write paths within RAIDframe. They also resolve the "panics with
RAID 5 sets with more than 3 components" issue which was present
(briefly) in the commits which were previously supposed to address
the malloc() issue.
With this new code the 5-component RAID 5 set panics are now gone.
It is also now also possible to swap to RAID 5.
The changes made are:
1) Introduce rf_AllocStripeBuffer() and rf_FreeStripeBuffer() to
allocate/free one stripe's worth of space. rf_AllocStripeBuffer() is
used in rf_MapUnaccessedPortionOfStripe() where it is not sufficient to
allocate memory using just rf_AllocBuffer(). rf_FreeStripeBuffer() is
called from rf_FreeRaidAccDesc(), well after the DAG is finished.
2) Add a set of emergency "stripe buffers" to struct RF_Raid_s.
Arrange for their initialization in rf_Configure(). In low-memory
situations these buffers will be returned by rf_AllocStripeBuffer()
and re-populated by rf_FreeStripeBuffer().
3) Move RF_VoidPointerListElem_t *iobufs from the dagHeader into
into struct RF_RaidAccessDesc_s. This is more consistent with the
original code, and will not result in items being freed "too early".
4) Add a RF_RaidAccessDesc_t *desc to RF_DagHeader_s so that we have a
way to find desc->iobufs.
5) Arrange for desc in the DagHeader to be initialized in InitHdrNode().
6) Don't cleanup iobufs in rf_FreeDAG() -- the freeing is now delayed
until rf_FreeRaidAccDesc() (which is how the original code handled the
allocList, and for which there seem to be some subtle, undocumented
assumptions).
7) Rename rf_AllocBuffer2() to be rf_AllocBuffer() and remove the
former rf_AllocBuffer(). Fix all callers of rf_AllocBuffer().
(This was how it was *supposed* to be after the last time these
changes were made, before they were backed out).
8) Remove RF_IOBufHeader and all references to it.
9) Remove desc->cleanupList and all references to it.
Fixes PR#20191
used in the event that we can't malloc a buffer of the appropriate
size in the traditional way. rf_AllocIOBuffer() and rf_FreeIOBuffer()
deal with allocating/freeing these structures. These buffers are
stored in a list on the 'iobuf' list. iobuf_count keeps track of how
many buffers are available, and numEmergencyBuffers is the effective
"high-water" mark for the freelist. The buffers allocated by
rf_AllocIOBuffer() are stripe-unit sized, which is the maximum
size requested by any of the callers.
Add an iobufs entry to RF_DagHeader_s. Use it for keeping track of
buffers that get allocated from the free-list.
Add a "generic list" pool (VoidPointerListElement Pool) for elements
used to maintain a list of allocated memory. [It is somewhat less
than ideal to add another little pool to handle this...]
Teach rf_AllocBuffer() to use the new rf_AllocIOBuffer(). Modify
other Mallocs to use rf_AllocIOBuffer(), and to update dag_h->iobufs as
appropriate.
Update rf_FreeDAG() to handle cleanup of dag_h->iobufs.
While here, add some missing pool_destroy() calls for a number of pools.
With these changes, it should (in theory) be possible to swap on
RAID 5 sets again. That said, I've not had any success there yet --
but the last issue I saw at least wasn't in RAIDframe. :-}
[There is room for this code to become a bit more consise, but I
wanted to do a checkpoint here with something known to work :) ]
rf_PrintUserStats() was mean for the simulator, and doesn't provide
any real info in kernel-space, especially for reconstructs.
Reconstructing actually renders the stats even more useless, since it
resets them all to zero before the reconstruct starts!
- since rf_PrintUserStats() is no longer used, nuke it along with the
routines that feed it. Nothing was using this code, and if we ever
need it again, we know where to find it.
(other than NULL when raidPtr is initialized). That means
SignalReconDone() never does anything useful. Bye-bye!
Say good-bye to recon_done_procs and recon_done_procs_mutex (and its
initializer) as well.
of strenuous agreement, and some general agreement, this commit is
going ahead because it's now starting to block some other changes I
wish to make.]
Remove most of the support for the concept of "rows" from RAIDframe.
While the "row" interface has been exported to the world, RAIDframe
internals have really only supported a single row, even though they
have feigned support of multiple rows.
Nothing changes in configuration land -- config files still need to
specify a single row, etc. All auto-config structures remain fully
forward/backwards compatible.
The only visible difference to the average user should be a
reduction in the size of a GENERIC kernel (i386) by 4.5K. For those
of us trolling through RAIDframe kernel code, a lot of the driver
configuration code has become a LOT easier to read.
XXX: Current code may have problems if kernel memory is completely depleted.
This is, unfortunately, not the only chunk of RAIDframe code to have
this problem, and will have to be dealt with. :(
rework how completed requests are handled. In particular, instead of
doing all sorts of work and locking in interrupt context, completed
requests are now queued. A new kernel thread (rf_RaidIOThread) now
handles calling rf_DiskIOComplete() and (req->CompleteFunc)() for each
completed request. There is still work to be done to make RAIDframe
LOCKDEBUG friendly, but this change is a huge step forward.
Reviewed by (and many thanks to): thorpej
the stuff that used to live in rf_types.h, rf_raidframe.h, rf_layout.h,
rf_netbsd.h, rf_raid.h, rf_decluster,h, and a few other places.
Believe it or not, when this is all done, things will be cleaner.
No functional changes to RAIDframe.
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.
- 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
- 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)
- 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.
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.
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.
Carnegie Mellon University. Full RAID implementation, including
levels 0, 1, 4, 5, 6, parity logging, and a few other goodies.
Ported to NetBSD by Greg Oster.
perry