]> git.hungrycats.org Git - linux/log
linux
21 years agoMerge http://nfsclient.bkbits.net/linux-2.5
Trond Myklebust [Mon, 9 Feb 2004 20:24:19 +0000 (21:24 +0100)]
Merge http://nfsclient.bkbits.net/linux-2.5
into fys.uio.no:/home/linux/bitkeeper/nfsclient-2.5

21 years ago[PATCH] Fix pc300_tty.c -> implement tiocmset/tiocmget
Marcelo Tosatti [Mon, 9 Feb 2004 01:12:42 +0000 (17:12 -0800)]
[PATCH] Fix pc300_tty.c -> implement tiocmset/tiocmget

PC300 MLPPP support is currently marked broken in 2.6.x.

To fix that, attached patch implements tiocmset/tiocmget methods on the
pc300_tty.c driver, which is the new method tty drivers are supposed to
use instead ioctl.

This fixes two related issues in the ioctl handler:

- ioctl requesting RTS signal would affect DTR signal
- The RTS signal is now handled.

Bonus: Throw out unused ioctl handler

21 years ago[PATCH] fix OOPS for multiple IDE PCI modules and CONFIG_PROC_FS=y
Bartlomiej Zolnierkiewicz [Sun, 8 Feb 2004 23:36:39 +0000 (15:36 -0800)]
[PATCH] fix OOPS for multiple IDE PCI modules and CONFIG_PROC_FS=y

The problem is that when loading next IDE PCI module memory used by
previously registered ide_pci_host_proc_list entry is already unmapped
(because of __{dev}initdata).  This doesn't happen in built-in case
because this memory is freed after all drivers are initialized.
Fix it by removing __{dev}initdata from all ide_pci_host_proc_t.

Thanks to Andre Tomt <andre@tomt.net> for help in debugging this.

21 years ago[PATCH] ide-cd: incorrect use of sector_div
Benjamin Herrenschmidt [Sun, 8 Feb 2004 23:32:15 +0000 (15:32 -0800)]
[PATCH] ide-cd: incorrect use of sector_div

The recent change to ide-cd using sector div is passing
the wrong type to do_div().

21 years agoMerge nathans@xfs.org:/export/hose/bkroot/xfs-linux-2.6
Nathan Scott [Tue, 10 Feb 2004 03:37:08 +0000 (14:37 +1100)]
Merge nathans@xfs.org:/export/hose/bkroot/xfs-linux-2.6
into sgi.com:/source2/xfs-linux-2.6

21 years ago[XFS] Update XFS config entries - add security entry, update ACL entry.
Nathan Scott [Tue, 10 Feb 2004 03:11:43 +0000 (14:11 +1100)]
[XFS] Update XFS config entries - add security entry, update ACL entry.

21 years ago[XFS] Make more xfs errors trappable with panic_mask
Eric Sandeen [Tue, 10 Feb 2004 02:56:25 +0000 (13:56 +1100)]
[XFS] Make more xfs errors trappable with panic_mask

SGI Modid: xfs-linux:xfs-kern:164051a

21 years ago[XFS] Fix a trivial compiler warning, remove some no-longer-used macros.
Nathan Scott [Tue, 10 Feb 2004 02:52:55 +0000 (13:52 +1100)]
[XFS] Fix a trivial compiler warning, remove some no-longer-used macros.

SGI Modid: xfs-linux:xfs-kern:165155a

21 years ago[XFS] Fix data loss when writing into unwritten extents while memory is being reclaimed.
Nathan Scott [Tue, 10 Feb 2004 02:39:45 +0000 (13:39 +1100)]
[XFS] Fix data loss when writing into unwritten extents while memory is being reclaimed.

SGI Modid: xfs-linux:xfs-kern:166324a

21 years ago[XFS] Add back a missing pflags check in releasepage.
Nathan Scott [Tue, 10 Feb 2004 02:31:45 +0000 (13:31 +1100)]
[XFS] Add back a missing pflags check in releasepage.

SGI Modid: xfs-linux:xfs-kern:166323a

21 years ago[XFS] Fix gcc 3.5 compilation for real
Christoph Hellwig [Tue, 10 Feb 2004 02:12:37 +0000 (13:12 +1100)]
[XFS] Fix gcc 3.5 compilation for real

SGI Modid: xfs-linux:xfs-kern:165783a

21 years ago[XFS] Fix compile warning, ensure _pagebuf_lookup_pages return value is inited.
Nathan Scott [Tue, 10 Feb 2004 02:10:01 +0000 (13:10 +1100)]
[XFS] Fix compile warning, ensure _pagebuf_lookup_pages return value is inited.

SGI Modid: xfs-linux:xfs-kern:166301a

21 years ago[XFS] Avoid NULL returns from pagebuf_get
Christoph Hellwig [Tue, 10 Feb 2004 02:07:39 +0000 (13:07 +1100)]
[XFS] Avoid NULL returns from pagebuf_get

SGI Modid: xfs-linux:xfs-kern:166207a

21 years ago[XFS] plug a pagebuf leak
Christoph Hellwig [Tue, 10 Feb 2004 01:53:44 +0000 (12:53 +1100)]
[XFS] plug a pagebuf leak

SGI Modid: xfs-linux:xfs-kern:166109a

21 years ago[XFS] Remove PBF_MAPPABLE
Christoph Hellwig [Tue, 10 Feb 2004 01:33:42 +0000 (12:33 +1100)]
[XFS] Remove PBF_MAPPABLE

SGI Modid: xfs-linux:xfs-kern:166087a

21 years ago[XFS] Use list_move for moving pagebufs between lists, not list_add/list_del
Nathan Scott [Tue, 10 Feb 2004 00:39:10 +0000 (11:39 +1100)]
[XFS] Use list_move for moving pagebufs between lists, not list_add/list_del

SGI Modid: xfs-linux:xfs-kern:166046a

21 years ago[XFS] Remove the lockable/not lockable buffer distinction.
Christoph Hellwig [Tue, 10 Feb 2004 00:20:58 +0000 (11:20 +1100)]
[XFS] Remove the lockable/not lockable buffer distinction.
All metadata buffers are lockable these days.

SGI Modid: xfs-linux:xfs-kern:166042a

21 years ago[XFS] Fix buffer teardown on _pagebuf_lookup_pages failure
Christoph Hellwig [Tue, 10 Feb 2004 00:14:28 +0000 (11:14 +1100)]
[XFS] Fix buffer teardown on _pagebuf_lookup_pages failure

SGI Modid: xfs-linux:xfs-kern:166032a

21 years agoMake SET_INPUT_KEYCODE return the old value, and clean up
Linus Torvalds [Sun, 8 Feb 2004 11:02:20 +0000 (03:02 -0800)]
Make SET_INPUT_KEYCODE return the old value, and clean up
users of this that were very confused indeed.

21 years agoDon't read i8042 data if no data exists
Linus Torvalds [Sun, 8 Feb 2004 10:35:08 +0000 (02:35 -0800)]
Don't read i8042 data if no data exists

21 years agoMerge bk://linux-pnp.bkbits.net/pnp-2.6
Linus Torvalds [Sun, 8 Feb 2004 07:03:49 +0000 (23:03 -0800)]
Merge bk://linux-pnp.bkbits.net/pnp-2.6
into home.osdl.org:/home/torvalds/v2.5/linux

21 years agoMerge http://linux.bkbits.net/linux-2.5
Adam Belay [Sun, 8 Feb 2004 17:00:24 +0000 (17:00 +0000)]
Merge http://linux.bkbits.net/linux-2.5
into neo.rr.com:/home/ambx1/linux/bk/linus-2.5

21 years ago[PNP]: Cleanup Kconfig
Adam Belay [Sun, 8 Feb 2004 15:52:11 +0000 (15:52 +0000)]
[PNP]: Cleanup Kconfig

This patch cleans up the kconfig options for the pnp subsystem.  It
updates the comments and makes pnpbios proc support an optional
feature.

21 years ago[PNP]: Add additonal sysfs entries
Adam Belay [Sun, 8 Feb 2004 15:47:23 +0000 (15:47 +0000)]
[PNP]: Add additonal sysfs entries

This patch adds some aditional information to sysfs for pnp cards.  It
should be useful for userland tools.

21 years ago[PNP]: Card matching code fix
Adam Belay [Sun, 8 Feb 2004 15:44:15 +0000 (15:44 +0000)]
[PNP]: Card matching code fix

This patch updates the matching code to ensure that all requested
devices are present on the card, even if they are in use.  It is
necessary for some ALSA drivers to work properly because early vendors
would have different sets of devices on the same card ids.  It is from
Takashi Iwai <tiwai@suse.de>.

21 years ago[PNP]: file2alias support
Adam Belay [Sun, 8 Feb 2004 15:40:52 +0000 (15:40 +0000)]
[PNP]: file2alias support

This patch updates file2alias.c to support pnp ids.  It is from
Takashi Iwai <tiwai@suse.de>.

21 years ago[PNP] Move ID declarations
Adam Belay [Sun, 8 Feb 2004 15:36:08 +0000 (15:36 +0000)]
[PNP] Move ID declarations

This patch moves the PnP ID declarations to mod_devicetable.h like
most of the other buses.  It is from Takashi Iwai <tiwai@suse.de>.

21 years ago[PNP]: Avoid static requests
Adam Belay [Sun, 8 Feb 2004 15:33:34 +0000 (15:33 +0000)]
[PNP]: Avoid static requests

Recently many PnPBIOS bugs have been triggered by static resource
information requests.  This patch makes an effort to further avoid
making them.

21 years ago[PNP]: Disable resources fix
Adam Belay [Sun, 8 Feb 2004 15:31:19 +0000 (15:31 +0000)]
[PNP]: Disable resources fix

Some PnPBIOSes do not follow the specifications with regard to
disabling devices.  This patch preserves the tag bits, while zeroing
the resource settings.  Previously we would zero the entire buffer.
It has been tested and appears to correct the issue while remaining
compatible with unbroken PnPBIOSes.

21 years ago[PNP]: Resource flags update
Adam Belay [Sun, 8 Feb 2004 15:28:23 +0000 (15:28 +0000)]
[PNP]: Resource flags update

This patch reorganizes resource flags to ensure that manual resource
settings are properly recognized.  This fix is necessary for many ALSA
drivers.  It also prevents comparisons between unset resource
structures.  The bug was discovered by Rene Herman
<rene.herman@keyaccess.nl>, who also wrote an initial version of this
patch.  I made further improvements to ensure that the pnp subsystem
was compatible with this initial change.

21 years ago[PNP]: Fix Serial PnP driver
Adam Belay [Sun, 8 Feb 2004 15:23:01 +0000 (15:23 +0000)]
[PNP]: Fix Serial PnP driver

The serial driver currently fails to unregister its pnp driver upon
module unload.  This patch corrects the problem by calling
pnp_unregister_driver and implementing a proper remove function.

21 years agoMerge bk://kernel.bkbits.net/davem/net-2.6
Linus Torvalds [Sun, 8 Feb 2004 07:01:45 +0000 (23:01 -0800)]
Merge bk://kernel.bkbits.net/davem/net-2.6
into home.osdl.org:/home/torvalds/v2.5/linux

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Sun, 8 Feb 2004 05:22:05 +0000 (21:22 -0800)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/sparc-2.6
David S. Miller [Sun, 8 Feb 2004 05:20:20 +0000 (21:20 -0800)]
Merge davem@nuts.davemloft.net:/disk1/BK/sparc-2.6
into kernel.bkbits.net:/home/davem/sparc-2.6

21 years ago[AF_UNIX]: Print out paths correctly in /proc/net/unix, matching 2.4.x
Hirofumi Ogawa [Sun, 8 Feb 2004 05:03:26 +0000 (21:03 -0800)]
[AF_UNIX]: Print out paths correctly in /proc/net/unix, matching 2.4.x

21 years ago[IPVS]: Unify URL referencing in Kconfig documentation.
Petri T. Koistinen [Sun, 8 Feb 2004 05:02:12 +0000 (21:02 -0800)]
[IPVS]: Unify URL referencing in Kconfig documentation.

21 years ago[IPV6]: Unify URL referencing in Kconfig documentation.
Petri T. Koistinen [Sun, 8 Feb 2004 05:01:43 +0000 (21:01 -0800)]
[IPV6]: Unify URL referencing in Kconfig documentation.

21 years ago[NET_SCHED]: Unify URL referencing in Kconfig documentation.
Petri T. Koistinen [Sun, 8 Feb 2004 05:01:10 +0000 (21:01 -0800)]
[NET_SCHED]: Unify URL referencing in Kconfig documentation.

21 years ago[SCTP]: Unify URL referencing in Kconfig documentation.
Petri T. Koistinen [Sun, 8 Feb 2004 05:00:41 +0000 (21:00 -0800)]
[SCTP]: Unify URL referencing in Kconfig documentation.

21 years ago[ECONET]: Use LL_RESERVED_SPACE() where applicable. Noticed by yoshfuji.
David S. Miller [Sun, 8 Feb 2004 04:58:29 +0000 (20:58 -0800)]
[ECONET]: Use LL_RESERVED_SPACE() where applicable.  Noticed by yoshfuji.

21 years ago[IPV6]: Kill broken and unused IPV6_AUTHHDR support.
Hideaki Yoshifuji [Sun, 8 Feb 2004 04:54:14 +0000 (20:54 -0800)]
[IPV6]: Kill broken and unused IPV6_AUTHHDR support.

21 years ago[SUNRPC]: Use completions instead of sleep_on for rpciod.
Christoph Hellwig [Sun, 8 Feb 2004 04:43:57 +0000 (20:43 -0800)]
[SUNRPC]: Use completions instead of sleep_on for rpciod.

The rpciod shutdown code gives ugly sleep_on without BKL warnings in
-mm.  And it looks indeed somewhat racy.

The easy fix would be to simply use a completion as in the patch below,
but that removes all the signal fuzzing semantics the current code has.
I don't really understand why we want to cancel the operation by
signals, but I think it'd be better to leave that to people familar with
the code anyway..

21 years agoMerge nuts.davemloft.net:/disk1/BK/sparcwork-2.6
David S. Miller [Sun, 8 Feb 2004 03:00:50 +0000 (19:00 -0800)]
Merge nuts.davemloft.net:/disk1/BK/sparcwork-2.6
into nuts.davemloft.net:/disk1/BK/sparc-2.6

21 years agoMerge nuts.davemloft.net:/disk1/BK/network-2.6
David S. Miller [Sun, 8 Feb 2004 02:56:12 +0000 (18:56 -0800)]
Merge nuts.davemloft.net:/disk1/BK/network-2.6
into nuts.davemloft.net:/disk1/BK/net-2.6

21 years ago[PATCH] ide-cd invalidate toc cache on last close
Jens Axboe [Sun, 8 Feb 2004 02:52:09 +0000 (18:52 -0800)]
[PATCH] ide-cd invalidate toc cache on last close

Make sure that TOC gets reloaded after every last close of the device.

21 years agoMerge bk://linux-scsi.bkbits.net/scsi-for-linus-2.6
Linus Torvalds [Sun, 8 Feb 2004 02:49:37 +0000 (18:49 -0800)]
Merge bk://linux-scsi.bkbits.net/scsi-for-linus-2.6
into home.osdl.org:/home/torvalds/v2.5/linux

21 years ago[PATCH] cciss: rmmod oops fix
Andrew Morton [Sun, 8 Feb 2004 01:44:08 +0000 (17:44 -0800)]
[PATCH] cciss: rmmod oops fix

From: mikem@beardog.cca.cpqcorp.net

This patch fixes an Oops when unloading the driver.  Bug fix.  Please
consider this for inclusion.

All of the patches sent out are needed to get the driver in the 2.6 tree up
to the level of the driver that is in the 2.4 tree, excluding this patch
which is not required in 2.4.

More patches will be coming.  They include multi-path failover support,
support for more than 8 controllers, and msi support.  Presently working on a
per logical volume queueing scheme.

21 years ago[PATCH] cciss: use pci_module_init()
Andrew Morton [Sun, 8 Feb 2004 01:43:58 +0000 (17:43 -0800)]
[PATCH] cciss: use pci_module_init()

From: mikem@beardog.cca.cpqcorp.net

This patch uses the pci_module_init wrapper for hot plug cases.  It is in the
2.4.  tree.

21 years ago[PATCH] cciss: improve /proc presentation
Andrew Morton [Sun, 8 Feb 2004 01:43:49 +0000 (17:43 -0800)]
[PATCH] cciss: improve /proc presentation

From: mikem@beardog.cca.cpqcorp.net

This patch changes the way we fill out the /proc files we create.  It now has
human readable volume sizes, RAID levels, etc.  Also removes some fields that
were orginally for debug purposes.  This is in the 2.4 tree.

21 years ago[PATCH] cciss: printk format fix
Andrew Morton [Sun, 8 Feb 2004 01:43:40 +0000 (17:43 -0800)]
[PATCH] cciss: printk format fix

From: mikem@beardog.cca.cpqcorp.net

This patch changes a format specifier to unsigned to prevent the number of
blocks being displayed as a negative value on very large volumes.

21 years ago[PATCH] cciss: avoid reading PCI config space
Andrew Morton [Sun, 8 Feb 2004 01:43:30 +0000 (17:43 -0800)]
[PATCH] cciss: avoid reading PCI config space

From: mikem@beardog.cca.cpqcorp.net

This patch replaces reading directly form PCI config space where possible.
Most of what we need is in the pdev struct.  This is in 2.4.

21 years ago[PATCH] cciss: intialisation oops fix
Andrew Morton [Sun, 8 Feb 2004 01:43:21 +0000 (17:43 -0800)]
[PATCH] cciss: intialisation oops fix

From: mikem@beardog.cca.cpqcorp.net

This patch moves the check of the controller to before trying to enable it.
If a controller is disabled the system will Oops without this fix.  This in
the 2.4 tree.

21 years ago[PATCH] cciss: disble prefetching in ASIC
Andrew Morton [Sun, 8 Feb 2004 01:43:12 +0000 (17:43 -0800)]
[PATCH] cciss: disble prefetching in ASIC

From: mikem@beardog.cca.cpqcorp.net

This patch addresses a bug in the ASIC on the 6400 series controllers.  When
prefetching from host memory we grab an extra 750 or so bytes of data.  If
this occurs on a memory boundary the machine will MCA.  This bug affects IPF
and Alpha based platforms.  It is not known to be a problem on x86.

Prefetch will be disabled via the f/w.  We need to enable it on x86 to
address a _big_ performance hit on RAID 1 operations.

It is in the 2.4 tree.

21 years ago[PATCH] cciss: IRQ sharing fix
Andrew Morton [Sun, 8 Feb 2004 01:43:02 +0000 (17:43 -0800)]
[PATCH] cciss: IRQ sharing fix

From: mikem@beardog.cca.cpqcorp.net

This patch fixes a bug when sharing IRQs with another controller that
receives a lot of interrupts.  Without this check we will panic the system
when unloading and reloading the driver.  This is in 2.4.

21 years ago[PATCH] cciss: Add support for SA 6i embedded controller
Andrew Morton [Sun, 8 Feb 2004 01:42:53 +0000 (17:42 -0800)]
[PATCH] cciss: Add support for SA 6i embedded controller

From: mikem@beardog.cca.cpqcorp.net

This patch adds support for the next generation embedded cciss controller.
It also bumps the version and changes the author to HP.  This patch is in
2.4.

21 years ago[PATCH] cciss: Fix freeing of incorrect IO memory address
Andrew Morton [Sun, 8 Feb 2004 01:42:44 +0000 (17:42 -0800)]
[PATCH] cciss: Fix freeing of incorrect IO memory address

From: mikem@beardog.cca.cpqcorp.net

This patch fixes a bug where under certain error conditions we bail and try
to free our I/O memory.  This patch is in the 2.4 tree.

21 years ago[PATCH] cciss: PCI BAR sizing fix
Andrew Morton [Sun, 8 Feb 2004 01:42:35 +0000 (17:42 -0800)]
[PATCH] cciss: PCI BAR sizing fix

From: <mikem@beardog.cca.cpqcorp.net>

This patch eliminates the bad assumption that all of our PCI BARs will always
be 32-bits.  Tested against the 2.6.2 kernel.

This is required to support the Pinnacles architecture.  It is already in the
2.4 tree.

(This needs to be converted to use pci_request_regions())

21 years ago[PATCH] vx222_ops.c warning fix
Andrew Morton [Sun, 8 Feb 2004 01:42:26 +0000 (17:42 -0800)]
[PATCH] vx222_ops.c warning fix

Print a size_t correctly.

21 years ago[PATCH] alpha: fix build error due to __attribute_const__ conversion
Andrew Morton [Sun, 8 Feb 2004 01:42:17 +0000 (17:42 -0800)]
[PATCH] alpha: fix build error due to __attribute_const__ conversion

Include the needed compiler.h.  Also swizze the return type and the attribute
around to match everyone else.

21 years ago[PATCH] bitmap_snprintf() optimization
Andrew Morton [Sun, 8 Feb 2004 01:42:08 +0000 (17:42 -0800)]
[PATCH] bitmap_snprintf() optimization

From: Joe Korty <joe.korty@ccur.com>

Now that bitmap_parse is part of bitmap.[ch], it is allowed to manipulate
bits directly as the other bitmap routines do.

21 years ago[PATCH] Fix __filemap_fdatawrite() comment
Andrew Morton [Sun, 8 Feb 2004 01:41:59 +0000 (17:41 -0800)]
[PATCH] Fix __filemap_fdatawrite() comment

It was old.

21 years ago[PATCH] Fix ppa/imm warnings
Andrew Morton [Sun, 8 Feb 2004 01:41:50 +0000 (17:41 -0800)]
[PATCH] Fix ppa/imm warnings

From: <viro@parcelfarce.linux.theplanet.co.uk>

Simplify the code, eliminate an unused variable warning.

21 years ago[PATCH] add device id to radeonfb
Andrew Morton [Sun, 8 Feb 2004 01:41:41 +0000 (17:41 -0800)]
[PATCH] add device id to radeonfb

From: Andreas Steinmetz <ast@domdv.de>

The attached patch adds the pci id 5961 to radeonfb.  Without the patch my
9200 displays only a blank screen.  lspci output below.

05:00.0 VGA compatible controller: ATI Technologies Inc Radeon RV280
[Radeon 9200] (rev 01) (prog-if 00 [VGA])
         Subsystem: Giga-byte Technology: Unknown device 4018
         Flags: bus master, 66Mhz, medium devsel, latency 64, IRQ 16
         Memory at e0000000 (32-bit, prefetchable) [size=128M]
         I/O ports at b800 [size=256]
         Memory at feaf0000 (32-bit, non-prefetchable) [size=64K]
         Expansion ROM at feac0000 [disabled] [size=128K]
         Capabilities: [58] AGP version 3.0
         Capabilities: [50] Power Management version 2

21 years ago[PATCH] v850: make __delay function handle a loop count of zero
Andrew Morton [Sun, 8 Feb 2004 01:41:31 +0000 (17:41 -0800)]
[PATCH] v850: make __delay function handle a loop count of zero

From: <miles@mcspd15.ucom.lsi.nec.co.jp> (Miles Bader)

make __delay function handle a loop count of zero.  In practice, this only
seems to occur in odd debugging situations, but it's quite annoying then.

21 years ago[PATCH] v850: Define ARCH_HAS_*_EXTABLE macros for v850
Andrew Morton [Sun, 8 Feb 2004 01:41:22 +0000 (17:41 -0800)]
[PATCH] v850: Define ARCH_HAS_*_EXTABLE macros for v850

From: <miles@mcspd15.ucom.lsi.nec.co.jp> (Miles Bader)

Define ARCH_HAS_*_EXTABLE macros for v850

21 years ago[PATCH] v850: Add some #includes for the v850 to eliminate some compiler warnings
Andrew Morton [Sun, 8 Feb 2004 01:41:12 +0000 (17:41 -0800)]
[PATCH] v850: Add some #includes for the v850 to eliminate some compiler warnings

From: <miles@mcspd15.ucom.lsi.nec.co.jp> (Miles Bader)

Add some #includes for the v850 to eliminate some compiler warnings

21 years agoNFSv4: Add support for POSIX file locking.
Trond Myklebust [Sat, 7 Feb 2004 16:03:03 +0000 (17:03 +0100)]
NFSv4: Add support for POSIX file locking.

21 years agoNFSv4: Fix a bug which was causing Oopses if the client was
Trond Myklebust [Sat, 7 Feb 2004 16:02:21 +0000 (17:02 +0100)]
NFSv4: Fix a bug which was causing Oopses if the client was
mounting more than one partition from the same server.

21 years agoNFSv4: Share open_owner structs between several different
Trond Myklebust [Sat, 7 Feb 2004 16:01:32 +0000 (17:01 +0100)]
NFSv4: Share open_owner structs between several different
processes. Reduces the load on the server.

21 years agoNFSv4: Atomic open(). Fixes races w.r.t. opening files.
Trond Myklebust [Sat, 7 Feb 2004 16:00:45 +0000 (17:00 +0100)]
NFSv4: Atomic open(). Fixes races w.r.t. opening files.

21 years agoRPC: Ensure that we disconnect TCP sockets if there has been no
Trond Myklebust [Sat, 7 Feb 2004 16:00:03 +0000 (17:00 +0100)]
RPC: Ensure that we disconnect TCP sockets if there has been no
NFS traffic for the last 5 minutes. This code also affects
NFSv2/v3.

21 years agoRPC/NFSv4: Allow lease RENEW calls to be soft (i.e. to time
Trond Myklebust [Sat, 7 Feb 2004 15:59:20 +0000 (16:59 +0100)]
RPC/NFSv4: Allow lease RENEW calls to be soft (i.e. to time
out) despite the mount being hard.

21 years agoNFSv4: Basic code for recovering file OPEN state after a server
Trond Myklebust [Sat, 7 Feb 2004 15:58:38 +0000 (16:58 +0100)]
NFSv4: Basic code for recovering file OPEN state after a server
reboot.

21 years agoNFSv4: Preparation for the server reboot recovery code.
Trond Myklebust [Sat, 7 Feb 2004 15:57:57 +0000 (16:57 +0100)]
NFSv4: Preparation for the server reboot recovery code.

21 years agoNFSv4: Don't translate those NFSv4 errors that are needed by
Trond Myklebust [Sat, 7 Feb 2004 15:57:13 +0000 (16:57 +0100)]
NFSv4: Don't translate those NFSv4 errors that are needed by
the kernel itself into EIO.

Fix a signed/unsigned bug in nfs4_increment_seqid.

21 years agoNFSv4: Convert SETCLIENTID and SETCLIENTID_CONFIRM to be
Trond Myklebust [Sat, 7 Feb 2004 15:56:12 +0000 (16:56 +0100)]
NFSv4: Convert SETCLIENTID and SETCLIENTID_CONFIRM to be
standalone operations. Ensure that SETCLIENTID_CONFIRM always
returns the lease timeout length.

21 years agoNFSv4: Split out the code for retrieving static server
Trond Myklebust [Sat, 7 Feb 2004 15:54:20 +0000 (16:54 +0100)]
NFSv4: Split out the code for retrieving static server
information out of the GETATTR compound.

21 years agoNFSv4: Convert the lease renewal daemon from being
Trond Myklebust [Sat, 7 Feb 2004 15:53:15 +0000 (16:53 +0100)]
NFSv4: Convert the lease renewal daemon from being
per-mountpoint to being per-server. Instead of running it on
top of rpciod, convert it to use keventd. This mean we can use
the struct nfs4_client semaphores for ordering purposes.

21 years agoNFSv4: Convert the RENEW operation from using nfs4_compound, to
Trond Myklebust [Sat, 7 Feb 2004 15:52:03 +0000 (16:52 +0100)]
NFSv4: Convert the RENEW operation from using nfs4_compound, to
being a standalone RPC call in preparation for the renew daemon
overhaul.

21 years agoNFSv4/RPCSEC_GSS: Make Frank's server->client_sys feature use
Trond Myklebust [Sat, 7 Feb 2004 15:50:51 +0000 (16:50 +0100)]
NFSv4/RPCSEC_GSS: Make Frank's server->client_sys feature use
RPC cloning in order to avoid duplicating sockets etc. Make
NFSv4 share a single socket for all communication to the same
server.

21 years agoRPC: Add support for sharing the same RPC transport and
Trond Myklebust [Sat, 7 Feb 2004 15:49:04 +0000 (16:49 +0100)]
RPC: Add support for sharing the same RPC transport and
credential caches between different mountpoints by allowing
cloning of the rpc_client struct.

21 years agoRPC: xdr_encode_pages either leaves the tail iovec pointing to
Trond Myklebust [Sat, 7 Feb 2004 15:47:04 +0000 (16:47 +0100)]
RPC: xdr_encode_pages either leaves the tail iovec pointing to
null or, if padding onthe page data is needed, sets it to point
to a little bit of static data. This is a problem if we're
expecting to later append some data in gss_wrap_req. Modify
xdr_encode_pages to make tail point to the end of the data in
head, as xdr_inline_pages and xdr_write_pages do.

21 years agoRPCSEC_GSS: Move the gss sequence number history from the task
Trond Myklebust [Sat, 7 Feb 2004 15:45:27 +0000 (16:45 +0100)]
RPCSEC_GSS: Move the gss sequence number history from the task
structure to the request structure, where it makes more sense.
In particular, when we start storing more sequence number
history (necessary to process responses to resent requests
correctly), this will make it easier to initialize the
necessary data structure in the right place (in
xprt_request_init).

21 years agoRPCSEC_GSS: Client-side only support for rpcsec_gss integrity
Trond Myklebust [Sat, 7 Feb 2004 15:44:01 +0000 (16:44 +0100)]
RPCSEC_GSS: Client-side only support for rpcsec_gss integrity
protection. Since this requires checksumming an entire request,
instead of just the header, and since the request may include,
for example, pages with write data, we modify the gss_api
routines to pass xdr_bufs instead of xdr_netobjs where
necessary.

We add rpcauth_wrap_req and rpcauth_unwrap_resp to rpcauth.c,
wrappers for the new rpc cred ops crwrap_req and crunwrap_req,
which are called just before encoding, and just after decoding,
respectively.

21 years agoRPCSEC_GSS: Instead of having gss_get_mic allocate memory for
Trond Myklebust [Sat, 7 Feb 2004 15:35:27 +0000 (16:35 +0100)]
RPCSEC_GSS: Instead of having gss_get_mic allocate memory for
the mic, require the caller to pass an output buffer whose data
pointer already points to preallocated memory.

21 years agoRPCSEC_GSS: Miscellaneous cleanups of the krb5 code required
Trond Myklebust [Sat, 7 Feb 2004 15:34:13 +0000 (16:34 +0100)]
RPCSEC_GSS: Miscellaneous cleanups of the krb5 code required
for the integrity checksumming mode.

21 years agoRPCSEC_GSS: Fix yet more memory leaks.
Trond Myklebust [Sat, 7 Feb 2004 15:32:34 +0000 (16:32 +0100)]
RPCSEC_GSS: Fix yet more memory leaks.

21 years agoRPCSEC_GSS: Fix two more memory leaks found by the
Trond Myklebust [Sat, 7 Feb 2004 15:31:43 +0000 (16:31 +0100)]
RPCSEC_GSS: Fix two more memory leaks found by the
Stanford checker.

21 years agoRPCSEC_GSS: Oops. Major memory leak here.
Trond Myklebust [Sat, 7 Feb 2004 15:30:50 +0000 (16:30 +0100)]
RPCSEC_GSS: Oops. Major memory leak here.

21 years agoRPCSEC_GSS: Make it safe to share crypto tfms among
Trond Myklebust [Sat, 7 Feb 2004 15:30:11 +0000 (16:30 +0100)]
RPCSEC_GSS: Make it safe to share crypto tfms among
multiple threads.

21 years agoNFSv4: Bugfixes and cleanups for the NFSv4 client
Trond Myklebust [Sat, 7 Feb 2004 15:29:20 +0000 (16:29 +0100)]
NFSv4: Bugfixes and cleanups for the NFSv4 client
name to uid mapper. Includes a fix by Tim Woods to
deal with a caching bug in the case where a user
and a group share the same numerical id and/or name.

21 years agoNFSv4: Bugfixes for the NFSv4 client name to uid mapper.
Trond Myklebust [Sat, 7 Feb 2004 15:27:44 +0000 (16:27 +0100)]
NFSv4: Bugfixes for the NFSv4 client name to uid mapper.
Fixes a memory-scribble problem.

21 years agoNFSv4: Fix an Oopsable condition if we fail to get the
Trond Myklebust [Sat, 7 Feb 2004 15:26:47 +0000 (16:26 +0100)]
NFSv4: Fix an Oopsable condition if we fail to get the
root directory under NFSv4.

21 years agoRPCSEC_GSS: Make the upcalls detect if the userland daemon
Trond Myklebust [Sat, 7 Feb 2004 15:25:51 +0000 (16:25 +0100)]
RPCSEC_GSS: Make the upcalls detect if the userland daemon
dies while processing a request.

21 years agoRPCSEC_GSS: More fixes to the upcall mechanism.
Trond Myklebust [Sat, 7 Feb 2004 15:24:53 +0000 (16:24 +0100)]
RPCSEC_GSS: More fixes to the upcall mechanism.

21 years agoNFSv4/RPCSEC_GSS: Ensure that RPC userland upcalls time out
Trond Myklebust [Sat, 7 Feb 2004 15:23:52 +0000 (16:23 +0100)]
NFSv4/RPCSEC_GSS: Ensure that RPC userland upcalls time out
correctly if the corresponding userland daemon is not up
and running.

21 years ago[PATCH] fix duplication of DMA {black,white}list in icside.c
Bartlomiej Zolnierkiewicz [Sat, 7 Feb 2004 12:48:53 +0000 (04:48 -0800)]
[PATCH] fix duplication of DMA {black,white}list in icside.c

Always compile ide-dma.c if CONFIG_BLK_DEV_IDEDMA=y, mark PCI specific code
with CONFIG_BLK_DEV_IDEDMA_PCI for now (it should migrate to ide_pcidma.c
over a time).  This fixes a small bug - in_drive_list() from icside.c used
!strstr() instead of strstr() so it was missing two entries from a blacklist.

21 years ago[PATCH] Fix imm/ppa initializing bug in driver updates
Alexander Viro [Sat, 7 Feb 2004 12:47:32 +0000 (04:47 -0800)]
[PATCH] Fix imm/ppa initializing bug in driver updates

Let's play spot the bug here...

Amazing that it survives in modular case...

21 years ago[PATCH] fix too early probing of default IDE ports
Bartlomiej Zolnierkiewicz [Sat, 7 Feb 2004 08:43:26 +0000 (00:43 -0800)]
[PATCH] fix too early probing of default IDE ports

It was broken by IDE modular fixes in 2.6.2.

21 years ago[PATCH] fix ns87415.c for PA-RISC Super I/O chip
Bartlomiej Zolnierkiewicz [Sat, 7 Feb 2004 05:05:55 +0000 (21:05 -0800)]
[PATCH] fix ns87415.c for PA-RISC Super I/O chip

From: Matthew Wilcox <willy@debian.org>