]> git.neil.brown.name Git - history.git/log
history.git
22 years agoLinux 2.6.0-test6 v2.6.0-test6
Linus Torvalds [Sat, 27 Sep 2003 10:46:14 +0000 (03:46 -0700)]
Linux 2.6.0-test6

22 years agoRevert NDEBUG move in NCR5380 - g_NCR5380 includes the file (ugh) and
Linus Torvalds [Sat, 27 Sep 2003 10:38:25 +0000 (03:38 -0700)]
Revert NDEBUG move in NCR5380 - g_NCR5380 includes the file (ugh) and
wants 'phases'

Cset exclude: geert@linux-m68k.org|ChangeSet|20030912135245|42657

22 years agoMerge bk://kernel.bkbits.net/jgarzik/misc-2.5
Linus Torvalds [Sat, 27 Sep 2003 02:43:54 +0000 (19:43 -0700)]
Merge bk://kernel.bkbits.net/jgarzik/misc-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

22 years agoMerge redhat.com:/spare/repo/linux-2.5
Jeff Garzik [Sat, 27 Sep 2003 03:56:29 +0000 (23:56 -0400)]
Merge redhat.com:/spare/repo/linux-2.5
into redhat.com:/spare/repo/misc-2.5

22 years agoMerge bk://kernel.bkbits.net/jgarzik/net-drivers-2.5
Linus Torvalds [Sat, 27 Sep 2003 02:41:58 +0000 (19:41 -0700)]
Merge bk://kernel.bkbits.net/jgarzik/net-drivers-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

22 years ago[PATCH] select MII
Adrian Bunk [Sat, 27 Sep 2003 03:14:23 +0000 (23:14 -0400)]
[PATCH] select MII

The patch below switches MII to be select'ed instead of including it in
the Makefile.

Note that this patch requires a recent Linus' tree with the select CRC32
patch included.

diffstat output:

 drivers/net/Kconfig          |   11 +++++++++++
 drivers/net/Makefile         |   28 +++++++++++-----------------
 drivers/net/pcmcia/Kconfig   |    1 +
 drivers/net/tulip/Kconfig    |    1 +
 drivers/usb/net/Kconfig      |    2 ++
 drivers/usb/net/Makefile.mii |    6 ------
 6 files changed, 26 insertions(+), 23 deletions(-)

Tangential to the patch I observed a small problem (not fixed in the patch):
MII depends on NET_ETHERNET, but USB_PEGASUS and USB_USBNET depend
only on NET.

cu
Adrian

22 years ago[PATCH] select ZLIB_{IN,DE}FLATE
Adrian Bunk [Sat, 27 Sep 2003 02:40:51 +0000 (19:40 -0700)]
[PATCH] select ZLIB_{IN,DE}FLATE

Similar to the patch Brian Gerst sent for CRC32, this patch changes
ZLIB_{IN,DE}FLATE to be select'ed.

The dependencies are much better maintainable that way.

22 years ago[PATCH] wan/z8530 deadlocks
Stephen Hemminger [Sat, 27 Sep 2003 02:32:06 +0000 (22:32 -0400)]
[PATCH] wan/z8530 deadlocks

Existing code for drivers/net/wan/z8530 is riddled with self-deadlocks
and irq flag confusion.  For example:
z8530_init -> do_z8530_init -> write_zsreg
self deadlocks on the channel lock.

Several places acquire both the channel and dma lock and then
reuse the same irq flags variable - ouch.

This code at least, correctly probes (for no device case) on SMP.
Other paths verified by inspection.

22 years ago[PATCH] sealevel -- syncppp startup fix
Stephen Hemminger [Sat, 27 Sep 2003 02:31:59 +0000 (22:31 -0400)]
[PATCH] sealevel -- syncppp startup fix

The sealevel driver called sppp_attach before checking that board existed
and never called detach in the error path.

My change is to call sppp_attach from the netdev->init hook which happens
later in the process, and call detach from the uninit hook.

Also, changed the structure element 'netdev' to 'pppdev' to avoid confusion.

Here is the fix against 2.6.0-test5 latest

22 years ago[netdrvr smc91c92_cs] select proper bank for MII registers
Jun Komuro [Sat, 27 Sep 2003 02:30:42 +0000 (22:30 -0400)]
[netdrvr smc91c92_cs] select proper bank for MII registers

22 years ago[netdrvr] build fixes
Jun Komuro [Sat, 27 Sep 2003 02:30:07 +0000 (22:30 -0400)]
[netdrvr] build fixes

For e1000 and baycom_epp.

Contributed by Jeff Garzik.

22 years ago[PATCH] janitor: e100: cleanup #includes
Randy Dunlap [Sat, 27 Sep 2003 01:52:14 +0000 (21:52 -0400)]
[PATCH] janitor: e100: cleanup #includes

Hi,
Please apply to 2.6.0-test5-current.

Thanks,
--
~Randy

From: Randy Hron <rwhron@earthlink.net>

Remove unneeded include of version.h.
Test compiled against 2.6.0-test5-bk9.

 linux-260-t5bk12-kj-rddunlap/drivers/net/e100/e100.h |    1 -
 1 files changed, 1 deletion(-)

22 years ago[PATCH] janitor: hermes: delete verify_area call
Randy Dunlap [Sat, 27 Sep 2003 01:52:07 +0000 (21:52 -0400)]
[PATCH] janitor: hermes: delete verify_area call

Hi,
Please apply to 2.6.0-test5-current.

Thanks,
--
~Randy

From: Domen Puncer <domen@coderock.org>

IMO, that verify_area wasn't needed.

 linux-260-t5bk12-kj-rddunlap/drivers/net/wireless/orinoco.c |    4 ----
 1 files changed, 4 deletions(-)

22 years ago[e100] trying to pci_alloc before pci_enable
Scott Feldman [Sat, 27 Sep 2003 01:52:01 +0000 (21:52 -0400)]
[e100] trying to pci_alloc before pci_enable

* Some archs don't like calling pci_alloc_consistent before calling
pci_enable_device.  (Imagine that).  This corrects that.

22 years ago[e100] h/w can't do IPv6 checksum offloading
Scott Feldman [Sat, 27 Sep 2003 01:51:54 +0000 (21:51 -0400)]
[e100] h/w can't do IPv6 checksum offloading

* Driver was advertising HW_CSUM, but hardware is only capable of IP_CSUM.

22 years ago[e100] PRO/10+ not configured properly
Scott Feldman [Sat, 27 Sep 2003 01:51:47 +0000 (21:51 -0400)]
[e100] PRO/10+ not configured properly

* PRO/10+ (10 Mbps-only card) was not configured properly so it didn't
pass traffic.

22 years ago[e1000] misc
Scott Feldman [Sat, 27 Sep 2003 01:51:41 +0000 (21:51 -0400)]
[e1000] misc

* removed unused var, ASSERT macro
* missed a free_netdev() in probe cleanup undo.

22 years ago[e1000] better propagation of error codes
Scott Feldman [Sat, 27 Sep 2003 01:51:34 +0000 (21:51 -0400)]
[e1000] better propagation of error codes

* Better propagation of error codes during probe/open paths.
  {Janice Girouard (janiceg@us.ibm.com)]

22 years ago[e1000] force 1000/full on SERDES connected to back-plane
Scott Feldman [Sat, 27 Sep 2003 01:51:28 +0000 (21:51 -0400)]
[e1000] force 1000/full on SERDES connected to back-plane

* Bug fix: SERDES devices might be connected to back-plan switch that
  doesn't support auto-neg, so add the capability to force 1000/full.

22 years ago[e1000] flow control updates
Scott Feldman [Sat, 27 Sep 2003 01:51:21 +0000 (21:51 -0400)]
[e1000] flow control updates

* handle ethtool force flow control
* correctly set flow control hi/low watermarks based on size of Rx FIFO
  area.  The size can change if doing Jumbo Frames or, in the case of
  82547, is smaller to start with.
* was not properly forcing flow control settings to fc_none if using
  strict IEEE flow control override.

22 years ago[netdrvr sk98lin] use seq_file for /proc
Stephen Hemminger [Sat, 27 Sep 2003 01:51:14 +0000 (21:51 -0400)]
[netdrvr sk98lin] use seq_file for /proc

Replace proc_read with seq_file interface which is much cleaner.

Also, instead of searching the list of devices looking for the
name which is slow and would break if interface name changes;
store a pointer to device in the proc entry, and retrieve
it from the information saved by single_open.

Formatting and other behaviours are retained.

22 years ago[netdrvr skge] handle proc_fs errors.
Stephen Hemminger [Sat, 27 Sep 2003 01:51:08 +0000 (21:51 -0400)]
[netdrvr skge] handle proc_fs errors.

Existing code does not
- check if proc_fs functions return error.
- use proc_mkdir to make the directory.

Note pSkRootDir defined twice.

22 years ago[netdrvr sk98lin] build on smp fix
Stephen Hemminger [Sat, 27 Sep 2003 01:51:01 +0000 (21:51 -0400)]
[netdrvr sk98lin] build on smp fix

There is no exported variable called smp_num_cpus! so this driver
won't build on SMP.  Since the local variable is never used anyway
just get rid of it.

This applies against 2.6.0-test5-bk13 which has the last vendor
driver update.

22 years ago[PATCH] ppc64 typo fix (kudos to Anton)
Alexander Viro [Fri, 26 Sep 2003 17:07:17 +0000 (10:07 -0700)]
[PATCH] ppc64 typo fix (kudos to Anton)

*duh*

22 years ago[PATCH] select CRC32
Brian Gerst [Fri, 26 Sep 2003 17:07:08 +0000 (10:07 -0700)]
[PATCH] select CRC32

Use "select CRC32" in Kconfig instead of makefile includes.

22 years agoMerge http://lia64.bkbits.net/to-linus-2.5
Linus Torvalds [Fri, 26 Sep 2003 14:20:15 +0000 (07:20 -0700)]
Merge http://lia64.bkbits.net/to-linus-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

22 years agoMerge bk://linux-scsi.bkbits.net/scsi-for-linus-2.6
Linus Torvalds [Fri, 26 Sep 2003 14:19:04 +0000 (07:19 -0700)]
Merge bk://linux-scsi.bkbits.net/scsi-for-linus-2.6
into home.osdl.org:/home/torvalds/v2.5/linux

22 years agoMerge bk://ppc.bkbits.net/for-linus-ppc
Linus Torvalds [Fri, 26 Sep 2003 14:17:24 +0000 (07:17 -0700)]
Merge bk://ppc.bkbits.net/for-linus-ppc
into home.osdl.org:/home/torvalds/v2.5/linux

22 years agoPPC32: Make scripts/mkuboot.sh executable
Paul Mackerras [Sat, 27 Sep 2003 21:40:10 +0000 (07:40 +1000)]
PPC32: Make scripts/mkuboot.sh executable

22 years agoPPC32: Fix the definition of PMU_IOC_GRAB_BACKLIGHT.
Paul Mackerras [Sat, 27 Sep 2003 21:32:44 +0000 (07:32 +1000)]
PPC32: Fix the definition of PMU_IOC_GRAB_BACKLIGHT.

22 years agoPPC32: Fix up the CPU frequency scaling questions in arch/ppc/Kconfig.
Paul Mackerras [Sat, 27 Sep 2003 21:25:58 +0000 (07:25 +1000)]
PPC32: Fix up the CPU frequency scaling questions in arch/ppc/Kconfig.

This is from Ben Herrenschmidt's tree.

22 years agoMerge bk://stop.crashing.org/linux-2.6-uImage
Paul Mackerras [Sat, 27 Sep 2003 20:11:30 +0000 (06:11 +1000)]
Merge bk://stop.crashing.org/linux-2.6-uImage
into samba.org:/home/paulus/kernel/for-linus-ppc

22 years agoMerge changes from Tom Rini
Paul Mackerras [Sat, 27 Sep 2003 19:46:47 +0000 (05:46 +1000)]
Merge changes from Tom Rini

22 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Sat, 27 Sep 2003 19:08:58 +0000 (05:08 +1000)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

22 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Sat, 27 Sep 2003 09:08:45 +0000 (19:08 +1000)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

22 years agoPPC32: Adjust ucontext to conform with historical practice and with glibc.
Paul Mackerras [Sat, 27 Sep 2003 09:07:22 +0000 (19:07 +1000)]
PPC32: Adjust ucontext to conform with historical practice and with glibc.

This moves the uc_sigmask field back to where it was before.  We now don't
assume the registers are at a fixed offset in the ucontext, but instead
access them through the uc_regs field as in the past.  This corresponds with
a recent change in glibc CVS.

22 years agoMerge
Paul Mackerras [Sat, 27 Sep 2003 07:31:46 +0000 (17:31 +1000)]
Merge

22 years ago[PATCH] scsi_mid_low_api.txt update
James Bottomley [Fri, 26 Sep 2003 14:11:05 +0000 (09:11 -0500)]
[PATCH] scsi_mid_low_api.txt update

From: Douglas Gilbert <dougg@torque.net>

Attached is an update of the scsi_mid_low_api text
document. The first attachment is a gzipped patch against
lk 2.6.0-test4 and the second one is the file gzipped.

Changes:
   - add "Calling context:" entry to interface functions
   - add "Might block:" entry to mid level functions
   - drop scsi_set_device_offline() + command()
   - change references to reflect transfer of headers to
     include/scsi/scsi_*.h
   - try to define what a "scsi host" is
   - typos + grammar

Thanks to Randy Dunlap, Alan Stern amd Christoph Hellwig
for their input. Comments welcome.

22 years agoMerge raven.il.steeleye.com:/home/jejb/BK/scsi-misc-2.5
James Bottomley [Fri, 26 Sep 2003 13:27:48 +0000 (08:27 -0500)]
Merge raven.il.steeleye.com:/home/jejb/BK/scsi-misc-2.5
into raven.il.steeleye.com:/home/jejb/BK/scsi-for-linus-2.6

22 years ago[PATCH] GCC 3.3.x/3.4 compatiblity fix in include/linux/init.h
Bernardo Innocenti [Fri, 26 Sep 2003 10:58:02 +0000 (03:58 -0700)]
[PATCH] GCC 3.3.x/3.4 compatiblity fix in include/linux/init.h

GCC 3.4 miscompiles the kernel because it silently optimizes away
data placed in the .init.setup section by the __setup() macro.

__attribute__((unused)) does only avoid the warning, but doesn't
mark the data as being used.

Since GCC 3.3, __attribute__((used)) should be applied to such
variables.  The __attribute_used__ macro from linux/compiler.h already
takes care of compiler differences for us.

In this patch, I've gone a step further and proactively fixed that in
all places.

22 years ago[PATCH] fix kernel BUG using multipath
Steven Dake [Fri, 26 Sep 2003 10:55:31 +0000 (03:55 -0700)]
[PATCH] fix kernel BUG using multipath

Thanks Matt and Jens for the debug help on the multipath problem.

This solves the problem and makes multipath work properly.

There are two types of "flags" that are used in a block io request,
bi_flags, and bi_rw.  bi_flags is used for flags to the block level
code, and bi_rw is used for flags to the low level device drivers.

The code in the multipath driver used the wrong flag in the wrong field.
In this case, the flag FASTFAIL (value 3) was being set to the bi_flags
field.  FASTFAIL is a hint to the low level driver that it should try to
fail out quickly.  Unfortunately, the value 3 is also BIO_SEG_VALID,
which is a flag to the block subsystem that the segments shouldn't be
recalculated.  The result was that the wrong field was set, telling the
block layer not to recalculate the segments resulting in phys and hw
segments of 0.  Not good.

22 years ago[PATCH] sysfs dput crash fix
Patrick Mochel [Fri, 26 Sep 2003 07:58:11 +0000 (00:58 -0700)]
[PATCH] sysfs dput crash fix

This is a new patch that includes Martin's, and makes create_dir() return
an int. That matches sysfs_create_file(), and saves a couple of
conversions back and forth between pointer and error.

22 years ago[PATCH] ia64: remove CONFIG_PCI ifdefs in SN2 code
Christoph Hellwig [Fri, 26 Sep 2003 07:01:17 +0000 (00:01 -0700)]
[PATCH] ia64: remove CONFIG_PCI ifdefs in SN2 code

It's not like it would compile without PCI support anyway, and these
ifdefs are _really_ strange.

22 years ago[PATCH] ia64: simplify SN2 interrupt allocation
Christoph Hellwig [Fri, 26 Sep 2003 07:00:19 +0000 (00:00 -0700)]
[PATCH] ia64: simplify SN2 interrupt allocation

Same issues as the dma mapping code, but as SN2 interrupt handling is
totally static and all irqs are allocated on bootup an never freed it's
not a speed issue.

I really wonder what someone has to smoke to call a that expensive
dispatcher just to cast the data structure back to the lowlevel one
directly afterwards...

22 years ago[PATCH] ia64: sn_ML_intr.c is a freakin mess
Christoph Hellwig [Fri, 26 Sep 2003 06:59:31 +0000 (23:59 -0700)]
[PATCH] ia64: sn_ML_intr.c is a freakin mess

fed through Lindent and dead codepathes eliminated.

22 years ago[PATCH] ia64: simplify and speedup SN2 dma mapping
Christoph Hellwig [Fri, 26 Sep 2003 06:58:45 +0000 (23:58 -0700)]
[PATCH] ia64: simplify and speedup SN2 dma mapping

Currently the the SN2 dma mapping code uses a file systen lookup (cat /r/napali/tmp/p10 |diffstat -p1) to
find the operation vector for the underlying dma mapping implementation.

But not only there is exactly one of those implementation, also data
structures from this lowlevel driver are used in the actual Linux dma
mapping routines and quite a few assumption are made about it.  So let's
just call directly in the lowlevel code and give up this utter bullshit.

22 years ago[PATCH] ia64: kill snia_pciio_*
Christoph Hellwig [Fri, 26 Sep 2003 06:57:27 +0000 (23:57 -0700)]
[PATCH] ia64: kill snia_pciio_*

These strange mixups of Linux and IRIX PCI interfaces were used in SGI's
2.4 tree for their port of the IRIX scsi code.  They have no business to
stgay in 2.6.

22 years agoia64: Patch by Christoph Hellwig: kill .hcl entry in SN hwgfs.
David Mosberger [Fri, 26 Sep 2003 06:56:41 +0000 (23:56 -0700)]
ia64: Patch by Christoph Hellwig: kill .hcl entry in SN hwgfs.

22 years ago[PATCH] ia64: kill dead SN code from ml_iograph.c
Christoph Hellwig [Fri, 26 Sep 2003 06:54:15 +0000 (23:54 -0700)]
[PATCH] ia64: kill dead SN code from ml_iograph.c

More IRIX creepism

22 years ago[PATCH] ia64: kill sn2 inventory stubs
Christoph Hellwig [Fri, 26 Sep 2003 06:53:14 +0000 (23:53 -0700)]
[PATCH] ia64: kill sn2 inventory stubs

Just some crappy stubs that sneaked over from IRIX.

22 years ago[PATCH] ia64: kill SN kdba_io.c
Christoph Hellwig [Fri, 26 Sep 2003 06:52:24 +0000 (23:52 -0700)]
[PATCH] ia64: kill SN kdba_io.c

Keith shared my opionion that it's completly useless.  Also everyone
seems to favour keeping kdb stuff out of mainline.

22 years agoia64: Patch by Christoph Hellwig: SN2 stopped abusing devfs in 2.5, clean
David Mosberger [Fri, 26 Sep 2003 06:51:29 +0000 (23:51 -0700)]
ia64: Patch by Christoph Hellwig: SN2 stopped abusing devfs in 2.5, clean
up the leftovers.

22 years agoRemove killed SCSI_IOCTL_TAGGED_{ENABLE|DISABLE} from compat_ioctl.h
James Bottomley [Fri, 26 Sep 2003 06:50:05 +0000 (01:50 -0500)]
Remove killed SCSI_IOCTL_TAGGED_{ENABLE|DISABLE} from compat_ioctl.h

22 years agoia64: Patch by Christoph Hellwig: None of the exported symbols is referenced
David Mosberger [Fri, 26 Sep 2003 06:43:36 +0000 (23:43 -0700)]
ia64: Patch by Christoph Hellwig: None of the exported symbols is referenced
by a module, even more the file doesn't compile when
CONFIG_IA64_SGI_SN_DEBUG is set.

22 years agoia64: Patch by Christoph Hellwig: Kill two SN headers never references in
David Mosberger [Fri, 26 Sep 2003 06:35:52 +0000 (23:35 -0700)]
ia64: Patch by Christoph Hellwig: Kill two SN headers never references in
the current tree.

22 years ago[PATCH] fix qla1280 compiles
Christoph Hellwig [Fri, 26 Sep 2003 06:10:59 +0000 (01:10 -0500)]
[PATCH] fix qla1280 compiles

Duh!  The scsi_cmnd.flags removal broke qla1280.

It checked for the flags beeing 0, which it always was for a long
time.  It looks a bit fishy thus, so if someone wants to audit
srb.flags audit please do so.

22 years ago[PATCH] janitor: cleanup includes in sym53c416
Randy Dunlap [Fri, 26 Sep 2003 06:04:44 +0000 (01:04 -0500)]
[PATCH] janitor: cleanup includes in sym53c416

From: Randy Hron <rwhron@earthlink.net>

Remove unneeded version.h include.

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/sym53c416.c |    1 -
 1 files changed, 1 deletion(-)

22 years ago[PATCH] janitor: cleanup includes in osst
Randy Dunlap [Fri, 26 Sep 2003 05:57:58 +0000 (00:57 -0500)]
[PATCH] janitor: cleanup includes in osst

From: Randy Hron <rwhron@earthlink.net>

Remove unneccessary version.h #include.

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/osst.c |    1 -
 1 files changed, 1 deletion(-)

22 years ago[PATCH] janitor: cleanup includes in megaraid
Randy Dunlap [Fri, 26 Sep 2003 05:55:53 +0000 (00:55 -0500)]
[PATCH] janitor: cleanup includes in megaraid

From: Randy Hron <rwhron@earthlink.net>

Remove unneeded version.h include in megaraid.

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/megaraid.h |    1 -
 1 files changed, 1 deletion(-)

22 years ago[PATCH] janitor: scsi/a3000: cleanup includes
Randy Dunlap [Fri, 26 Sep 2003 05:43:45 +0000 (00:43 -0500)]
[PATCH] janitor: scsi/a3000: cleanup includes

From: Randy Hron <rwhron@earthlink.net>

Remove unneeded version.h include.

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/a3000.c |    1 -
 1 files changed, 1 deletion(-)

22 years ago[PATCH] janitor: cleanup includes in in2000
Randy Dunlap [Fri, 26 Sep 2003 05:36:04 +0000 (00:36 -0500)]
[PATCH] janitor: cleanup includes in in2000

From: Randy Hron <rwhron@earthlink.net>

Remove unneeded version.h include.

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/in2000.h |    2 --
 1 files changed, 2 deletions(-)

22 years ago[PATCH] janitor: cleanup includes in dpt_i2o
Randy Dunlap [Fri, 26 Sep 2003 05:29:41 +0000 (00:29 -0500)]
[PATCH] janitor: cleanup includes in dpt_i2o

From: Randy Hron <rwhron@earthlink.net>

Remove doubled #includes.
Add linux/version.h

 linux-260-t5bk12-kj-rddunlap/drivers/scsi/dpt/dpti_i2o.h |    7 +------
 1 files changed, 1 insertion(+), 6 deletions(-)

22 years agoMark PM_DISK_PARTITION as depending on PM_DISK, so as to
Linus Torvalds [Fri, 26 Sep 2003 03:57:26 +0000 (20:57 -0700)]
Mark PM_DISK_PARTITION as depending on PM_DISK, so as to
avoid an annoying nonsense configuration question.

22 years agoAvoid compiler warning by using the proper types in "min()".
Linus Torvalds [Fri, 26 Sep 2003 03:56:28 +0000 (20:56 -0700)]
Avoid compiler warning by using the proper types in "min()".

22 years ago[PATCH] Another small x86-64 merge
Andi Kleen [Fri, 26 Sep 2003 01:56:18 +0000 (18:56 -0700)]
[PATCH] Another small x86-64 merge

 - Remove obsolete wrong do_magic prototype
 - Remove the dependency between ACPI_SLEEP and SOFTWARE_SUSPEND
   some people complained about. Works the same as i386 now.
 - Fix another gcc 3.3 warning in reboot.c
 - Default NMI watchdog to LOCAL_APIC because that works on more
   boxes

22 years ago[PATCH] janitor: serial/tx3912: remove unneeded verify_area calls
Randy Dunlap [Fri, 26 Sep 2003 01:46:43 +0000 (18:46 -0700)]
[PATCH] janitor: serial/tx3912: remove unneeded verify_area calls

From: Domen Puncer <domen@coderock.org>

22 years ago[PATCH] janitor: intermezzo: clean up #includes
Randy Dunlap [Fri, 26 Sep 2003 01:46:24 +0000 (18:46 -0700)]
[PATCH] janitor: intermezzo: clean up #includes

From: Randy Hron <rwhron@earthlink.net>

Remove duplicate #includes from fs/intermezzo.

22 years ago[PATCH] janitor: isdn: remove verify_area calls
Randy Dunlap [Fri, 26 Sep 2003 01:46:05 +0000 (18:46 -0700)]
[PATCH] janitor: isdn: remove verify_area calls

From: Domen Puncer <domen@coderock.org>

The verify_area's are not needed.

22 years ago[PATCH] janitor: cleanup includes in tc/zs
Randy Dunlap [Fri, 26 Sep 2003 01:45:48 +0000 (18:45 -0700)]
[PATCH] janitor: cleanup includes in tc/zs

From: Randy Hron <rwhron@earthlink.net>

Remove unnecessary include of version.h in drivers/tc/

22 years ago[PATCH] janitor: h8300: put_user for error handling
Randy Dunlap [Fri, 26 Sep 2003 01:45:28 +0000 (18:45 -0700)]
[PATCH] janitor: h8300: put_user for error handling

From: Domen Puncer <domen@coderock.org>

verify_area isn't needed.

22 years ago[PATCH] jantior: coda: userspace error handling
Randy Dunlap [Fri, 26 Sep 2003 01:45:12 +0000 (18:45 -0700)]
[PATCH] jantior: coda: userspace error handling

From: Domen Puncer <domen@coderock.org>

Drop verify_area, using copy*user already (instead).
Also user -EFAULT for error instead of -EINVAL.

22 years ago[PATCH] janitor: clean up newlines
Randy Dunlap [Fri, 26 Sep 2003 01:44:53 +0000 (18:44 -0700)]
[PATCH] janitor: clean up newlines

From: maximilian attems <janitor@sternwelten.at>

this patches catches obvious mistyping of new line:
/n versus \n
adds a printk loglevel to one of the erroneous lines

22 years ago[PATCH] janitor: rio_linux: user get/put_user for errors (not
Randy Dunlap [Fri, 26 Sep 2003 01:44:28 +0000 (18:44 -0700)]
[PATCH] janitor: rio_linux: user get/put_user for errors (not

From: Domen Puncer <domen@coderock.org>

22 years ago[PATCH] janitor: cleanup includes in fs/
Randy Dunlap [Fri, 26 Sep 2003 01:44:07 +0000 (18:44 -0700)]
[PATCH] janitor: cleanup includes in fs/

From: Randy Hron <rwhron@earthlink.net>

Remove duplicate #includes in fs/
Remove version.h includes where they aren't necessary.
Test compiled on 2.6.0-test5-bk9.

22 years ago[PATCH] janitor: isdn: remove unneeded verify_area calls
Randy Dunlap [Fri, 26 Sep 2003 01:43:48 +0000 (18:43 -0700)]
[PATCH] janitor: isdn: remove unneeded verify_area calls

From: Domen Puncer <domen@coderock.org>

Deleted unneeded verify_area calls.
Also commented a label.

22 years ago[PATCH] jantior: sx: use get/put_user (remove verify_area)
Randy Dunlap [Fri, 26 Sep 2003 01:43:35 +0000 (18:43 -0700)]
[PATCH] jantior: sx: use get/put_user (remove verify_area)

From: Domen Puncer <domen@coderock.org>

Use get_user/put_user to check user addressing,
don't need to use verify_area also.

22 years agoMerge bk://kernel.bkbits.net/davem/sparc-2.5
Linus Torvalds [Fri, 26 Sep 2003 01:41:09 +0000 (18:41 -0700)]
Merge bk://kernel.bkbits.net/davem/sparc-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

22 years agoMerge bk://kernel.bkbits.net/davem/net-2.5
Linus Torvalds [Fri, 26 Sep 2003 01:12:16 +0000 (18:12 -0700)]
Merge bk://kernel.bkbits.net/davem/net-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

22 years ago[PATCH] s390: remove outdated code.
Martin Schwidefsky [Fri, 26 Sep 2003 01:03:39 +0000 (18:03 -0700)]
[PATCH] s390: remove outdated code.

Remove outdated ipl records from arch/s390/boot. They are now part of
the s390-tools package.

22 years ago[PATCH] s390: documentation.
Martin Schwidefsky [Fri, 26 Sep 2003 01:03:31 +0000 (18:03 -0700)]
[PATCH] s390: documentation.

s390 documentation changes.

22 years ago[PATCH] s390: vt220 console.
Martin Schwidefsky [Fri, 26 Sep 2003 01:03:21 +0000 (18:03 -0700)]
[PATCH] s390: vt220 console.

Add support for vt220 console over sclp.

22 years ago[PATCH] s390: qeth driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:03:13 +0000 (18:03 -0700)]
[PATCH] s390: qeth driver.

 - Create symlinks between netdev and groupdev.
 - Remove initialization of device.name.
 - Call qeth_free_card on removal.
 - Remove async hsi.
 - Remove contig memusage.
 - Add check for -EFAULT to copy_from_user/copy_to_user.
 - Inlining some functions to save kernel stack space.
 - vlan header fixes.
 - Replace atomic_return_sub with atomic_add_return.

22 years ago[PATCH] s390: lcs driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:03:04 +0000 (18:03 -0700)]
[PATCH] s390: lcs driver.

 - Add type and timeout attribute.
 - Create symlinks between netdev and groupdev.
 - Remove initialization of device.name.

22 years ago[PATCH] s390: iucv driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:56 +0000 (18:02 -0700)]
[PATCH] s390: iucv driver.

 - Move iucv bus and root device initialization from netiucv to iucv.
 - Fix race condition in iucv_connect.
 - Add 'user' attribute to netiucv driver.

22 years ago[PATCH] s390: ctc driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:47 +0000 (18:02 -0700)]
[PATCH] s390: ctc driver.

 - Add type attribute.
 - Remove initialization of device.name.
 - Remove unnecessary include.

22 years ago[PATCH] s390: tape driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:38 +0000 (18:02 -0700)]
[PATCH] s390: tape driver.

 - Fix use of tape block request queue pointer.
 - Remove unnecessary includes.

22 years ago[PATCH] s390: dasd partitions.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:30 +0000 (18:02 -0700)]
[PATCH] s390: dasd partitions.

Fix bug in CMS label recognition in ibm.c

22 years ago[PATCH] s390: dasd driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:21 +0000 (18:02 -0700)]
[PATCH] s390: dasd driver.

 - Do add_disk even for unformatted devices to be able to format them.
 - Remove initialization of device.name.
 - Export some functions.

22 years ago[PATCH] s390: xpram driver.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:12 +0000 (18:02 -0700)]
[PATCH] s390: xpram driver.

 - Make xpram work on 64 bit machines.
 - Use new-style module_param.

22 years ago[PATCH] s390: Kconfig.
Martin Schwidefsky [Fri, 26 Sep 2003 01:02:03 +0000 (18:02 -0700)]
[PATCH] s390: Kconfig.

 - Use common drivers/block/Kconfig instead of own config options, move
   s390 block device config options to drivers/s390/block/Kconfig and
   include it from drivers/block/Kconfig.
 - Fix configuration combination CONFIG_IPV6=m and CONFIG_QETH=y.
 - New default configuration.

22 years ago[PATCH] s390: sysfs_create_group.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:53 +0000 (18:01 -0700)]
[PATCH] s390: sysfs_create_group.

Make use of sysfs_create_group in s390 device drivers.

22 years ago[PATCH] s390: system call restart bug.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:45 +0000 (18:01 -0700)]
[PATCH] s390: system call restart bug.

Fix restarting of system calls done by use of the execute instruction.

22 years ago[PATCH] s390: system tick misaccounting.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:37 +0000 (18:01 -0700)]
[PATCH] s390: system tick misaccounting.

Fix system tick misaccounting problem.

22 years ago[PATCH] s390: micro optimizations.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:27 +0000 (18:01 -0700)]
[PATCH] s390: micro optimizations.

 - Put cpu number to lowcore.
 - Put percpu_offset to lowcore.
 - Put current pointer to lowcore.
 - Replace barrier() with cpu_relax().

22 years ago[PATCH] s390: 31 bit compat.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:19 +0000 (18:01 -0700)]
[PATCH] s390: 31 bit compat.

 - Fix emulation of sys_sysinfo and sys_clone.
 - Add code for -ERESTART_RESTARTBLOCK in signal emulation.
 - Fix ptrace peek/poke for 31 bit programs under a 64 bit kernel.
 - Fix typos in cp_stat64.

22 years ago[PATCH] s390: common i/o layer.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:09 +0000 (18:01 -0700)]
[PATCH] s390: common i/o layer.

 - Remove initialization of device.name.
 - Don't do put_device after failed get_device in get_ccwdev_by_busid.
 - Fix read_dev_chars and read_conf_data.
 - Call interrupt function of ccw device if path verification has been started.
 - Replace atomic_return_add by atomic_add_return in qdio.
 - Use wait_event instead of homegrown wait loop.
 - Fix reestablish queue problem.
 - Add ungroup attribute to ccw_group devices and add links from each
   ccw device of a group to the group device.
 - Use BUS_ID_SIZE instead of DEVICE_ID_SIZE.
 - Delay path verification if a basic sense is required.
 - Move qdio shutdown code from qdio_free to qdio_shutdown.

22 years ago[PATCH] s390: arch fixes.
Martin Schwidefsky [Fri, 26 Sep 2003 01:01:00 +0000 (18:01 -0700)]
[PATCH] s390: arch fixes.

 - Fix cflags for z990 compiles.
 - Rename resume to __switch_to to avoid name clash.
 - Fix show_trace and show_stack.
 - Add alignments to linker script.
 - Add atomic64_t and related funtions.
 - Add include/asm-s390/local.h
 - Fix 31 bit get_user for 8 byte values.
 - Fix show_regs oops.
 - Add a couple of might_sleep() calls.
 - Fix loading of modules with a BIG symbol table.
 - Fix inline asm constraint in __get_user_asm_1
 - Fix nested irq_enter bug on shutdown.
 - Add sched_clock function.

22 years ago[PATCH] fix AS crappy performance
Nick Piggin [Fri, 26 Sep 2003 00:55:16 +0000 (17:55 -0700)]
[PATCH] fix AS crappy performance

A while ago some guy complained that his server connections / sec was
way down due to AS. The server would fork a process which would then read
some tiny bit of random data and hang around for a while doing nothing.
This defeated all AS's heuristics, and due to a new process being a good
anticipation candidate by default, his performance went bad.

The fix was just to treat new processes as being unsuitable for
anticipation. Fixed. Unfortunately ls, cat, grep, small programs
starting up, find | xargs, etc only submit a couple of requests, and they
really want anticipation to be on.

This patch keeps a simple goodness value per queue based on how successful
anticipation has been for new processes.

22 years ago[PATCH] fix AS hangs
Nick Piggin [Fri, 26 Sep 2003 00:55:08 +0000 (17:55 -0700)]
[PATCH] fix AS hangs

Clears the next_arq cache if a request is repositioned and added as an alias
behind another.  Should fix the BUGs which Mary Edie Meredith has been
reporting.

Also, fixes a few random (but harmless) cases where last_merge was being set
and causing warnings.

When 2 requests are merged, and one is thrown away, move its alias list onto
the merged request! This is likely to be the cause of Mary's hangs.

22 years agoMerge bk://kernel.bkbits.net/vojtech/input
Linus Torvalds [Fri, 26 Sep 2003 00:41:03 +0000 (17:41 -0700)]
Merge bk://kernel.bkbits.net/vojtech/input
into home.osdl.org:/home/torvalds/v2.5/linux