- 06 Jan, 2006 40 commits
-
-
NeilBrown authored
With this patch it is possible to poll /proc/mdstat to detect arrays appearing or disappearing, to detect failures, recovery starting, recovery completing, and devices being added and removed. It is similar to the poll-ability of /proc/mounts, though different in that: We always report that the file is readable (because face it, it is, even if only for EOF). We report POLLPRI when there is a change so that select() can detect it as an exceptional event. Not only are these exceptional events, but that is the mechanism that the current 'mdadm' uses to watch for events (It also polls after a timeout). (We also report POLLERR like /proc/mounts). Finally, we only reset the per-file event counter when the start of the file is read, rather than when poll() returns an event. This is more robust as it means that an fd will continue to report activity to poll/select until the program clearly responds to that activity. md_new_event takes an 'mddev' which isn't currently used, but it will be soon. Signed-off-by:
Neil Brown <neilb@suse.de> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
NeilBrown authored
Add in correct read-error handling for resync and read-only situations. When read-only, we don't over-write, so we need to mark the failed drive in the r10_bio so we don't re-try it. During resync, we always read all blocks, so if there is a read error, we simply over-write it with the good block that we found (assuming we found one). Note that the recovery case still isn't handled in an interesting way. There is nothing useful to do for the 2-copies case. If there are 3 or more copies, then we could try reading from one of the non-missing copies, but this is a bit complicated and very rarely would be used, so I'm leaving it for now. Signed-off-by:
-
NeilBrown authored
Largely just a cross-port from raid1. Signed-off-by:
-
NeilBrown authored
We are inadvertently setting the R1BIO_Uptodate bit on read errors when we decide not to try correcting (because there are no other working devices). This means that the read error is reported to the client as success. Signed-off-by:
-
NeilBrown authored
Where performing a user-requested 'check' or 'repair', we read all readable devices, and compare the contents. We only write to blocks which had read errors, or blocks with content that differs from the first good device found. Signed-off-by:
-
NeilBrown authored
Also keep count on the number of errors found. Signed-off-by:
-
NeilBrown authored
There is this "FIXME" comment with a typo in it!! that been annoying me for days, so I just had to remove it. conf->disks[i].rdev should only be accessed if - we know we hold a reference or - the mddev->reconfig_sem is down or - we have a rcu_readlock handle_stripe was referencing rdev in three places without any of these. For the first two, get an rcu_readlock. For the last, the same access (md_sync_acct call) is made a little later after the rdev has been claimed under and rcu_readlock, if R5_Syncio is set. So just use that access... However R5_Syncio isn't really needed as the 'syncing' variable contains the same information. So use that instead. Issues, comment, and fix are identical in raid5 and raid6. Signed-off-by:
-
NeilBrown authored
Signed-off-by:
-
NeilBrown authored
Handling of read errors during resync is separate from handling of read errors during normal IO in raid1. A previous patch added support for read errors during normal IO. This one adds support for read errors during resync or recovery. The key differences are that we don't need to freeze the array, because the normal handling of resync means that this part of the array will be idle except for resync, and the read/overwrite/re-read is needed in a separate piece of code. Signed-off-by:
-
NeilBrown authored
We are dereferencing ->rdev without an rcu lock! Signed-off-by:
-
NeilBrown authored
On a read-error we suspend the array, then synchronously read the block from other arrays until we find one where we can read it. Then we try writing the good data back everywhere and make sure it works. If any write or subsequent read fails, only then do we fail the device out of the array. To be able to suspend the array, we need to also keep track of how many requests are queued for handling by raid1d. Signed-off-by:
-
NeilBrown authored
This is a simple port of match functionality across from raid5. If we get a read error, we don't kick the drive straight away, but try to over-write with good data first. Signed-off-by:
-
NeilBrown authored
raid6 currently does not check the P/Q syndromes when doing a resync, it just calculates the correct value and writes it. Doing the check can reduce writes (often to 0) for a resync, and it is needed to properly implement the echo check > sync_action operation. This patch implements the appropriate checks and tidies up some related code. It also allows raid6 user-requested resync to bypass the intent bitmap. Signed-off-by:
-
NeilBrown authored
Signed-off-by:
-
NeilBrown authored
This is important because bitmap_create uses mddev->resync_max_sectors and that doesn't have a valid value until after the array has been initialised (with pers->run()). [It doesn't make a difference for current personalities that support bitmaps, but will make a difference for raid10] This has the added advantage of meaning with can move the thread->timeout manipulation inside the bitmap.c code instead of sprinkling identical code throughout all personalities. Signed-off-by:
-
NeilBrown authored
See patch to md.txt for more details Signed-off-by:
-
NeilBrown authored
Resync code: A test that isn't needed, a 'compute_block' that makes more sense elsewhere (And then doesn't need a test), a couple of BUG_ONs to confirm the change makes sense. Printks: A few were missing KERN_* Also fix a typo in a comment.. Signed-off-by: -
NeilBrown authored
raid10 needs to put up a barrier to new requests while it does resync or other background recovery. The code for this is currently open-coded, slighty obscure by its use of two waitqueues, and not documented. This patch gathers all the related code into 4 functions, and includes a comment which (hopefully) explains what is happening. Signed-off-by:
-
NeilBrown authored
raid1 needs to put up a barrier to new requests while it does resync or other background recovery. The code for this is currently open-coded, slighty obscure by its use of two waitqueues, and not documented. This patch gathers all the related code into 4 functions, and includes a comment which (hopefully) explains what is happening. Signed-off-by:
-
Darrick J. Wong authored
I've been attempting to set up a (Host)RAID mirror with dm_mirror on 2.6.14.3, and I've been having a strange little problem. The configuration in question is a set of 9GB SCSI disks that have 17942584 sectors. I set up the dm_mirror table as such: 0 17942528 mirror core 2 2048 nosync 2 8:48 0 8:64 0 If I'm not mistaken, this sets up a 9GB RAID1 mriror with 1MB stripes across both SCSI disks. The sector count of the dm device is less than the size of the disks, so we shouldn't fall off the end. However, I always get the messages like this in dmesg when I set up the dm table: attempt to access beyond end of device sdd: rw=0, want=17958656, limit=17942584 Clearly, something is trying to read sectors past the end of the drive. I traced it down to the __rh_recovery_prepare function in dm-raid1.c, which gets called when we're putting the mirror set together. This function calls the dirty region log's get_resync_work function to see if there's any resync that needs to be done, and queues up any areas that are out of sync. The log's get_resync_work function is actually a pointer to the core_get_resync_work function in dm-log.c. The core_get_resync_work function queries a bitset lc->sync_bits to find out if there are any regions that are out of date (i.e. the bit is 0), which is where the problem occurs. If every bit in lc->sync_bits is 1 (which is the case when we've just configured a new RAID1 with the nosync option), the find_next_zero_bit does NOT return the size parameter (lc->region_count in this case), it returns the size parameter rounded up to the nearest multiple of 32! I don't know if this is intentional, but i386 and x86_64 both exhibit this behavior. In any case, the statement "if (*region == lc->region_count)" looks like it's supposed to catch the case where are no regions to resync and return 0. Since find_next_zero_bit apparently has a habit of returning a value that's larger than lc->region_count, the enclosed patch changes the equality test to a greater-than test so that we don't try to resync areas outside of the RAID1 region. Seeing as the HostRAID metadata lives just past the end of the RAID1 data, mucking around in that area is not a good idea. I suppose another way to fix this would be to amend find_next_zero_bit so that it doesn't return values larger than "size", but I don't know if there's a reason for the current behavior. Signed-Off-By:
Darrick J. Wong <djwong@us.ibm.com> Acked-by:
Alasdair G Kergon <agk@redhat.com> Signed-off-by:
-
Stefan Rompf authored
Zap the memory before freeing it so we don't leave crypto information around in memory. Signed-off-by:
Stefan Rompf <stefan@loplof.de> Acked-by:
Clemens Fruhwirth <clemens@endorphin.org> Acked-by:
-
Adrian Bunk authored
This patch #if 0's the not yet implemented global function kcopyd_cancel(). Signed-off-by:
Adrian Bunk <bunk@stusta.de> Acked-by:
-
Alasdair G Kergon authored
Add ioctl DM_SKIP_LOCKFS_FLAG for userspace to request that lock_fs is bypassed when suspending a device. There's no change to the behaviour of existing code that doesn't know about the new flag. Signed-off-by:
-
Alasdair G Kergon authored
Devices only needs syncing when creating snapshots, so make this optional when suspending a device. Signed-off-by:
-
Alasdair G Kergon authored
Rename frozen_bdev to suspended_bdev and move the bdget outside lockfs. (This prepares for making lockfs optional.) Signed-off-by:
-
Jonathan E Brassow authored
This patch introduces a new field to the mirror_set (default_mirror) to store the default mirror. (A subsequent patch will allow us to change the default mirror in the event of a failure.) Signed-off-by:
-
Alasdair G Kergon authored
Use %llu not %Lu in sscanf/printf format strings. Signed-off-by:
-
Andrew Stribblehill authored
This patch removes an unused #define. Signed-off-by:
-
Alasdair G Kergon authored
More snapshot metadata reading into separate function, to prepare for changing the place it gets called from. Signed-off-by:
-
goggin, edward authored
After changing the name of a mapped device, trigger a dm event. (For userspace multipath tools.) Signed-off-by:
-
David Teigland authored
Add dm_get_dev() to get a mapped device given its dev_t. Signed-off-by:
-
David Teigland authored
Abstract dm_find_md() from dm_get_mdptr() to allow use elsewhere. Signed-off-by:
-
Neil Brown authored
A typical nfsd call trace is nfsd -> svc_process -> nfsd_dispatch -> nfsd3_proc_write -> nfsd_write ->nfsd_vfs_write -> vfs_writev These add up to over 300 bytes on the stack. Looking at each of these, I see that nfsd_write (which includes nfsd_vfs_write) contributes 0x8c to stack usage itself!! It turns out this is because it puts a 'struct iattr' on the stack so it can kill suid if needed. The following patch saves about 50 bytes off the stack in this call path. Signed-off-by:
-
David Shaw authored
Both vfs_getattr and i_op->fsync return error statuses which nfsd was largely ignoring. This as noticed when exporting directories using fuse. This patch cleans up most of the offences, which involves moving the call to vfs_getattr out of the xdr encoding routines (where it is too late to report an error) into the main NFS procedure handling routines. There is still a called to vfs_gettattr (related to the ACL code) where the status is ignored, and called to nfsd_sync_dir don't check return status either. Signed-off-by:
-
Olaf Kirch authored
I submitted this one previously - svc_tcp_recvfrom currently returns any errors to the caller, including ECONNRESET and the like. This is something svc_recv isn't able to deal with: len = svsk->sk_recvfrom(rqstp); [...] if (len == 0 || len == -EAGAIN) { [...] return -EAGAIN; } [...] return len; The nfsd main loop will exit when it sees an error code other than EAGAIN. The following patch fixes this problem svc_recv is not equipped to deal with error codes other than EAGAIN, and will propagate anything else (such as ECONNRESET) up to nfsd, causing it to exit. Signed-off-by:Olaf Kirch <okir@suse.de> Cc: Trond Myklebust <trond.myklebust@fys.uio.no> Cc: Neil Brown <neilb@cse.unsw.edu.au> Signed-off-by:
-
Jan Kara authored
Split the checkpoint list of the transaction into two lists. In the first list we keep the buffers that need to be submitted for IO. In the second list are kept buffers that were already submitted and we just have to wait for the IO to complete. This should simplify a handling of checkpoint lists a bit and can eventually be also a performance gain. Signed-off-by:
Jan Kara <jack@suse.cz> Signed-off-by:
-
Jayachandran C authored
This patch fixes an issue reported by Coverity in kernel/module.c Error reported: Cannot reach this line of code "else return ptr;" Patch description: This is the error path, so 'err' will be negative, the else case is not required, this patch removes it. Signed-off-by:
Jayachandran C. <c.jayachandran@gmail.com> Signed-off-by:
-
Paul E. McKenney authored
I missed a use of list_for_each_rcu_safe() in -mm tree. Here is an updated patch to fix it. This time tested on a machine that actually uses IPMI... (Thanks to Serge Hallyn for spotting this.) Signed-off-by:
"Paul E. McKenney" <paulmck@us.ibm.com> Cc: Corey Minyard <minyard@acm.org> Cc: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by:
-
Adrian Bunk authored
"extern inline" doesn't make much sense. Signed-off-by:
-
Adrian Bunk authored
Unless I miss something, this should be the simplest way to express the intended dependencies. Signed-off-by:
-
