Commit Graph

84 Commits

Author SHA1 Message Date
Poul-Henning Kamp
5777c5b989 NO_GEOM cleanup: unifdef 2003-01-30 13:12:31 +00:00
Poul-Henning Kamp
16bcbe8cf7 Implement MDIOCLIST which returns the unit numbers of configured md(4)
devices.

We use the md_pad[] array and if there are more units than its size the
last returned unit number will be -1, but the number of units returned
is correct.
2003-01-27 07:58:18 +00:00
Alfred Perlstein
44956c9863 Remove M_TRYWAIT/M_WAITOK/M_WAIT. Callers should use 0.
Merge M_NOWAIT/M_DONTWAIT into a single flag M_NOWAIT.
2003-01-21 08:56:16 +00:00
Poul-Henning Kamp
26d48b4044 OK Ok, so I didn't check the NO_GEOM case for the final version...
Stumbled on by:	bde
2003-01-13 20:19:04 +00:00
Poul-Henning Kamp
a4f8615810 Enable the new h0h0magic code which on GEOM kernels make the md(4)
driver a _real_ GEOM driver.
2003-01-13 17:31:46 +00:00
Poul-Henning Kamp
0f8500a5b3 Add a mutex around the per unit bioqueue.
Only grab giant in the per unit kthread for SWAP and VNODE backed devices.

Initialize the bioq before the kthread gets a chance to study it.

Don't lock Giant in mddone_swap, we shouldn't need it.
2003-01-13 08:50:23 +00:00
Poul-Henning Kamp
64bfc43b26 Remove the printf which announces the creation of malloc disks: it is
inconsistent when we do not do it for swap or vnode.

We still printf for preloaded disks because of the weak debugging
options people have in embedded/tiny environments where this is
usually used.
2003-01-13 08:01:09 +00:00
Poul-Henning Kamp
6f4f00f114 Add code to make md(4) a GEOM device driver instead of relying in
the disk mini-layer.

This is currently not enabled.
2003-01-12 21:16:49 +00:00
Poul-Henning Kamp
a522a15950 Shift things around a bit in preparation for future evilness. 2003-01-12 17:39:29 +00:00
Ian Dowse
e176446dc8 Move the check for the MD_SHUTDOWN flag to before the tsleep() call
in the per-device kthread. This ensures that synchronisation with
mddestroy() succeeds even if the kthread was not waiting in tsleep()
at the time of the wakeup(). Among other things, this fixes the
problem of mdconfig getting stuck when an attempt is made to use a
zero-length file as a vnode-type backing store.

Approved by:	re
2002-11-30 22:03:53 +00:00
Poul-Henning Kamp
8689acc48c We want /dev/md0 for ramdisk roots, not /dev/md0c.
Sponsored by:	DARPA & NAI Labs
2002-10-21 20:08:28 +00:00
Poul-Henning Kamp
975b628f27 Use ENOSPC error return, not ENOMEM.
Use %jd rather than %lld.
2002-10-20 20:50:31 +00:00
Jake Burkholder
81bb0b95b1 MODINFO_SIZE metadata has type size_t, not unsigned. This makes preloaded
md root work on sparc64.
2002-10-13 18:19:22 +00:00
Scott Long
316ec49abd Some kernel threads try to do significant work, and the default KSTACK_PAGES
doesn't give them enough stack to do much before blowing away the pcb.
This adds MI and MD code to allow the allocation of an alternate kstack
who's size can be speficied when calling kthread_create.  Passing the
value 0 prevents the alternate kstack from being created.  Note that the
ia64 MD code is missing for now, and PowerPC was only partially written
due to the pmap.c being incomplete there.
Though this patch does not modify anything to make use of the alternate
kstack, acpi and usb are good candidates.

Reviewed by:	jake, peter, jhb
2002-10-02 07:44:29 +00:00
Poul-Henning Kamp
1cff889a46 Put the casts on the right hand side of =. 2002-09-28 14:17:24 +00:00
Peter Grehan
a0c6726472 Initialize fwsectors/fwheads to allow the DIOCGFWSECTORS and
DIOCGFWHEADS ioctls to return meaningful values to disklabel/newfs

Approved by: phk
2002-09-22 10:07:18 +00:00
Poul-Henning Kamp
7812d86f03 (This commit touches about 15 disk device drivers in a very consistent
and predictable way, and I apologize if I have gotten it wrong anywhere,
getting prior review on a patch like this is not feasible, considering
the number of people involved and hardware availability etc.)

If struct disklabel is the messenger: kill the messenger.

Inside struct disk we had a struct disklabel which disk drivers used to
communicate certain metrics to the disklayer above (GEOM or the disk
mini-layer).  This commit changes this communication to use four
explicit fields instead.

Amongst the benefits is that the fields do not get overwritten by
wrong or bogus on-disk disklabels.

Once that is clear, <sys/disk.h> which is included in the drivers
no longer need to pull <sys/disklabel.h> and <sys/diskslice.h> in,
the few places that needs them, have gotten explicit #includes for
them.

The disklabel inside struct disk is now only for internal use in
the disk mini-layer, so instead of embedding it, we malloc it as
we need it.

This concludes (modulus any mistakes) the series of disklabel related
commits.

I belive it all amounts to a NOP for all the rest of you :-)

Sponsored by:   DARPA & NAI Labs.
2002-09-20 19:36:05 +00:00
Archie Cobbs
4a6a94d8d8 Replace (ab)uses of "NULL" where "0" is really meant. 2002-08-22 21:24:01 +00:00
Maxime Henrion
6569d6f3ec Yet another warning fix for 64 bits platforms.
Reviewed by:	phk
2002-06-24 12:07:02 +00:00
Poul-Henning Kamp
58f3c42e6d mdcreate_vnode() isn't correctly clearing things out of the linked
list if the file is of 0 size or mdsetcred() fails.

Submitted by:	Martin Faxer <gmh003532@brfmasthugget.se>
2002-06-15 19:18:43 +00:00
Maxim Sobolev
d8a186ebbb - Whitespace only: use return statement consistentlt (return (foo), not
return(foo)), kill extra blank names between function names;
- fix format string in printf(): devtoname() returns string, not pointer.
2002-06-10 19:25:21 +00:00
Ian Dowse
5c97ca54e5 Use a per-device worker thread to avoid blocking in mdstrategy()
until the I/O completes. This fixes some easily reproducable deadlocks
that occur when using md(4) with GEOM.

Reviewed by:	phk
2002-06-03 22:09:04 +00:00
Poul-Henning Kamp
fcf867e9f7 Mis-edit in last commit. 2002-05-26 09:57:59 +00:00
Poul-Henning Kamp
fde2a2e414 Be a bit smarter about rewriting data so we don't loose too much performance.
Sponsored by: DARPA & NAI Labs.
2002-05-26 09:38:51 +00:00
Poul-Henning Kamp
f43b2bac72 Use an umazone per unit for allocating the sectors for malloc backing.
Clean up things properly when we unconfigure malloc backed units.

Sponsored by:	DARPA & NAI Labs.
2002-05-26 06:48:55 +00:00
Poul-Henning Kamp
c6517568df Give the "malloc" backing of md(4) an adaptive multilevel index tree to
remove the need for a contiguous array with pointers to all the sectors.

Try to make failure to malloc(9) memory a non-hang situation.

Eventually this will allow us to test the 64bit cleanness of the disk
I/O patch, but more work is outstanding here and elsewhere.

Sponsored by:	DARPA & NAI Labs.
2002-05-25 20:44:20 +00:00
Poul-Henning Kamp
9589c2561c Fix a memory-leak when configuring a vnode backed md(4) device fails.
Submitted by:	Martin Faxér <gmh003532@brfmasthugget.se>
MFC after:	4 weeks
2002-05-03 17:55:10 +00:00
Jeff Roberson
421e6a659e Remove unused include. 2002-03-20 09:55:07 +00:00
Bruce Evans
1ab0b5f920 The previous commit missed fixing 2 old printf format errors and
introduced a format printf error.
2002-03-19 04:07:29 +00:00
Andrew Gallatin
54ed0c3221 Fix printf warning caused by recent changes in bio_pblkno's type. 2002-03-19 01:45:04 +00:00
Kirk McKusick
0d2af52141 Introduce the new 64-bit size disk block, daddr64_t. Change
the bio and buffer structures to have daddr64_t bio_pblkno,
b_blkno, and b_lblkno fields which allows access to disks
larger than a Terabyte in size. This change also requires
that the VOP_BMAP vnode operation accept and return daddr64_t
blocks. This delta should not affect system operation in
any way. It merely sets up the necessary interfaces to allow
the development of disk drivers that work with these larger
disk block addresses. It also allows for the development of
UFS2 which will use 64-bit block addresses.
2002-03-15 18:49:47 +00:00
John Baldwin
a854ed9893 Simple p_ucred -> td_ucred changes to start using the per-thread ucred
reference.
2002-02-27 18:32:23 +00:00
Poul-Henning Kamp
e087ce2dd2 Staticize the malloc definitions.
Obtained from:	~bde/sys.dif.gz
2002-02-10 21:42:44 +00:00
Poul-Henning Kamp
3ca627fefa Gah! last commit botched indentation, fix indentation and some other
white-space nits while at it.
2002-01-21 20:57:03 +00:00
Poul-Henning Kamp
b4a4f93c5e Restructure slightly, eliminating some repetitive source lines and
making GEOM patches simpler and more readable at the same time.
2002-01-21 20:50:06 +00:00
Dima Dorfman
53d745bc7c Actually make use of the md_version field of 'struct mdio'. In order
not to needlessly break compatibility, decrement MDIOVERSION to 0.

Approved by:	phk
2001-12-20 06:38:21 +00:00
Matthew Dillon
7e76bb562e Implement IO_NOWDRAIN and B_NOWDRAIN - prevents the buffer cache from blocking
in wdrain during a write.  This flag needs to be used in devices whos
strategy routines turn-around and issue another high level I/O, such as
when MD turns around and issues a VOP_WRITE to vnode backing store, in order
to avoid deadlocking the dirty buffer draining code.

Remove a vprintf() warning from MD when the backing vnode is found to be
in-use.  The syncer of buf_daemon could be flushing the backing vnode at
the time of an MD operation so the warning is not correct.

MFC after:	1 week
2001-11-05 18:48:54 +00:00
John Baldwin
bd78cece5d Change the kernel's ucred API as follows:
- crhold() returns a reference to the ucred whose refcount it bumps.
- crcopy() now simply copies the credentials from one credential to
  another and has no return value.
- a new crshared() primitive is added which returns true if a ucred's
  refcount is > 1 and false (0) otherwise.
2001-10-11 23:38:17 +00:00
John Baldwin
9935282d50 Use crhold() instead of crdup(). The md(4) driver doesn't modify the ucred
that it uses, so it merely needs to bump its refcount to make it immutable
rather than obtain its own copy.
2001-10-09 16:37:51 +00:00
Julian Elischer
b40ce4165d KSE Milestone 2
Note ALL MODULES MUST BE RECOMPILED
make the kernel aware that there are smaller units of scheduling than the
process. (but only allow one thread per process at this time).
This is functionally equivalent to teh previousl -current except
that there is a thread associated with each process.

Sorry john! (your next MFC will be a doosie!)

Reviewed by: peter@freebsd.org, dillon@freebsd.org

X-MFC after:    ha ha ha ha
2001-09-12 08:38:13 +00:00
Maxim Sobolev
166cd0b42f OOPS, remove local change that somehow slipped into a commit (I swear that
I already deleted it some time ago). This should fix problem people have with
unsefined reference to `MD_PRELOAD_COMPRESSED'.

Submitted by:	Manfred Antar <null@pozo.com>
2001-08-27 17:48:37 +00:00
Maxim Sobolev
9d4b594558 - On module unload try to detach all configured disks and let unload proceed
if all disks were detached sucessfully;
- use consistent style for return statements and fix several others style
  inconsistencies.

Reviewed by:	ru
Approved by:	phk
2001-08-27 13:25:47 +00:00
Dima Dorfman
e0cebb4027 There is no MD_OBJET disk type, it's actually MD_SWAP. I guess the
former was either a previous or proposed name that kind of snuck in.
2001-08-16 03:04:49 +00:00
Dima Dorfman
26a0ee75c6 Introduce a force option, MD_FORCE, that instructs the driver to
bypass some extra anti-foot-shooting measures.  Currently, its only
effect is to allow detaching a device while it's still open (e.g.,
mounted).  This is useful for testing how the system reacts to a disk
suddenly going away, which can happen with some removeable media.

At this point, the force option is only checked on detach, so it
would've been possible to allow the option to be passed with the
MDIOCDETACH operation.  This was not done to allow the possibility of
having the force flag influence other tests in the future, which may
not necessarily deal with detaching the device.

Reviewed by:	sobomax
Approved by:	phk
2001-08-07 19:23:16 +00:00
Maxim Sobolev
fe603109d1 - Deny detaching requests until device is still open, otherwise it is possible
to hang or panic kernel by detaching disk from which fs is mounted;
- replace "md" with MD_NAME in yet another place.

Reviewed by:	phk
Approved by:	phk
2001-08-02 10:19:13 +00:00
Thomas Moestl
3d3c27fedf Make sure the total number of sectors is not 0 for a vnode-type md to
avoid a division by zero which would occur on open() in this case.

Reviewed by:	phk
2001-07-26 20:05:20 +00:00
Dima Dorfman
10b0e058bb Use MD_NAME and MDCTL_NAME constants where appropriate. 2001-07-18 13:32:38 +00:00
Matthew Dillon
0cddd8f023 With Alfred's permission, remove vm_mtx in favor of a fine-grained approach
(this commit is just the first stage).  Also add various GIANT_ macros to
formalize the removal of Giant, making it easy to test in a more piecemeal
fashion. These macros will allow us to test fine-grained locks to a degree
before removing Giant, and also after, and to remove Giant in a piecemeal
fashion via sysctl's on those subsystems which the authors believe can
operate without Giant.
2001-07-04 16:20:28 +00:00
John Baldwin
277be4c219 We don't need the vm lock to perform a few simple calculations on the
md device's softc.
2001-06-25 18:24:52 +00:00
Poul-Henning Kamp
8adeb35aff Remove MFS compat bits. 2001-05-29 18:49:23 +00:00