769ea563
At config time, initialize a small amount of space for the disk queues
...
corresponding to the hot spares. We'll need that space when a spare is
actually added. Huge Thanks to Martin Laubach for helping track this down.
1999-06-04 01:51:00 +00:00
3af0ce9e19
Don't count two problems with a single component as two
...
component failures! Also: remember to configure the disk queues for
the spares -- failure to do that is Bad(tm).
Huge Thanks to Martin Laubach for helping track this down.
1999-06-04 01:50:01 +00:00
daf0e5b05c
Replace two instances of TNF copyright with one (was replicated for two
...
separate contributers).
1999-05-13 21:46:17 +00:00
4882ae464f
Fix a couple of compiler warnings on the Alpha. Pointed out by Tim Rightnour.
1999-04-12 19:39:59 +00:00
9f560505d4
Allow re-writing parity on a RAID 0 to trivially succeed, and to set the
...
clean bit. This is somewhat bogus as RAID 0 does not have any parity,
but is a slightly cleaner than other solutions, and makes the handling
of clean bits for RAID 0 consistent with the handling of clean bits at
other RAID levels.
1999-04-07 14:17:10 +00:00
14de6c2418
Cleanup shutdown code - garbage collect (now) unused shutdownhooks() stuff.
1999-03-27 01:26:37 +00:00
efe288e520
Correct a component label problem where a RAID 1 set with a
...
failed component would not configure properly in certain situations.
1999-03-18 03:02:38 +00:00
2ec2c2045f
Cleanup rf_RewriteParity() and rf_VerifyParity(). Nuke unused stuff,
...
and make more readable.
1999-03-14 22:10:46 +00:00
414b788837
Disallow interrupting parity re-writes or copybacks by removing PCATCH
...
from the tsleep()'s (they probably shouldn't have been there in the
first place!). Making parity re-writing and copybacks interruptable
will require re-designing how a few things are done (e.g. how memory
is freed for structures shipped off to routines that run asynchronously
relative to the calling routine). Fix a few other tsleep's while we're at it.
1999-03-14 21:53:31 +00:00
cd0a696176
Cleanup raidattach: move rf_boot's functionality to raidattach, and
...
nuke rf_boot(). Cleanup raidopen().
1999-03-09 03:53:18 +00:00
e39fe5e2fe
Cleanup unused cruft.
1999-03-09 03:52:41 +00:00
29daf0c77d
rf_update_component_labels() now lives here. Re-work clean-bit
...
support a little: clean-bits now also get frobbed on partition
opening/closing, rather than just at device configuration and device
unconfiguration. Schedule shutdownhooks() stuff for nukage at a later
date, since it isn't really necessary any more.
1999-03-09 02:59:25 +00:00
585c77594e
Move rf_update_component_labels() from here to rf_netbsdkintf.c.
1999-03-09 02:57:29 +00:00
98d8c12355
Update for recent changes including component label support, clean
...
bits, rebuilding components in-place, adding hot spares, shutdownhooks, etc.
1999-03-02 03:18:48 +00:00
ee9187f136
After a vn_close(), set the vnode * to NULL so we don't try closing it
...
again later.
1999-02-27 03:43:20 +00:00
a76f56098c
Do the VOP_UNLOCK() on an initialized vp, rather than random garbage.
1999-02-27 01:48:13 +00:00
65be57243f
Minor cleanups and code rearranging. Now knows a little about component
...
labels and how to deal with hot-adding spare components. Still work in
progress -- component labels and hot-adding are not enabled (yet).
1999-02-24 00:00:03 +00:00
be9eca67c8
Cleanup/remove unused cruft. First kick at component labels and clean bits.
...
Still work in progress. New code is there, but not enabled yet.
1999-02-23 23:57:53 +00:00
07b6b89dce
Move vnode cleanup code into one place. Cleanup a bit.
1999-02-23 23:55:29 +00:00
0d29172def
Do a VOP_UNLOCK() before a vn_close().
1999-02-23 23:53:37 +00:00
6d72c8f0e7
Minor cleanups that got missed last time through.
1999-02-11 01:23:32 +00:00
0014588545
Phase 2 of the RAIDframe cleanup. The source is now closer to KNF
...
and is much easier to read. No functionality changes.
1999-02-05 00:06:06 +00:00
58f7de552d
Nuke more bits of RAIDframe "demo" code. We're not "demoing" here,
...
we're doing the Real Thing!
1999-01-26 04:40:03 +00:00
472e649cd8
More simulator bits going bye-bye.
1999-01-26 04:05:01 +00:00
c52883118d
This stuff is no longer needed. Mostly simulator bits.
1999-01-26 03:56:21 +00:00
a37be85529
rf_ccmn.h no longer needed. Nuke it's inclusion from rf_disks.c.
1999-01-26 03:49:49 +00:00
d94a2c5b2a
These files are not used in the kernel driver. We don't need them any more.
1999-01-26 03:07:10 +00:00
1eecf8e491
RAIDframe cleanup, phase 1. Nuke simulator support, user-land driver,
...
out-dated comments, and other unneeded stuff. This helps prepare
for cleaning up the rest of the code, and adding new functionality.
No functional changes to the kernel code in this commit.
1999-01-26 02:33:49 +00:00
edd259d12e
Make it so raidframe will only perform synchronous writes, and async
...
reads. This avoids a problem where many writes will cause the driver
to allocate way too much memory.
This needs to change to a queueing system later, which will provide a
way to limit the memory consumed by the driver.
Without these changes, raidframe would use 24M or more on my machine when
the buffer cache dumped all its dirty blocks. Now it uses around 200k
or so.
1999-01-15 17:55:52 +00:00
d8f3aefec7
Use M_RAIDFRAME.
1999-01-14 22:49:05 +00:00
4128f9df1e
Duh... check that malloc() hasn't returned a NULL *before* we zero the buffer,
...
not after. Pointed out by Michael Graff.
1999-01-14 20:29:38 +00:00
e392d933c8
Clean up some debugging stuff, and make it compile without warnings
...
with -DDEBUG on alphas.
1998-12-22 20:03:14 +00:00
1b04131eff
Try a bit harder to detect the physical failure of a component, and to
...
mark that component as dead. Physical failures are handled much better now.
1998-12-03 15:14:40 +00:00
3f3d8e85e6
If a configuration fails due to a non-existant spare, remember to release
...
any previously allocated components before bailing. Credit to Manuel Bouyer
for noticing this.
1998-12-03 15:06:25 +00:00
oster