]> git.hungrycats.org Git - linux/log
linux
21 years agoLinux 2.6.4-rc1 v2.6.4-rc1
Linus Torvalds [Fri, 27 Feb 2004 06:17:12 +0000 (22:17 -0800)]
Linux 2.6.4-rc1

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Fri, 27 Feb 2004 05:39:13 +0000 (21:39 -0800)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years agoMerge http://linux-lksctp.bkbits.net/lksctp-2.5.work
David S. Miller [Fri, 27 Feb 2004 05:25:15 +0000 (21:25 -0800)]
Merge http://linux-lksctp.bkbits.net/lksctp-2.5.work
into nuts.davemloft.net:/disk1/BK/net-2.6

21 years ago[SUNRPC]: Use '%z' for size_t printk in svcauth_gss.c
David S. Miller [Fri, 27 Feb 2004 04:30:30 +0000 (20:30 -0800)]
[SUNRPC]: Use '%z' for size_t printk in svcauth_gss.c

21 years agoMerge bk://gkernel.bkbits.net/misc-2.5
Linus Torvalds [Fri, 27 Feb 2004 03:40:17 +0000 (19:40 -0800)]
Merge bk://gkernel.bkbits.net/misc-2.5
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years ago[IDE] Create DECLARE_PIIX_DEV declarator, to eliminate a large amount
Jeff Garzik [Fri, 27 Feb 2004 08:03:51 +0000 (03:03 -0500)]
[IDE] Create DECLARE_PIIX_DEV declarator, to eliminate a large amount
of redundancy, and to help prevent future errors.

21 years ago[PATCH] ppc64: virtual IO bus updates
Hollis Blanchard [Fri, 27 Feb 2004 03:30:41 +0000 (19:30 -0800)]
[PATCH] ppc64: virtual IO bus updates

This makes GregKH happy by removing the device name from the
device.bus_id field (and replacing it with a "name" sysfs attribute).

It also renames the parent device from "vdevice" to "vio", making the
/sys/bus and /sys/devices hierarchies consistent.

21 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-rmk
Linus Torvalds [Fri, 27 Feb 2004 03:29:59 +0000 (19:29 -0800)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-rmk
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years ago[ARM] Make free_memmap() use PFNs instead of physical addresses.
Russell King [Fri, 27 Feb 2004 18:16:49 +0000 (18:16 +0000)]
[ARM] Make free_memmap() use PFNs instead of physical addresses.

21 years ago[ARM] amba_{request,release}_regions
Russell King [Fri, 27 Feb 2004 17:18:54 +0000 (17:18 +0000)]
[ARM] amba_{request,release}_regions

Add documentation for the above two functions.  Allow a NULL name to
be passed to amba_request_regions.  In this case, we use the driver's
name associated with the device.

21 years ago[ARM] Add resources and platform devices for SA11x0 serial ports.
Russell King [Fri, 27 Feb 2004 17:04:55 +0000 (17:04 +0000)]
[ARM] Add resources and platform devices for SA11x0 serial ports.

21 years ago[ARM] Allow decompressor to use "cache type" register
Russell King [Fri, 27 Feb 2004 16:37:12 +0000 (16:37 +0000)]
[ARM] Allow decompressor to use "cache type" register

This allows the decompressor to automatically detect the geometry of
the cache while booting to ensure correct behaviour over a wider
number of ARM CPUs.

21 years ago[ARM] Optimise readsl.
Russell King [Fri, 27 Feb 2004 15:50:02 +0000 (15:50 +0000)]
[ARM] Optimise readsl.

This optimises the underlying string-long read function for little
endian platforms.

21 years ago[ARM] Improve bad IRQ reporting.
Russell King [Fri, 27 Feb 2004 15:41:20 +0000 (15:41 +0000)]
[ARM] Improve bad IRQ reporting.

This gives greater information on the cause of the bad IRQ, allowing
the cause to be more effectively traced.

21 years ago[ARM] Update mach-types definitions file.
Russell King [Fri, 27 Feb 2004 15:27:48 +0000 (15:27 +0000)]
[ARM] Update mach-types definitions file.

21 years ago[ARM] Remove obsolete sysctl PM interface.
Russell King [Fri, 27 Feb 2004 15:22:13 +0000 (15:22 +0000)]
[ARM] Remove obsolete sysctl PM interface.

The old sysctl PM interface has been replaced by the driver models
interface.  This has been gone for a while in the -rmk tree and no
one complained, so there's no point keeping it around.

21 years ago[ARM] Add SA11x0 sched_clock() implementation.
Russell King [Fri, 27 Feb 2004 15:16:18 +0000 (15:16 +0000)]
[ARM] Add SA11x0 sched_clock() implementation.

21 years ago[ARM] Update sa1111-pcibuf for dmapool changes.
Russell King [Fri, 27 Feb 2004 14:02:35 +0000 (14:02 +0000)]
[ARM] Update sa1111-pcibuf for dmapool changes.

- use dev_dbg for device-centric debugging messages
- use pr_debug for general debugging messages
- use dmapools instead of pcipools
- use NULL rather than 0 for NULL pointers
- use enum dma_data_direction rather than int
- use DMA_* direction definitions rather than PCI_DMA_*
- only check for sane DMA direction on mapping functions, but
  check that DMA direction matches when unmapping/syncing.

21 years ago[ARM] Cleanup MODULE_* macros
Russell King [Fri, 27 Feb 2004 10:16:19 +0000 (10:16 +0000)]
[ARM] Cleanup MODULE_* macros

Thanks to Adrian Bunk.

Remove kernel 2.0 #ifdef's from arm code, and move MODULE_* to
the end of the file.  Add an appropriate MODULE_LICENSE().

21 years agoAdd d_type information to legacy readdir system call
Linus Torvalds [Fri, 27 Feb 2004 02:28:57 +0000 (18:28 -0800)]
Add d_type information to legacy readdir system call

21 years agoppc64: fix non-iSeries build
Linus Torvalds [Fri, 27 Feb 2004 02:28:08 +0000 (18:28 -0800)]
ppc64: fix non-iSeries build

21 years ago[SCTP] Fix incorrect INIT process termination with sinit_max_init_timeo.
Sridhar Samudrala [Fri, 27 Feb 2004 01:45:39 +0000 (17:45 -0800)]
[SCTP] Fix incorrect INIT process termination with sinit_max_init_timeo.

Currently, sinit_max_init_timeo is treated as the maximum timeout
value for INIT retransmissions and the INIT process is aborted when
the timeout reaches this value. But as per the SCTP sockets API draft,
sinit_max_init_timeo only limits the timeout. The INIT process should
be aborted only after MAX_INIT_RETRANSMITS.

21 years ago[PATCH] ppc64: defconfig update
Anton Blanchard [Fri, 27 Feb 2004 01:29:02 +0000 (17:29 -0800)]
[PATCH] ppc64: defconfig update

Add an iseries defconfig and update the pseries one.

21 years ago[PATCH] ppc64: remove CONFIG_VETH
Anton Blanchard [Fri, 27 Feb 2004 01:28:51 +0000 (17:28 -0800)]
[PATCH] ppc64: remove CONFIG_VETH

CONFIG_VETH isnt used anymore so we lets get rid of it.

21 years ago[PATCH] ppc64: fix pci hotplug compile error
Anton Blanchard [Fri, 27 Feb 2004 01:28:39 +0000 (17:28 -0800)]
[PATCH] ppc64: fix pci hotplug compile error

Fix a compile error in the ppc64 pci hotplug code.

21 years ago[PATCH] ppc64: TLB flush rework
Anton Blanchard [Fri, 27 Feb 2004 01:28:28 +0000 (17:28 -0800)]
[PATCH] ppc64: TLB flush rework

ppc64 tlb flush rework from Paul Mackerras

Instead of doing a double pass of the pagetables, we batch things
up in the pte flush routines and then shoot the batch down in
flush_tlb_pending.

Our page aging was broken, we never flushed entries out of the ppc64
hashtable. We now flush in ptep_test_and_clear_young.

A number of other things were fixed up in the process:

- change ppc64_tlb_batch to per cpu data
- remove some LPAR debug code
- be more careful with ioremap_mm inits
- clean up arch/ppc64/mm/init.c, create tlb.c

21 years ago[PATCH] ppc64: fix kernel access of user pages
Anton Blanchard [Fri, 27 Feb 2004 01:28:17 +0000 (17:28 -0800)]
[PATCH] ppc64: fix kernel access of user pages

Set the ks bit on userspace segments otherwise the kernel can read/write
into userspace mprotected pages.

21 years ago[PATCH] ppc64:Implement support for Apple DART IOMMU (PowerMac G5)
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:16:59 +0000 (17:16 -0800)]
[PATCH] ppc64:Implement support for Apple DART IOMMU (PowerMac G5)

This allows us to use more than 2Gb of RAM on a PowerMac G5.  You can
also boot with iommu=off to disable it, but then you get back the 2Gb
RAM limitation.

21 years ago[PATCH] ppc64: Fix drivers/ide when using an IOMMU
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:16:46 +0000 (17:16 -0800)]
[PATCH] ppc64: Fix drivers/ide when using an IOMMU

Make IDE advertise only 128 entries of SG table on archs with an IOMMU.

The current IOMMU implementations of pci_map_sg() may produce segments
that don't match the boundary requirements of IDE, thus causing the
driver to break them up.  The BIO is supposed to account for that,
however, it cannot account for a pci_map_sg producing a segment of the
requested size, but with incorrect alignement, thus we may still try to
break up the list in more entries than is supported by the HW.

A similar fix already went in libata.  The "real" long term fix will be
to move the boundary requirements to struct device so that pci_map_sg()
can respect them when producing the sglist.  In the meantime, this
band-aid works around the problem.

21 years ago[PATCH] ppc64: iommu rewrite
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:16:34 +0000 (17:16 -0800)]
[PATCH] ppc64: iommu rewrite

Lots of things renamed, sillicaps killed, stuffs moved around and common
code properly extracted from implementation specific code, new
allocator, etc...  The code is overall a lot simpler, faster, less prone
to fail, and a lot more manageable.  I didn't use "bk mv", there is no
need to keep the old history attached to the new file.

21 years ago[PATCH] ppc64: Fix /dev/mem idea of what is memory
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:15:56 +0000 (17:15 -0800)]
[PATCH] ppc64: Fix /dev/mem idea of what is memory

This adds a hack to /dev/mem (along with the other ones already there)
to prevent mapping cacheable of the IO hole.

Without this, XFree blows up on machines with enough memory to go past
the IO hole.  It also tries to prevent memory from beeing mapped
uncached.  Cache paradoxes are evil and can kill the CPU.

The necessary page_is_ram() call was added by the previous patch doing
the proper IO hole accounting.

21 years ago[PATCH] ppc64: physical RAM accounting fix
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:15:45 +0000 (17:15 -0800)]
[PATCH] ppc64: physical RAM accounting fix

On machines with an IO hole (like Apple G5 with more than 2Gb of RAM,
but also some pSeries) we failed to properly account for the real
amount of physical RAM and inform the zone allocator of our hole size.

During the process, I included Anton slaughtering of the guard page we
had in the first 256Mb kernel segment, thus allowing this segment to be
mapped with large pages as it should be.

21 years ago[PATCH] radeonfb: small cleanup of common register init
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:15:36 +0000 (17:15 -0800)]
[PATCH] radeonfb: small cleanup of common register init

This removes the unused common_regs_m6 and add one more register to be
cleared, according to latest XFree code from ATI.

21 years ago[PATCH] ppc64: Fix warning on pmac build
Benjamin Herrenschmidt [Fri, 27 Feb 2004 01:15:25 +0000 (17:15 -0800)]
[PATCH] ppc64: Fix warning on pmac build

Fix a stupid warning in PowerMac SMP build on ppc64

21 years ago[PATCH] Fix broken PIIX build
Jeff Garzik [Thu, 26 Feb 2004 18:42:06 +0000 (10:42 -0800)]
[PATCH] Fix broken PIIX build

Fix the build.

Pass me a brown paper bag.

21 years agoMerge bk://kernel.bkbits.net/davem/sparc-2.6
Linus Torvalds [Thu, 26 Feb 2004 15:33:17 +0000 (07:33 -0800)]
Merge bk://kernel.bkbits.net/davem/sparc-2.6
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years agoMerge bk://gkernel.bkbits.net/libata-2.5
Linus Torvalds [Thu, 26 Feb 2004 14:58:43 +0000 (06:58 -0800)]
Merge bk://gkernel.bkbits.net/libata-2.5
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years ago[libata ata_piix] Update PCI quirk with new Intel SATA devices
Jeff Garzik [Thu, 26 Feb 2004 17:36:03 +0000 (12:36 -0500)]
[libata ata_piix] Update PCI quirk with new Intel SATA devices

And also make a note in ata_piix.c so this is not forgotten
in the future.

21 years ago[libata ata_piix] add support for ICH6
Jeff Garzik [Thu, 26 Feb 2004 16:45:39 +0000 (11:45 -0500)]
[libata ata_piix] add support for ICH6

21 years ago[libata] limit S/G table size to 128 entries
Jeff Garzik [Thu, 26 Feb 2004 16:26:13 +0000 (11:26 -0500)]
[libata] limit S/G table size to 128 entries

Agreeing with Ben H's recommendation, I reduce the max-hw-segments
number to 128, to match max-phys-segments number.  This shouldn't
affect performance because 128*64K is far above the max transfer size
for most current IDE devices.  Even 128*4K is OK.

21 years ago[PATCH] fbdev cursor fix
Andrew Morton [Thu, 26 Feb 2004 14:52:21 +0000 (06:52 -0800)]
[PATCH] fbdev cursor fix

From: James Simmons <jsimmons@infradead.org>

This patch creates a separate cursor image drawing region and regular drawing
region.

21 years ago[PATCH] rename other MODULE_VERSION users
Andrew Morton [Thu, 26 Feb 2004 14:52:09 +0000 (06:52 -0800)]
[PATCH] rename other MODULE_VERSION users

Avoid cpp clash with the new MODULE_VERSION

21 years ago[PATCH] Add a MODULE_VERSION macro
Andrew Morton [Thu, 26 Feb 2004 14:51:58 +0000 (06:51 -0800)]
[PATCH] Add a MODULE_VERSION macro

From: Rusty Russell <rusty@au1.ibm.com>

The way it works is that the .mod file contains the name of the module (as
before), but succeeding lines are the constituent parts (assumed to be .c
files, which usually works: if they use MODULE_VERSION in a file for which
this isn't true we'll get a warning).

As we postprocess modules, we look in the .modinfo section for a
"version=", which is placed by the MODULE_VERSION() macro.  This will be of
form "version=<macroarg>" "\0" [24 chars] "\0".  The 24 chars are replaced
by the md4 sum of the .c files and any files they #include using '#include
"file"' which are found in the current directory.  Whitespace is collapsed
outside strings, and comments are ignored for purposes of the sum.

The result is a .modinfo entry such as

version=1.16ac-rustytest B13E9451C4CA3B89577DEFF

At the kernel summit, various people asked for a MODULE_VERSION macro to
store module strings (for later access through sysfs).  A simple md4 is
needed to identify changes in modules which, inevitably, do not update the
version.  It skips whitespace and comments, and includes #includes which
are in the same dir.

The module versions should be set according to this definition, based on
the RPM one, or CVS Revision tags.  Violators will be shot.

 [<epoch>`:']<version>[`-'<extraversion>]
 <epoch>: A (small) unsigned integer which allows you to start versions
          anew. If not mentioned, it's zero.  eg. "2:1.0" is after
     "1:2.0".
 <version>: The <version> may contain only alphanumerics.
 <extraversion>: Like <version>, but inserted for local
          customizations, eg "rh3" or "rusty1".

Comparison of two versions (assuming same epoch):

Split each into all-digit and all-alphabetical parts.  Compare each one one
at a time: digit parts numerically, alphabetical in ASCII order.  So 0.10
comes after 0.9.

21 years ago[PATCH] kNFSd: return more than one page of directory entries.
Andrew Morton [Thu, 26 Feb 2004 14:51:46 +0000 (06:51 -0800)]
[PATCH] kNFSd: return more than one page of directory entries.

From: NeilBrown <neilb@cse.unsw.edu.au>

From: Bruce Allan <bwa@us.ibm.com>

This requires checking if the entry will fit in the current page and, if it
doesn't, encoding it into the next page, and then copying it back to lie
across the boundary between two pages.

21 years ago[PATCH] kNFSd: make sure sunrpc init routines called before gss init routines.
Andrew Morton [Thu, 26 Feb 2004 14:51:34 +0000 (06:51 -0800)]
[PATCH] kNFSd: make sure sunrpc init routines called before gss init routines.

From: NeilBrown <neilb@cse.unsw.edu.au>

make sure sunrpc init routines called before gss init routines.

21 years ago[PATCH] kNFSd: Make the calculation in the first READ_BUF easier to understand.
Andrew Morton [Thu, 26 Feb 2004 14:51:23 +0000 (06:51 -0800)]
[PATCH] kNFSd: Make the calculation in the first READ_BUF easier to understand.

From: NeilBrown <neilb@cse.unsw.edu.au>

Don't rely on the assumption that
    sizeof(stateid_t) == sizeof(stateid_opaque_t) + 4.

21 years ago[PATCH] kNFSd: Fix out-of-spec error return in attribute decoding.
Andrew Morton [Thu, 26 Feb 2004 14:51:12 +0000 (06:51 -0800)]
[PATCH] kNFSd: Fix out-of-spec error return in attribute decoding.

From: NeilBrown <neilb@cse.unsw.edu.au>

Fix out-of-spec error return in attribute decoding.

21 years ago[PATCH] kNFSd: Miscellaneous fixes to stateid-based replay
Andrew Morton [Thu, 26 Feb 2004 14:51:01 +0000 (06:51 -0800)]
[PATCH] kNFSd: Miscellaneous fixes to stateid-based replay

From: NeilBrown <neilb@cse.unsw.edu.au>

Save replays for LOCK operations, make sure errors are replayed correctly.

21 years ago[PATCH] kNFSd: Fix an out-of-spec error in nfsd4_remove.
Andrew Morton [Thu, 26 Feb 2004 14:50:50 +0000 (06:50 -0800)]
[PATCH] kNFSd: Fix an out-of-spec error in nfsd4_remove.

From: NeilBrown <neilb@cse.unsw.edu.au>

Fix an out-of-spec error in nfsd4_remove.

21 years ago[PATCH] kNFSd: Fix an out-of-spec readlink error return.
Andrew Morton [Thu, 26 Feb 2004 14:50:39 +0000 (06:50 -0800)]
[PATCH] kNFSd: Fix an out-of-spec readlink error return.

From: NeilBrown <neilb@cse.unsw.edu.au>

Fix an out-of-spec readlink error return.

21 years ago[PATCH] kNFSd: Enforce open_downgrade requirement
Andrew Morton [Thu, 26 Feb 2004 14:50:27 +0000 (06:50 -0800)]
[PATCH] kNFSd: Enforce open_downgrade requirement

From: NeilBrown <neilb@cse.unsw.edu.au>

The spec requires that the access and deny bits requested in open_downgrade
represent the union of the bits for some subset of the OPENS for the given
open_owner and file.  Enforce that requirement.

21 years ago[PATCH] kNFSd: Corrects an error return for OP_OPEN_CONFIRM.
Andrew Morton [Thu, 26 Feb 2004 14:50:16 +0000 (06:50 -0800)]
[PATCH] kNFSd: Corrects an error return for OP_OPEN_CONFIRM.

From: NeilBrown <neilb@cse.unsw.edu.au>

Corrects an error return for OP_OPEN_CONFIRM.

21 years ago[PATCH] kNFSd: Add a check in OP_LOCK for new lockowners to ensure that the open...
Andrew Morton [Thu, 26 Feb 2004 14:50:05 +0000 (06:50 -0800)]
[PATCH] kNFSd: Add a check in OP_LOCK for new lockowners to ensure that the open stateid is

From: NeilBrown <neilb@cse.unsw.edu.au>

owned by the lockowner clientid.

21 years ago[PATCH] kNFSd: fix an error return for OP_CREATE
Andrew Morton [Thu, 26 Feb 2004 14:49:54 +0000 (06:49 -0800)]
[PATCH] kNFSd: fix an error return for OP_CREATE

From: NeilBrown <neilb@cse.unsw.edu.au>

fix an error return for OP_CREATE

21 years ago[PATCH] kNFSd: add OP_ILLEGAL, and fix processing of compounds with out of bounds...
Andrew Morton [Thu, 26 Feb 2004 14:49:43 +0000 (06:49 -0800)]
[PATCH] kNFSd: add OP_ILLEGAL, and fix processing of compounds with out of bounds op numbers.

From: NeilBrown <neilb@cse.unsw.edu.au>

add OP_ILLEGAL, and fix processing of compounds with out of bounds op
numbers.

21 years ago[PATCH] kNFSd: Implement the nfsv4 RELEASE_LOCKOWNER operation.
Andrew Morton [Thu, 26 Feb 2004 14:49:31 +0000 (06:49 -0800)]
[PATCH] kNFSd: Implement the nfsv4 RELEASE_LOCKOWNER operation.

From: NeilBrown <neilb@cse.unsw.edu.au>

Implement the nfsv4 RELEASE_LOCKOWNER operation.

21 years ago[PATCH] kNFSd: move fh_dup2 and fix it
Andrew Morton [Thu, 26 Feb 2004 14:49:20 +0000 (06:49 -0800)]
[PATCH] kNFSd: move fh_dup2 and fix it

From: NeilBrown <neilb@cse.unsw.edu.au>

fh_dup2 duplicates an entire svc_fh structure, so in addition to dget()'ing
the dentry, it also needs to bump the reference count on fh_export.

Moved it out of the header file into nfsfh.c for the simple reason that I
couldn't figure out how to reorganize the headers files right so that the
fh_dup2 could use the definition of svc_fh.  Since fh_dup2 is never actually
used outside of nfs4proc.c, it could just as well be moved there....

21 years ago[PATCH] kNFSd: Remove a comment that is no longer accurate
Andrew Morton [Thu, 26 Feb 2004 14:49:09 +0000 (06:49 -0800)]
[PATCH] kNFSd: Remove a comment that is no longer accurate

From: NeilBrown <neilb@cse.unsw.edu.au>

Remove a comment that is no longer accurate

21 years ago[PATCH] kNFSd: unlock-on-close fix
Andrew Morton [Thu, 26 Feb 2004 14:48:58 +0000 (06:48 -0800)]
[PATCH] kNFSd: unlock-on-close fix

From: NeilBrown <neilb@cse.unsw.edu.au>

add the bookeeping necessary to remove all locks held by an nfsv4 lockowner
upon CLOSE, or upon state expiration.  calls locks_remove_posix().

replace list_del_init() with list_del on nfsv4 state structures that are
being reaped.

21 years ago[PATCH] kNFSd: correct rename error returns.
Andrew Morton [Thu, 26 Feb 2004 14:48:48 +0000 (06:48 -0800)]
[PATCH] kNFSd: correct rename error returns.

From: NeilBrown <neilb@cse.unsw.edu.au>

unlike v2/v3, nfsv4 returns nfserr_exist in some situations where the
underlying filesystem returns nfserr_isdir, nfserr_notdir.

on rename, nfsv4 returns nfserr_notdir instead of nfserr_symlink.

21 years ago[PATCH] kNFSd: check lock length, return appropriate error
Andrew Morton [Thu, 26 Feb 2004 14:48:37 +0000 (06:48 -0800)]
[PATCH] kNFSd: check lock length, return appropriate error

From: NeilBrown <neilb@cse.unsw.edu.au>

check lock length, return appropriate error

21 years ago[PATCH] kNFSd: correct symlink related error returns.
Andrew Morton [Thu, 26 Feb 2004 14:48:26 +0000 (06:48 -0800)]
[PATCH] kNFSd: correct symlink related error returns.

From: NeilBrown <neilb@cse.unsw.edu.au>

unlike v2/v3, nfsv4 returns nfserr_inval when attempting to read, write,
commit or test lock a symlink.  nfsv4 does return nfserr_symlink on lookup
and open, so a simple fix in fh_verify() will not work.

21 years ago[PATCH] kNFSd: fixes an xdr error by removing the verifier from error return.
Andrew Morton [Thu, 26 Feb 2004 14:48:15 +0000 (06:48 -0800)]
[PATCH] kNFSd: fixes an xdr error by removing the verifier from error return.

From: NeilBrown <neilb@cse.unsw.edu.au>

21 years ago[PATCH] kNFSd: Correct error returns.
Andrew Morton [Thu, 26 Feb 2004 14:48:04 +0000 (06:48 -0800)]
[PATCH] kNFSd: Correct error returns.

From: NeilBrown <neilb@cse.unsw.edu.au>

Call decode_fattr on writable attributes to check for xdr errors, incorrect
utf8, etc.

21 years ago[PATCH] kNFSd: Fix for lookup-parent at pseudo root
Andrew Morton [Thu, 26 Feb 2004 14:47:53 +0000 (06:47 -0800)]
[PATCH] kNFSd: Fix for lookup-parent at pseudo root

From: NeilBrown <neilb@cse.unsw.edu.au>

tests if current_fh is the pseudo root for the client and returns
nfserr_noent if so.  need to call exp_pseudoroot because different clients
can have

21 years ago[PATCH] kNFSd: correctly tests and sets nfserr_nofilehandle for current and save fh.
Andrew Morton [Thu, 26 Feb 2004 14:47:42 +0000 (06:47 -0800)]
[PATCH] kNFSd: correctly tests and sets nfserr_nofilehandle for current and save fh.

From: NeilBrown <neilb@cse.unsw.edu.au>

nfsd4_restore_fh() returns nfserr_restorefh instead of nfserr_nofilehandle

21 years ago[PATCH] kNFSd: readdir error code fix
Andrew Morton [Thu, 26 Feb 2004 14:47:30 +0000 (06:47 -0800)]
[PATCH] kNFSd: readdir error code fix

From: NeilBrown <neilb@cse.unsw.edu.au>

changes nfserr_readdir_nospc to nfserr_toosmall, following rfc3530 which has
no nfserr_readdir_nospc.

error 10030 which was nfserr_readdir_nospc is actually nfserr_restorefh.

21 years ago[PATCH] kNFSd: When looking for a shareowner in the nfsd open, make sure we don't...
Andrew Morton [Thu, 26 Feb 2004 14:47:19 +0000 (06:47 -0800)]
[PATCH] kNFSd: When looking for a shareowner in the nfsd open, make sure we don't get a lockowner instead.

From: NeilBrown <neilb@cse.unsw.edu.au>

When looking for a shareowner in the nfsd open, make sure we don't get a
lockowner instead.

21 years ago[PATCH] kNFSd: Use only the uid when deciding whether a setclientid is being done...
Andrew Morton [Thu, 26 Feb 2004 14:47:08 +0000 (06:47 -0800)]
[PATCH] kNFSd: Use only the uid when deciding whether a setclientid is being done with the "same principal".

From: NeilBrown <neilb@cse.unsw.edu.au>

I would have assumed that we should also check for the same security
(pseudo)flavor, but that doesn't seem to be how Solaris does it, and since
the spec doesn't suggest including such information in the clientid, that
would make it impossible to switch pseudoflavors.

21 years ago[PATCH] kNFSd: NFSdV4 fixes for replaying open requests.
Andrew Morton [Thu, 26 Feb 2004 14:46:54 +0000 (06:46 -0800)]
[PATCH] kNFSd: NFSdV4 fixes for replaying open requests.

From: NeilBrown <neilb@cse.unsw.edu.au>

Since the open op changes the current filehandle, we can't correctly replay
compounds containing opens unless we save the filehandle resulting from the
open as well as the encoded reply.

21 years ago[PATCH] kNFSd: When looking for a shareowner in the nfsd open, make sure we don't...
Andrew Morton [Thu, 26 Feb 2004 14:46:43 +0000 (06:46 -0800)]
[PATCH] kNFSd: When looking for a shareowner in the nfsd open, make sure we don't get a lockowner instead.

From: NeilBrown <neilb@cse.unsw.edu.au>

When looking for a shareowner in the nfsd open, make sure we don't get a
lockowner instead.

21 years ago[PATCH] kNFSd: Use higher-resolution time for the changeinfo, instead of using time...
Andrew Morton [Thu, 26 Feb 2004 14:46:32 +0000 (06:46 -0800)]
[PATCH] kNFSd: Use higher-resolution time for the changeinfo, instead of using time and filesize.

From: NeilBrown <neilb@cse.unsw.edu.au>

Use higher-resolution time for the changeinfo, instead of using time and
filesize.

21 years ago[PATCH] kNFSd: v4 exclusive open fix.
Andrew Morton [Thu, 26 Feb 2004 14:46:20 +0000 (06:46 -0800)]
[PATCH] kNFSd: v4 exclusive open fix.

From: NeilBrown <neilb@cse.unsw.edu.au>

Exclusive creates were failing with permission errors, because they don't set
the mode on the newly created file, and the permission check in fh_verify
(called at the end of do_open_lookup) was failing.  So pass in the
MAY_OWNER_OVERRIDE flag to allow the owner of the file to override the
permission check.

21 years ago[PATCH] kNFSd: NFSv4 locking state fix
Andrew Morton [Thu, 26 Feb 2004 14:46:09 +0000 (06:46 -0800)]
[PATCH] kNFSd: NFSv4 locking state fix

From: NeilBrown <neilb@cse.unsw.edu.au>

It's OK for the find_lockstateowner_str to fail; that just means we haven't
seen the lockowner, and will compare the request range to all locks on the
file.

21 years ago[PATCH] kNFSd: Nfsdv4 pointer cleanup
Andrew Morton [Thu, 26 Feb 2004 14:45:58 +0000 (06:45 -0800)]
[PATCH] kNFSd: Nfsdv4 pointer cleanup

From: NeilBrown <neilb@cse.unsw.edu.au>

From: "William A.(Andy) Adamson" <andros@citi.umich.edu>

Compare inode pointers instead of dentry pointers when checking an inode
refered to by a stateid against an inode refered to by a file handle.

Set st_vfs_set to 0 if the stateid/file handle check fails to avoid
referencing bad state when reaping stateid's.

21 years ago[PATCH] kNFSd: IDmap support for the NFSv4 server.
Andrew Morton [Thu, 26 Feb 2004 14:45:47 +0000 (06:45 -0800)]
[PATCH] kNFSd: IDmap support for the NFSv4 server.

From: NeilBrown <neilb@cse.unsw.edu.au>

Updated version which uses ascii-encoding of messages, from
http://www.citi.umich.edu/u/marius/linux-2.5.70-idmap-server-new.diff as of
October 14, 2003.

21 years ago[PATCH] kNFSd: gss api changes for integrity checking.
Andrew Morton [Thu, 26 Feb 2004 14:45:35 +0000 (06:45 -0800)]
[PATCH] kNFSd: gss api changes for integrity checking.

From: NeilBrown <neilb@cse.unsw.edu.au>

Doesn't actually add integrity support on the server yet; just adapts
server-side code to the gss api changes necessary to get integrity working on
the client.

21 years ago[PATCH] kNFSd: Add minimal server-side support for rpcsec_gss.
Andrew Morton [Thu, 26 Feb 2004 14:45:25 +0000 (06:45 -0800)]
[PATCH] kNFSd: Add minimal server-side support for rpcsec_gss.

From: NeilBrown <neilb@cse.unsw.edu.au>

From: "J. Bruce Fields" <bfields@fieldses.org>

Note that the user (or exportfs, on the user's behalf) allows a gss
pseudoflavor to be used to access an export by exporting to a special client
named "gss/pseudoflavor-name", e.g., "gss/krb5" or "gss/lipkey-i".

21 years ago[PATCH] nfsd: don't modify group_info structures
Andrew Morton [Thu, 26 Feb 2004 14:45:16 +0000 (06:45 -0800)]
[PATCH] nfsd: don't modify group_info structures

From: Neil Brown <neilb@cse.unsw.edu.au>

Make sure nfsd doesn't modify group_info structures as they might be
shared.

21 years ago[PATCH] knfsd: NGROUPS fixes
Andrew Morton [Thu, 26 Feb 2004 14:45:06 +0000 (06:45 -0800)]
[PATCH] knfsd: NGROUPS fixes

From: Neil Brown <neilb@cse.unsw.edu.au>

Tidy up new groups handling in nfsd.

Set up the group_info structure when decoding the RPC packet instead of in
nfsd.

21 years ago[PATCH] ppc64: Fix a sleeping with spinlock bug in ioremap
Andrew Morton [Thu, 26 Feb 2004 14:44:54 +0000 (06:44 -0800)]
[PATCH] ppc64: Fix a sleeping with spinlock bug in ioremap

From: Benjamin Herrenschmidt <benh@kernel.crashing.org>

ppc64 uses its own range allocator for ioremap (in order to allocate
things in a different space than normal vmalloc).  This is historic stuff,
we may get rid of it, but in the meantime, here's a patch turning the
spinlock in there into a semaphore so it doesn't blow up when doing
kmallocs.

21 years ago[PATCH] ppc64: move sg_dma_{len,address} macros
Andrew Morton [Thu, 26 Feb 2004 14:44:43 +0000 (06:44 -0800)]
[PATCH] ppc64: move sg_dma_{len,address} macros

From: Anton Blanchard <anton@samba.org>

From: Stephen Rothwell <sfr@canb.auug.org.au>

This moves the sg_dma_* macros to asm/scatterlist.h where they belong (and
where almost every other architecture has them).

21 years ago[PATCH] ppc64: use kallsyms_lookup_name() in xmon
Andrew Morton [Thu, 26 Feb 2004 14:44:29 +0000 (06:44 -0800)]
[PATCH] ppc64: use kallsyms_lookup_name() in xmon

From: olof@austin.ibm.com

Use the new kallsyms_lookup_name() in xmon on ppc64.  Ben said he might
backport these changes to ppc32 as well.

21 years ago[PATCH] Add kallsyms_lookupname()
Andrew Morton [Thu, 26 Feb 2004 14:44:18 +0000 (06:44 -0800)]
[PATCH] Add kallsyms_lookupname()

From: Rusty Russell <rusty@au1.ibm.com>

Attached patch adds a kallsyms_lookupname() function for lookups of a
symbol name to an address.  Debuggers such as ppc[64] xmon can use this.

It's intentionally not exported as a symbol for module use, since it
can be used to circumvent other symbol export restrictions.

21 years ago[PATCH] PPC64 iSeries virtual disk driver
Andrew Morton [Thu, 26 Feb 2004 14:44:07 +0000 (06:44 -0800)]
[PATCH] PPC64 iSeries virtual disk driver

From: Stephen Rothwell <sfr@canb.auug.org.au>

Add the PPC64 iSeries virtual disk driver

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/sparc-2.6
David S. Miller [Thu, 26 Feb 2004 14:32:27 +0000 (06:32 -0800)]
Merge davem@nuts.davemloft.net:/disk1/BK/sparc-2.6
into kernel.bkbits.net:/home/davem/sparc-2.6

21 years ago[SPARC]: Move ptrace_signal_deliver() implementation out of header file.
David S. Miller [Thu, 26 Feb 2004 14:28:02 +0000 (06:28 -0800)]
[SPARC]: Move ptrace_signal_deliver() implementation out of header file.

21 years ago[ARM] Add ARM architecture version 6 support.
Russell King [Thu, 26 Feb 2004 14:24:43 +0000 (14:24 +0000)]
[ARM] Add ARM architecture version 6 support.

This cset adds support ARM architecture version 6.

21 years agoMerge bk://gkernel.bkbits.net/misc-2.5
Linus Torvalds [Thu, 26 Feb 2004 14:23:02 +0000 (06:23 -0800)]
Merge bk://gkernel.bkbits.net/misc-2.5
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years agoAdd Intel PCI ids to IDE (PATA) driver.
Jeff Garzik [Thu, 26 Feb 2004 20:01:52 +0000 (15:01 -0500)]
Add Intel PCI ids to IDE (PATA) driver.

21 years agoAdd Intel PCI ids to old-OSS driver i810_audio
Jeff Garzik [Thu, 26 Feb 2004 19:54:44 +0000 (14:54 -0500)]
Add Intel PCI ids to old-OSS driver i810_audio

21 years agoAdd Intel ICH6 irq router.
Jeff Garzik [Thu, 26 Feb 2004 19:48:18 +0000 (14:48 -0500)]
Add Intel ICH6 irq router.

21 years agoAdd Intel ICH6 PCI ids to pci_ids.h.
Jeff Garzik [Thu, 26 Feb 2004 19:43:17 +0000 (14:43 -0500)]
Add Intel ICH6 PCI ids to pci_ids.h.

21 years ago[ARM] Optimise ARM720T Thumb abort unwinding.
Andre McCurdy [Thu, 26 Feb 2004 12:15:46 +0000 (12:15 +0000)]
[ARM] Optimise ARM720T Thumb abort unwinding.

Patch from: Andre.

Optimise data_thumb_pushpop and data_thumb_ldmstm population
counting code with a slightly more optimal algorithm than the
original as there are only 8 bits to count instead of 16.

21 years ago[PATCH] Improve code generation for x86 raid XOR functions
Alexandre Oliva [Thu, 26 Feb 2004 08:10:25 +0000 (00:10 -0800)]
[PATCH] Improve code generation for x86 raid XOR functions

This helps avoid doing push/pop pairs on register contents that we just
don't care about. After we've done the xor, the pointers are dead anyway.

21 years agoMerge bk://kernel.bkbits.net/davem/net-2.6
Linus Torvalds [Thu, 26 Feb 2004 07:16:43 +0000 (23:16 -0800)]
Merge bk://kernel.bkbits.net/davem/net-2.6
into ppc970.osdl.org:/home/torvalds/v2.5/linux

21 years ago[SPARC64]: Update defconfig.
David S. Miller [Thu, 26 Feb 2004 06:44:55 +0000 (22:44 -0800)]
[SPARC64]: Update defconfig.

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Thu, 26 Feb 2004 06:44:48 +0000 (22:44 -0800)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years agoMerge bk://bk.phunnypharm.org/sparc-2.6
David S. Miller [Thu, 26 Feb 2004 06:35:23 +0000 (22:35 -0800)]
Merge bk://bk.phunnypharm.org/sparc-2.6
into nuts.davemloft.net:/disk1/BK/sparc-2.6

21 years ago[SPARC64]: Fix compile warning from RW_LOCK_UNLOCKED with spinlock debug enabled.
Ben Collins [Thu, 26 Feb 2004 10:39:56 +0000 (05:39 -0500)]
[SPARC64]: Fix compile warning from RW_LOCK_UNLOCKED with spinlock debug enabled.