Corey Minyard [Mon, 7 Nov 2005 09:00:05 +0000 (01:00 -0800)]
[PATCH] ipmi: use rcu lock for using command receivers
Use rcu_read_lock for the cmd_rcvrs list, since that was what what
intended, anyway. This means that all the users of the cmd_rcvrs_lock are
tasks, so the irq disables are no longer required for that lock and it can
become a semaphore.
Signed-off-by: Corey Minyard <minyard@acm.org> Acked-by: "Paul E. McKenney" <paulmck@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 09:00:03 +0000 (01:00 -0800)]
[PATCH] ipmi: add timer thread
We must poll for responses to commands when interrupts aren't in use. The
default poll interval is based on using a kernel timer, which varies with HZ.
For character-based interfaces like KCS and SMIC though, that can be way too
slow (>15 minutes to flash a new firmware with KCS, >20 seconds to retrieve
the sensor list).
This creates a low-priority kernel thread to poll more often. If the state
machine is idle, so is the kernel thread. But if there's an active command,
it polls quite rapidly. This decrease a firmware flash time from 15 minutes
to 1.5 minutes, and the sensor list time to 4.5 seconds, on a Dell PowerEdge
x8x system.
The timer-based polling remains, to ensure some amount of responsiveness even
under high user process CPU load.
Checking for a stopped timer at rmmod now uses atomics and del_timer_sync() to
ensure safe stoppage.
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 09:00:02 +0000 (01:00 -0800)]
[PATCH] ipmi: kcs error0 delay
BMCs can get into ERROR0 state while flashing new firmware, particularly while
the BMC is erasing the next flash block, which may take a just under 2 seconds
on a Dell PowerEdge 2800 (1.75 seconds typical), during which time the
single-threaded firmware may not be able to process new commands. In
particular, clearing OBF may not take effect immediately.
We want it to delay in ERROR0 after clearing OBF a bit waiting for OBF to
actually be clear before proceeding.
This introduces a new return value from the LLDD's event loop,
SI_SM_CALL_WITH_TICK_DELAY. This means the calling thread/timer should
schedule_timeout() at least 1 tick, rather than busy-wait. This is a longer
delay than SI_SM_CALL_WITH_DELAY, which is typically a 250us busy-wait.
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 09:00:01 +0000 (01:00 -0800)]
[PATCH] ipmi: bt restart reset fixes
The current BT retry/reset mechanism fails to succeed on a PowerEdge 1650,
when the controller is wedged with B2H_ATN asserted at XACTION_START. If this
occurs, no further commands will ever succeed unless the state of the
controller is first cleared out.
Furthermore, the soft reset would only occur if the first command after insmod
was the one that timed out, not if a later command timed out.
This patch changes the retry/reset mechanism to be as follows:
Before retrying a command, clear the state of the BT controller such that the
flags represent ready for a new transaction. This increases the chance of
success of the restarted transaction.
After 2 retries, issue a soft reset and retry one more time before giving up
and reporting back a failure.
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Acked-by: Rocky Craig <rocky.craig@hp.com> Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 08:59:59 +0000 (00:59 -0800)]
[PATCH] ipmi: si start transaction hook
Some commands, on some system BMCs, don't respond at at all. This is seen on
Dell PowerEdge x6xx and x7xx systems with IPMI 1.0 BT controllers when a "Get
SDR" command is issued, with a length field of 0x3A, which happens to be the
length of about SDR entries. If another length is passed, this command
succeeds.
This patch adds general infrastructure for receiving commands before they're
passed down to the low-level drivers, such that they can be completed
immediately, or modified, prior to being sent to ->start_transaction().
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 08:59:57 +0000 (00:59 -0800)]
[PATCH] ipmi: watchdog parms in sysfs
Modify the IPMI watchdog parameters (the ones that make sense) to be exported
from sysfs. This is somewhat complicated because these parameters have
side-effects that must be handled.
Signed-off-by: Corey Minyard <minyard@acm.org> Cc: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 08:59:56 +0000 (00:59 -0800)]
[PATCH] ipmi: various si cleanup
A number of small changes for the various system interface drivers,
consolidated from a number of patches from Matt Domsch.
Clear B2H_ATN and drain the BMC message buffer on command timeout. This
prevents further commands from failing after a timeout.
Add bt_debug and smic_debug module parameters, expose them in sysfs. This
lets you enable and disable debugging messages at runtime.
Unsigned jiffies math in ipmi_si_intf.c causes a too-large value to be passed
to ->event() after jiffies wrap-around. The BT driver had caught this, but
didn't know how to fix it. Now all calls to ->event() use a sane value for
time.
Increase timeout for commands handed to the BT driver from 2 seconds to 5
seconds. This is necessary particularly when the previous command was a
"Clear SEL", as that command completes, yet the BMC isn't really ready to
handle another command yet.
Silence BT debugging messages which were being printed on the console.
Increase SMIC timeout form 1/10s to 2s. This is needed on Dell PowerEdge 2650
and PowerEdge 750 with ERA/O cards to allow commands to complete without
timing out.
Adds kcs_debug module param, to match behavior of BT and SMIC. This also
prevents messages from being sent to the console unless explicitly requested.
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Corey Minyard <minyard@acm.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Corey Minyard [Mon, 7 Nov 2005 08:59:54 +0000 (00:59 -0800)]
[PATCH] ipmi: use refcount in message handler
This patch is rather large, but it really can't be done in smaller chunks
easily and I believe it is an important change. This has been out and tested
for a while in the latest IPMI driver release. There are no functional
changes, just changes as necessary to convert the locking over (and a few
minor style updates).
The IPMI driver uses read/write locks to ensure that things exist while they
are in use. This is bad from a number of points of view. This patch removes
the rwlocks and uses refcounts and RCU lists to manage what the locks did.
Signed-off-by: Corey Minyard <minyard@acm.org> Cc: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Nick Piggin [Mon, 7 Nov 2005 08:59:53 +0000 (00:59 -0800)]
[PATCH] as: cooperating processes
Introduce the notion of cooperating processes (those that submit requests
close to one another), and use these statistics to make better choices about
whether or not to do anticipatory waiting.
Help and analysis from Seetharami Seelam <seelam@cs.utep.edu>
Performance testing from Seelam:
I set up my system and executed a couple of tests that I used for OLS. I
tested with AS, cooperative process patch merged in -mm tree (which I called
Nick, below) and the cooperative patch with modifications to as_update_iohist
(which I called Seelam).
I used a dual-processor (2.28GHz Pentium 4 Xeon) system, with 1 GB main memory
and 1 MB L2 cache, running Linux 2.6.9. Only a single processor is used for
the experiments. I used 7.2K RPM Maxtor 10GB drive configured with ext2 file
system.
Experiment 1 (ex1) consists of reading one Linux source trees using
find . -type f -exec cat '{}' ';' > /dev/null.
Experiment 2 (ex2) consists of reading two disjoint Linux source trees
using
find . -type f -exec cat '{}' ';' > /dev/null.
Experiment 3 (ex3) consists of streaming read of a 2GB file in the background
and 1 instance of the chunk reads in Experiment 1.
Timings for reading the Linux source are shown below:
AS Nick Seelam
ex1: 0m25.813s 0m27.859s 0m27.640s
ex2: 1m11.468s 1m13.918s 1m5.869s
ex3: 81m44.352s 10m38.572s 6m47.994s
The difference between the numbers in Experiment 3 must be due to the code in
as_update_iohist. (akpm: that's not part of this patch. So this patch is
"Nick").
Signed-off-by: Nick Piggin <npiggin@suse.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Miklos Szeredi [Mon, 7 Nov 2005 08:59:52 +0000 (00:59 -0800)]
[PATCH] FUSE: pass file handle in setattr
This patch passes the file handle supplied in iattr to userspace, in case the
->setattr() was invoked from sys_ftruncate(). This solves the permission
checking (or lack thereof) in ftruncate() for the class of filesystems served
by an unprivileged userspace process.
Miklos Szeredi [Mon, 7 Nov 2005 08:59:51 +0000 (00:59 -0800)]
[PATCH] FUSE: atomic create+open
This patch adds an atomic create+open operation. This does not yet work if
the file type changes between lookup and create+open, but solves the
permission checking problems for the separte create and open methods.
Miklos Szeredi [Mon, 7 Nov 2005 08:59:50 +0000 (00:59 -0800)]
[PATCH] FUSE: add access call
Add a new access call, which will only be called if ->permission is invoked
from sys_access(). In all other cases permission checking is delayed until
the actual filesystem operation.
Miklos Szeredi [Mon, 7 Nov 2005 08:59:49 +0000 (00:59 -0800)]
[PATCH] FUSE: bump interface minor version
Though the following changes are all backward compatible (from the kernel's as
well as the library's POV) change the minor version, so interested
applications can detect new features.
The sys_ptrace boilerplate code (everything outside the big switch
statement for the arch-specific requests) is shared by most architectures.
This patch moves it to kernel/ptrace.c and leaves the arch-specific code as
arch_ptrace.
Some architectures have a too different ptrace so we have to exclude them.
They continue to keep their implementations. For sh64 I had to add a
sh64_ptrace wrapper because it does some initialization on the first call.
For um I removed an ifdefed SUBARCH_PTRACE_SPECIAL block, but
SUBARCH_PTRACE_SPECIAL isn't defined anywhere in the tree.
Signed-off-by: Christoph Hellwig <hch@lst.de> Acked-by: Paul Mackerras <paulus@samba.org> Acked-by: Ralf Baechle <ralf@linux-mips.org> Acked-By: David Howells <dhowells@redhat.com> Acked-by: Russell King <rmk+kernel@arm.linux.org.uk> Acked-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Tim Schmielau [Mon, 7 Nov 2005 08:59:43 +0000 (00:59 -0800)]
[PATCH] fix remaining missing includes
Fix more include file problems that surfaced since I submitted the previous
fix-missing-includes.patch. This should now allow not to include sched.h
from module.h, which is done by a followup patch.
Signed-off-by: Tim Schmielau <tim@physik3.uni-rostock.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Peter Staubach [Mon, 7 Nov 2005 08:59:42 +0000 (00:59 -0800)]
[PATCH] memory leak in dentry_open()
There is a memory leak possible in dentry_open(). If get_empty_filp()
fails, then the references to dentry and mnt need to be released. The
attached patch adds the calls to dput() and mntput() to release these two
references.
Signed-off-by: Peter Staubach <staubach@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Adrian Bunk [Mon, 7 Nov 2005 08:59:35 +0000 (00:59 -0800)]
[PATCH] quota: small cleanups
- "extern inline" -> "static inline"
- every file should #include the headers containing the prototypes for
it's global functions
Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Marcelo Tosatti [Mon, 7 Nov 2005 08:59:34 +0000 (00:59 -0800)]
[PATCH] Only disallow _setting_ of function key string
Mikael Pettersson <mikpe@csd.uu.se> noted that the current 2.6-git (and 2.4)
patch to disallow KDSKBSENT for unpriviledged users should be less restrictive
allowing reading of current function key string entry, but not writing.
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David Gibson [Mon, 7 Nov 2005 08:59:33 +0000 (00:59 -0800)]
[PATCH] FUTEX_WAKE_OP: enhanced error handling
The code for FUTEX_WAKE_OP calls an arch callback,
futex_atomic_op_inuser(). That callback can return an error code, but
currently the caller assumes any error is EFAULT, and will try various
things to resolve the fault before eventually giving up with EFAULT
(regardless of the original error code). This is not a theoretical case -
arch callbacks currently return -ENOSYS if the opcode they are given is
bogus.
This patch alters the code to detect non-EFAULT errors and return them
directly to the user.
Of course, whether -ENOSYS is the correct return value for the bogus opcode
case, or whether EINVAL would be more appropriate is another question.
Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Ingo Molnar <mingo@elte.hu> Cc: Jamie Lokier <jamie@shareable.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Zach Brown [Mon, 7 Nov 2005 08:59:31 +0000 (00:59 -0800)]
[PATCH] aio: remove aio_max_nr accounting race
AIO was adding a new context's max requests to the global total before
testing if that resulting total was over the global limit. This let
innocent tasks get their new limit tested along with a racing guilty task
that was crossing the limit. This serializes the _nr accounting with a
spinlock It also switches to using unsigned long for the global totals.
Individual contexts are still limited to an unsigned int's worth of
requests by the syscall interface.
The problem and fix were verified with a simple program that spun creating
and destroying a context while holding on to another long lived context.
Before the patch a task creating a tiny context could get a spurious EAGAIN
if it raced with a task creating a very large context that overran the
limit.
Signed-off-by: Zach Brown <zach.brown@oracle.com> Cc: Benjamin LaHaise <bcrl@kvack.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David Howells [Mon, 7 Nov 2005 08:59:30 +0000 (00:59 -0800)]
[PATCH] Keys: Remove incorrect and obsolete '!' operators
The attached patch removes a couple of incorrect and obsolete '!' operators
left over from the conversion of the key permission functions from
true/false returns to zero/error returns.
Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Hans Reiser [Mon, 7 Nov 2005 08:59:29 +0000 (00:59 -0800)]
[PATCH] reiser4: add radix_tree_lookup_slot()
Reiser4 uses radix trees to solve a trouble reiser4_readdir has serving nfs
requests.
Unfortunately, radix tree api lacks an operation suitable for modifying
existing entry. This patch adds radix_tree_lookup_slot which returns pointer
to found item within the tree. That location can be then updated.
Both Nick and Christoph Lameter have patches which need this as well.
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Abhay Salunke [Mon, 7 Nov 2005 08:59:26 +0000 (00:59 -0800)]
[PATCH] dell_rbu: Adding BIOS memory floor support
This patch has the changes to support the memory floor fix done in Dell
BIOS. The BIOS incase of packet update mechanism would not accept packet
placed in memory below a cretain address. This address is by default 128K
but can change. The driver now can accept the memory floor if the user
chooses to make it will try to allocate contiguous physical memory above
the memory floor by allocating a set of packets till a valid memory
allocation is made. All the allocates then are freed. This repeats for
everty packet.
This patch was created by Michael E Brown and has been tested on 2.6.14-rc5
Signed-of-by: Michael E Brown <Michael_E_Brown@Dell.com> Signed-off-by: Abhay Salunke <abhay_salunke@dell.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Kylene Jo Hall [Mon, 7 Nov 2005 08:59:25 +0000 (00:59 -0800)]
[PATCH] tpm: Fix lack of driver_unregister in init failcases
driver_unregister is not being properly called when the init function
returns an error case. Restructured the return logic such that this and
the other cleanups all happen in one place. Preformed many of the cleanups
that Andrew Morton's patch on Thursday made in tpm_atmel.c. Fixed
Matthieu's concern about writing before discovery.
(akpm: rmk said:
This driver is buggy. You must not provide your own release function - it
doesn't solve the problem which the warning (which you get when you don't
provide one) is telling you about.
You should convert your device driver over to the replacement dynamic platform
support, once it is merged. IOW, something like:
pdev = platform_device_alloc("mydev", id);
if (pdev) {
err = platform_device_add_resources(pdev, &resources,
ARRAY_SIZE(resources));
if (err == 0)
err = platform_device_add_data(pdev, &platform_data,
sizeof(platform_data));
if (err == 0)
err = platform_device_add(pdev);
} else {
err = -ENOMEM;
}
if (err)
platform_device_put(pdev);
)
Signed-off-by: Kylene Jo Hall <kjhall@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Kirill Smelkov [Mon, 7 Nov 2005 08:59:23 +0000 (00:59 -0800)]
[PATCH] serial moxa: fix wrong BUG
There is a wrong BUG in mxser_close.
The BUG is triggered when tty->driver_data == NULL, But in fact this is not
a bug, because tty->driver->close is called even when tty->driver->open
fails.
LDD3 tells us to do nothing in such cases.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Bob Picco [Mon, 7 Nov 2005 08:59:19 +0000 (00:59 -0800)]
[PATCH] HPET, Maintainers
This patch identifies the HPET Maintainers. Clemens in taking over as
primary maintainer for the HPET driver. Clemens has i386 hardware with
HPET and is a better choice than me because of this. I've shared this
patch with all cc: recipients and there is agreement on ownership.
Hopefully this eliminates future confusion in terms of where HPET
maintenance is owned.
Signed-off-by: Bob Picco <bob.picco@hp.com> Cc: <clemens@ladisch.de> Cc: <venkatesh.pallipadi@intel.com> Cc: <bob.picco@hp.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Reported by Eddy Petrisor <eddy.petrisor@gmail.com>
fs/built-in.o(.text+0x35fdc): In function `hfs_mdb_put':
: undefined reference to `unload_nls'
fs/built-in.o(.text+0x35ff1): In function `hfs_mdb_put':
: undefined reference to `unload_nls'
fs/built-in.o(.text+0x367a5): In function `parse_options':
super.c: undefined reference to `load_nls'
fs/built-in.o(.text+0x367db):super.c: undefined reference to `load_nls'
fs/built-in.o(.text+0x36938):super.c: undefined reference to `load_nls_default'
Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org> Acked-by: Roman Zippel <zippel@linux-m68k.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] Remove hlist_for_each_rcu() API, convert existing use to hlist_for_each_entry_rcu
Remove the hlist_for_each_rcu() API, which is used only in one place, and
is trivially converted to hlist_for_each_entry_rcu(), making the code
shorter and more readable. Any out-of-tree uses may be similarly
converted.
Signed-off-by: "Paul E. McKenney" <paulmck@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Matt Helsley [Mon, 7 Nov 2005 08:59:16 +0000 (00:59 -0800)]
[PATCH] Process Events Connector
This patch adds a connector that reports fork, exec, id change, and exit
events for all processes to userspace. It replaces the fork_advisor patch
that ELSA is currently using. Applications that may find these events
useful include accounting/auditing (e.g. ELSA), system activity monitoring
(e.g. top), security, and resource management (e.g. CKRM).
Signed-off-by: Matt Helsley <matthltc@us.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Andrew Morton [Mon, 7 Nov 2005 08:59:15 +0000 (00:59 -0800)]
[PATCH] write_inode_now(): write inode if not BDI_CAP_NO_WRITEBACK
If the backing_dev_info doesn't have BDI_CAP_NO_WRITEBACK we're not supposed
to write back an inode's pages. But in this situation write_inode_now()
refuses to write the inode itself as well. Fix.
Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] Move Kprobes and Oprofile to "Instrumentation Support" menu
Andrew Morton suggested to move kprobes from kernel hacking menu, since
kernel hacking menu is in-appropriate for the Kprobes. This patch moves
Kprobes and Oprofile under instrumentation menu.
(akpm: it's not a natural fit, but things like djprobes and the s390 guys'
statistics library need a home)
Signed-of-by: Prasanna S Panchamukhi <prasanna@in.ibm.com> Cc: Philippe Elie <phil.el@wanadoo.fr> Cc: John Levon <levon@movementarian.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Heiko Carstens [Mon, 7 Nov 2005 08:59:12 +0000 (00:59 -0800)]
[PATCH] s390: merge common parts of head.S and head64.S
Merge common parts of head.S and head64.S into head.S and move architecture
specific parts to head31.S and head64.S respectively. Saves us ~500 lines
of duplicated assembly code.
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Access to FBA disks via DIAG fails for block sizes > 512 byte. The device
analysis code of the DIAG discipline does not properly initialize the DIAG250
device environment after completion of the analysis. This results in VM only
serving 512 bytes per block I/O request whereas Linux expects larger block
sizes. Add proper device environment setup to end of analysis code.
Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Future versions of gcc may remove initialization code for control blocks used
by the diag250 inline assembly due to incompletely specified constraints.
This may lead to erratic behavior. Fix the diag250 inline assembly
constraints.
Signed-off-by: Peter Oberparleiter <peter.oberparleiter@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The test_bit function returns a non-boolean value, it returns 0,1,2,4,...
instead of only 0 or 1. This causes wrongs results in the mincore system
call. Check against 0 to get a proper boolean value.
Signed-off-by: Christian Borntraeger <cborntra@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove pagex pseudo page fault code. It does not work together with the
system call speedup that makes the complete system call path enabled for
interrupts. To make pagex and the syscall speedup code work together we would
have to add code to the program check handler to do a critical section cleanup
like the asynchronous interrupt code. This would make program checks slower.
Not what we want.
Newer versions of z/VM have the improved pfault pseudo page fault interface.
This replaces the old pagex interface and does not have the problem. So its
better to just rip out the pagex code.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Cornelia Huck [Mon, 7 Nov 2005 08:59:05 +0000 (00:59 -0800)]
[PATCH] s390: ccwgroup online attribute
Make the interface for setting ccw group devices on-/offline consistent with
that for ccw devices: Check if the device driver provided a set_{on,off}line
function and just set the device on-/offline if not.
Signed-off-by: Cornelia Huck <cohuck@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Heiko Carstens [Mon, 7 Nov 2005 08:59:04 +0000 (00:59 -0800)]
[PATCH] s390: memory query wait psw
Don't switch back to 24 bit addressing mode when waiting for an external
interrupt and set the correct bit in wait PSW (external mask instead of I/O
mask).
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The calculation of the value return by next_timer_interrupt from jiffies to
jiffies_64 is racy against xtime updates. We need to protect the calculation
with read_seqbegin/read_seqretry.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: build host-binaries with the native host arch again
This patch reverts back the changes to HOSTCFLAGS and HOSTLDFLAGS
When we were building complete binaries to get constants (such as ptrace
register layout on stack) from host userspace headers, we needed to make the
arch for building HOST binaries match our one: i.e. on a 64bit system
compiling 32bit binaries, we compile 32-bit hostprogs and need, say, 32-bit
ncurses. Now we can revert that - that avoids problem with, say, menuconfig
and ncurses, on a system which can't compile well 32-bit programs.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: fix hardcoded ZONE_* constants in zone setup
Remove usage of hardcoded constants in paging_init().
By chance I spotted a bug in zones_setup involving a change to ZONE_*
constants, due to the ZONE_DMA32 patch from Andi Kleen (which is in -mm).
So, possibly, instead of zones_size[2] you will find zones_size[3] in the
code, but that change is wrong and this patch is still correct.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 7 Nov 2005 08:58:57 +0000 (00:58 -0800)]
[PATCH] uml: make tt mode-dependent options depend on MODE_TT
This makes some of the tt-specific options actually depend on CONFIG_MODE_TT.
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 7 Nov 2005 08:58:57 +0000 (00:58 -0800)]
[PATCH] uml: big memory fixes
A number of fixes to improve behavior when large physical memory sizes
are specified:
- libc files need -D_FILE_OFFSET_BITS=64 because there are unavoidable uses
of non-64 interfaces in libc
- some %d need to be %u
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Bodo Stroesser [Mon, 7 Nov 2005 08:58:55 +0000 (00:58 -0800)]
[PATCH] uml: maintain own LDT entries
Patch imlements full LDT handling in SKAS:
* UML holds it's own LDT table, used to deliver data on
modify_ldt(READ)
* UML disables the default_ldt, inherited from the host (SKAS3)
or resets LDT entries, set by host's clib and inherited in
SKAS0
* A new global variable skas_needs_stub is inserted, that
can be used to decide, whether stub-pages must be supported
or not.
* Uses the syscall-stub to replace missing PTRACE_LDT (therefore,
write_ldt_entry needs to be modified)
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ben Lahaise [Mon, 7 Nov 2005 08:58:52 +0000 (00:58 -0800)]
[PATCH] uml: switch_mm fix
Not quite, something along the lines of the patch below works correctly (and
makes aio performance not suffer from multiple second delays), as skas0 mode
correctly switches mm contexts, unlike TT (which should probably get nuked
from the kernel now that skas0 seems to be working).
Signed-off-by: Benjamin LaHaise <bcrl@linux.intel.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 7 Nov 2005 08:58:51 +0000 (00:58 -0800)]
[PATCH] uml: separate libc-dependent helper code
The serial UML OS-abstraction layer patch (um/kernel dir).
This moves all systemcalls from helper.c file under os-Linux dir
Signed-off-by: Gennady Sharapov <Gennady.V.Sharapov@intel.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 7 Nov 2005 08:58:50 +0000 (00:58 -0800)]
[PATCH] uml: separate libc-dependent early initialization
The serial UML OS-abstraction layer patch (um/kernel dir).
This moves all systemcalls from main.c file under os-Linux dir and joins mem.c
and um_arch.c files.
Signed-off-by: Gennady Sharapov <Gennady.V.Sharapov@intel.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The serial UML OS-abstraction layer patch (um/kernel dir).
This moves all systemcalls from uaccess_user.c file under os-Linux dir
Signed-off-by: Gennady Sharapov <Gennady.V.Sharapov@intel.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
ifa->ifa_address and ifa->ifa_mask are defined as __u32, but used as if they
were char[4].
Network code uses htons() to convert it. So UML's method to access these
fields is wrong for bigendians (e.g. s390)
I replaced bytewise copying by memcpy(), maybe even that might be removed, if
ifa->ifa_address/mask may be used immediately.
Signed-off-by: Bodo Stroesser <bstroesser@fujitsu-siemens.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Cc: <viro@parcelfarce.linux.theplanet.co.uk> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike noted that the assembly code for syscall stubs is misassembled with
GCC 3.2.3: the values copied in registers weren't preserved between one asm()
and the following one.
So I fixed the thing by rewriting the __asm__ constraints more like unistd.h
ones.
Note: in syscall6 case I had to add one more instruction (i.e. moving arg6 in
eax and shuffling things around) - it's needed for the function to be valid in
general (we can't load the value from the stack, relative to ebp, because we
change it), but could be avoided since we actually use a constant as param 6.
The only fix would be to turn stub_syscall6 to a macro and use a "i"
constraint for arg6 (i.e., specify it's a constant value).
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 7 Nov 2005 08:58:45 +0000 (00:58 -0800)]
[PATCH] uml: improve stub debugging
Add some more debugging information when a stub does something unexpected,
usually segfaulting. Now, it dumps out the stub's registers as well as the
signal.
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Heiko Carstens [Mon, 7 Nov 2005 08:58:38 +0000 (00:58 -0800)]
[PATCH] cpu hoptlug: avoid usage of smp_processor_id() in preemptible code
Replace smp_processor_id() with any_online_cpu(cpu_online_map) in order to
avoid lots of "BUG: using smp_processor_id() in preemptible [00000001]
code:..." messages in case taking a cpu online fails.
All the traces start at the last notifier_call_chain(...) in kernel/cpu.c.
Since we hold the cpu_control semaphore it shouldn't be any problem to access
cpu_online_map.
The reason why cpu_up failed is simply that the cpu that was supposed to be
taken online wasn't even there. That is because on s390 we never know when a
new cpu comes and therefore cpu_possible_map consists of only ones and doesn't
reflect reality.
Zwane Mwaikambo [Mon, 7 Nov 2005 08:58:33 +0000 (00:58 -0800)]
[PATCH] i386: LVT entries remaining unmasked on reboot
Excerpt from bugzilla entry
http://bugzilla.kernel.org/show_bug.cgi?id=5518
"i386 version of Reboot-through-BIOS is unsafe: it forgets to mask APIC LVT
interrupts before jumping to a BIOS entry point. As a result, BIOS ends up
bombarded with interrupts early on boot. The BIOS does not expect it since
following a "normal" hardware cpu reset, all APIC LVT registers have the
Mask bit (16) set and can't generate interrupts.
For example, the version of Phoenix BIOS used by VMware enables interrupts
for the first time before masking/clearing APIC LVT. The APIC Timer LVT
register is still set up for a timer interrupt delivery with a high vector
from the previous Linux incarnation (0xef in our case). The BIOS has not
fully initialized its IDT at this point and the real mode gate for 0xef
remains all zeros. Vector 0xef dispatches BIOS to address 0:0, BIOS takes
a #GP and eventually hangs.
machine_shutdown() does attempt to shut down APIC before jumping to BIOS,
but it is ineffective"
Signed-off-by: Zwane Mwaikambo <zwane@arm.linux.org.uk> Cc: "Seth, Rohit" <rohit.seth@intel.com> Cc: Zachary Amsden <zach@vmware.com> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:28 +0000 (00:58 -0800)]
[PATCH] sh: Use pfn_valid() for lazy dcache write-back on SH7705
SH7705 in extended cache mode has some left-over VALID_PAGE() cruft that it
checks when doing lazy dcache write-back. This has been gone for some time
(the last bits were in the discontig code, which should now also be gone --
this also fixes up a build error in the non-discontig case).
pfn_valid() gives the desired behaviour, so we switch to that.
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:24 +0000 (00:58 -0800)]
[PATCH] sh: Drop hp690 discontig support
There was only one board using this (hp690 specifically), and it just so
happens that it's only physically discontiguous at the "normal" P1 offset. If
we bump up the P1 offset, it's possible to hit a shadowed region of memory
where we suddenly become magically contiguous.
As people have been using this shadowed region workaround for quite some time
(and without any adverse effects), it's time to drop the left over discontig
bits that no longer have any practical use (it was always very much
hp690-centric to begin with).
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:23 +0000 (00:58 -0800)]
[PATCH] sh: pte_mkhuge() compile fix for !CONFIG_HUGETLB_PAGE
Presently it is bogus to call pte_mkhuge() outside of the CONFIG_HUGETLB_PAGE
context, as the only processors that support _PAGE_SZHUGE do so in the
hugetlbpage context only (and this is the only time that _PAGE_SZHUGE is even
defined). SH-2 and SH-3 do not support huge pages at all, and so it is not
possible to enable this.
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:22 +0000 (00:58 -0800)]
[PATCH] sh: SuperHyway support for SH4-202
This adds support for the relatively quirky (ie, not in line with any known
documentation, and amazed it works at all) SuperHyway implementation on
SH4-202. This depends on the earlier SuperHyway patch for multiple block
support and VCR refactoring.
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:21 +0000 (00:58 -0800)]
[PATCH] superhyway: multiple block support and VCR rework
This extends the API somewhat to allow for platform-specific VCR reading and
writing. Some platforms (like SH4-202) implement the VCR in a split VCRL and
VCRH, but end up being in reverse order or have other quirks that need to be
dealt with, so we add a set of superhyway_ops per-bus to accomodate this.
We also have to extend the per-device resources somewhat, as some devices now
conveniently split control and data blocks. So we allow a platform to
register its set of SuperHyway devices via superhyway_add_devices() with the
control block always ordered as the first resource (as this is the one that
userspace cares about).
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Paul Mundt [Mon, 7 Nov 2005 08:58:20 +0000 (00:58 -0800)]
[PATCH] sh: Drop deprecated support for custom ramdisk embedding
sh had its own support for embedding ramdisk images in to the kernel binary,
but people are using initramfs for this now, so we drop the ramdisk embedding.
Signed-off-by: Paul Mundt <lethal@linux-sh.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
projects / linux-2.6 / log