Linus Torvalds [Wed, 1 Aug 2007 03:40:50 +0000 (20:40 -0700)]
Merge branch 'release' of git://lm-sensors.org/kernel/mhoffman/hwmon-2.6
* 'release' of git://lm-sensors.org/kernel/mhoffman/hwmon-2.6:
hwmon: fscher read control bugfix
hwmon: (adm1031) Fix broken links in documentation
hwmon: make abituguru3_read_increment_offset() static
hwmon: Fix regression caused by typo in lm90.c
hwmon: (applesmc) add temperature sensors set for Macbook
hwmon: fscher control update bugfix
hwmon: fix dme1737 temp fault attribute
hwmon: Add missing __devexit tags in various drivers
hwmon: clean up duplicate includes
hwmon: fix lm78 detection regression
hwmon: fix array overruns in lm93.c
hwmon: add support for THMC50 and ADM1022
Len Brown [Wed, 1 Aug 2007 03:27:10 +0000 (23:27 -0400)]
ACPI: delete CONFIG_ACPI_PROCFS_SLEEP (again)
CONFIG_ACPI_PROCFS_SLEEP is a NO-OP -- delete it (again).
Apparently 296699de6bdc717189a331ab6bbe90e05c94db06 creating CONFIG_SUSPEND
and CONFIG_PM_SLEEP was based on an out-dated version of drivers/acpi/Kconfig,
as it erroneously restored this recently deleted config option.
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hpa/linux-2.6-x86setup:
[x86 setup] EDD: Fix the computation of the MBR sector buffer
[x86 setup] Newline after setup signature failure message
x86 boot code comments typos
Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (41 commits)
[RTNETLINK]: Fix warning for !CONFIG_KMOD
[IPV4] ip_options.c: kmalloc + memset conversion to kzalloc
[DECNET]: kmalloc + memset conversion to kzalloc
[NET]: ethtool_perm_addr only has one implementation
[NET]: ethtool ops are the only way
[PPPOE]: Improve hashing function in hash_item().
[XFRM]: State selection update to use inner addresses.
[IPSEC]: Ensure that state inner family is set
[TCP]: Bidir flow must not disregard SACK blocks for lost marking
[TCP]: Fix ratehalving with bidirectional flows
[PPPOL2TP]: Add CONFIG_INET Kconfig dependency.
[NET]: Page offsets and lengths need to be __u32.
[AF_UNIX]: Make code static.
[NETFILTER]: Make nf_ct_ipv6_skip_exthdr() static.
[PKTGEN]: make get_ipsec_sa() static and non-inline
[PPPoE]: move lock_sock() in pppoe_sendmsg() to the right location
[PPPoX/E]: return ENOTTY on unknown ioctl requests
[IPV6]: ipv6_addr_type() doesn't know about RFC4193 addresses.
[NET]: Fix prio_tune() handling of root qdisc.
[NET]: Fix sch_api to properly set sch->parent on the root.
...
David Brownell [Tue, 31 Jul 2007 07:39:45 +0000 (00:39 -0700)]
spi device setup gets better error checking
This updates some error reporting paths in SPI device setup:
- Move validation logic for SPI chipselects to spi_new_device(),
which is where it should always have been.
- In spi_new_device(), emit error messages if the device can't
be created. This is LOTS better than a silent failure; though
eventually, the calling convention should probably change to
use the <linux/err.h> conventions.
- Includes one previously-missing check: SPI masters must always
have at least one chipselect, even for dedicated busses which
always keep it selected!
It also adds a FIXME (IDR for dynamic ID allocation) so the issue doesn't live
purely in my mailbox.
Meelis Roos [Tue, 31 Jul 2007 07:39:41 +0000 (00:39 -0700)]
fix integer overflow warning in i2o_block
drivers/message/i2o/i2o_block.c: In function 'i2o_block_transfer':
drivers/message/i2o/i2o_block.c:837: warning: integer overflow in expression
msg->u.head[1] = cpu_to_le32(I2O_CMD_PRIVATE << 24 | HOST_TID << 12 | tid);
and I2O_CMD_PRIVATE is defined as 0xFF. This gets "0xFF0100 | tid" and fits
into 32-bit unsigned but not into 32-bit signed integer properly. Target
value is defined as u32 so the claculation does not fit during computation.
Change local variable tid to u32 so the whole expression is of u32 type and
fits well into u32 result.
This patch fix weird behaviour of UDF mounting procedure. To get UID
changed (for now) we have to type
mount -t udf -o uid=some_user,uid=ignore /dev/device /mnt/moun_point
and specifying two uid at once is strange a bit. So with the patch we are
able to mount without additional 'uid=ignore' option. The same for GID
option is done.
This patch will not break current mount scheme (with two option).
Btw this does fix (I hope) the following
[BUG 6124] mount of UDF fs ignores UID and GID options
http://bugzilla.kernel.org/show_bug.cgi?id=6124
This is a set of small fixes addressing points raised with the original
driver submission. In particular, __maybe_unused is used rather than a
local hack and sbd_ops is made const. Additionally I have made two local
string variables automatic as rodata space was wasted for pointers
unnecessarily.
Fix a use after free bug in kernel->userspace relay file support
Coverity spotted what looks like a real possible case of using a variable
after it has been freed. The problem is in
kernel/relay.c::relay_open_buf()
If the code hits "goto free_buf;" it ends up in this code :
free_buf:
relay_destroy_buf(buf); <--- calls kfree() on 'buf'.
free_name:
kfree(tmpname);
end:
return buf; <-- use after free of 'buf'.
I read through the callers and they all handle a NULL return from this
function as an error (and hitting the 'free_buf' label only happens on
failure to chan->cb->create_buf_file(), so that looks like a clear error to
me).
The patch simply sets 'buf' to NULL after the call to
relay_destroy_buf(buf); - as far as I can see that should take care of the
problem.
The patch also corrects a reference to a documentation file while
I was at it.
Note from Mathieu: the documentation reference change should have been
done in a separate patch, but I guess no one will really care.
WARNING: kernel/built-in.o(.text+0x16910): Section mismatch:
reference to .init.text: (between 'kthreadd' and 'init_waitqueue_head')
comes because kernel/kthread.c:kthreadd() is not __init but calls
kthreadd_setup() which is __init. But this is ok, because kthreadd_setup()
is only ever called at init time, and then kthreadd() proceeds into its
"for (;;)" loop. We could mark kthreadd __init_refok, but kthreadd_setup()
with just one callsite and 4 lines in it (it's been that small since 10ab825bdef8df51) doesn't need to be a separate function at all -- so let's
just move those four lines at beginning of kthreadd() itself.
david m. richter [Tue, 31 Jul 2007 07:39:12 +0000 (00:39 -0700)]
VFS: fix a race in lease-breaking during truncate
It is possible that another process could acquire a new file lease right
after break_lease() is called during a truncate, but before lease-granting
is disabled by the subsequent get_write_access(). Merely switching the
order of the break_lease() and get_write_access() calls prevents this race.
Mike Frysinger [Tue, 31 Jul 2007 07:39:11 +0000 (00:39 -0700)]
use __val in __get_unaligned
Use "__val" rather than "val" in the __get_unaligned macro in
asm-generic/unaligned.h. This way gcc wont warn if you happen to also name
something in the same scope "val".
Gabriel C [Tue, 31 Jul 2007 07:39:06 +0000 (00:39 -0700)]
broken lilo check on make install
On make install I get the this error:
...
sh /work/crazy/linux-git/linux-2.6/arch/i386/boot/install.sh 2.6.22-g4eb6bf6b arch/i386/boot/bzImage System.map "/boot"
/work/crazy/linux-git/linux-2.6/arch/i386/boot/install.sh: line 54:
/etc/lilo/install: No such file or directory
make[1]: *** [install] Error 127
...
I don't use and don't have lilo installed on this system. The attached
patch fixes the problem for me.
This is only called at init time and only happens if the BIOS screws
something up, so the leak is slight and it is probably not worth sending to
2.6.22.x. The driver would not initialize the interface in the case, and I
have no reports of this happening. I have booted and run tests on a system
with this patch. Note that the original patch was munged by the mailer,
here's a new one.
If we ever hit the "default:" case in the switch in try_init_dmi(),
then we'll leak the storage allocated with kzalloc() and assigned
to 'info'.
Randy Dunlap [Tue, 31 Jul 2007 07:39:04 +0000 (00:39 -0700)]
docbook bad file references
Fix docbook warnings:
Warning(linux-2.6.22-git12//drivers/base/power/main.c): no structured comments found
Warning(linux-2.6.22-git12//include/linux/splice.h): no structured comments found
Randy Dunlap [Tue, 31 Jul 2007 07:39:02 +0000 (00:39 -0700)]
i2c.h kernel-doc additions
Add kernel-doc notation in <linux/i2c.h> for:
Warning(linux-2.6.22-git12//include/linux/i2c.h:183): No description found for parameter 'driver'
Warning(linux-2.6.22-git12//include/linux/i2c.h:183): No description found for parameter 'usage_count'
Warning(linux-2.6.22-git12//include/linux/i2c.h:183): No description found for parameter 'list'
Warning(linux-2.6.22-git12//include/linux/i2c.h:183): No description found for parameter 'released'
Currently, the MPSC driver doesn't stop recieving characters when the CREAD
flag in termios->c_cflag is cleared. It should. Also, only start receiving
if its not already started.
Carlos Sanchez [Tue, 31 Jul 2007 07:38:59 +0000 (00:38 -0700)]
serial: MPSC: Remove race between Rx stop & restart
The patch in commit ID f7232056bff5fe2d3bfeab35252a66ebaeb5bbde stops (aborts)
the MPSC's receive engine just before restarting it. Unfortunately, it
doesn't wait for the abort to complete before restarting it which creates a
race between the abort and the restart. If the restart occurs first, the
in-progress abort stops it again and the rx engine remains stopped.
Instead, do the abort when the SDMA engine is being stopped. Make sure to
wait for the abort to complete before continuing.
It turned out that mounting a corrupted ISO image to a regular file may
succeed, e.g. if an image was prepared as follows:
$ dd if=correct.iso of=bad.iso bs=4k count=8
We then can mount it to a regular file:
# mount -o loop -t iso9660 bad.iso /tmp/file
But mounting it to a directory fails with -ENOTDIR, simply because
the root directory inode doesn't have S_IFDIR set and the condition
in graft_tree() is met:
if (S_ISDIR(nd->dentry->d_inode->i_mode) !=
S_ISDIR(mnt->mnt_root->d_inode->i_mode))
return -ENOTDIR
This is because the root directory inode was read from an incorrect
block. It's supposed to be read from sbi->s_firstdatazone, which is
an absolute value and gets messed up in the case of an incorrect image.
In order to somehow circumvent this we have to check that the root
directory inode is actually a directory after all.
As reported by Gustavo de Nardin <[email protected]>, while trying to
compile xosview (http://xosview.sourceforge.net/) with upstream kernel
headers being used you get the following errors:
serialmeter.cc:48:30: error: linux/serial_reg.h: No such file or directory
serialmeter.cc: In member function 'virtual void
SerialMeter::checkResources()':
serialmeter.cc:71: error: 'UART_LSR' was not declared in this scope
serialmeter.cc:71: error: 'UART_MSR' was not declared in this scope
...
David Howells [Tue, 31 Jul 2007 07:38:49 +0000 (00:38 -0700)]
AFS: fix file locking
Fix file locking for AFS:
(*) Start the lock manager thread under a mutex to avoid a race.
(*) Made the locking non-fair: New readlocks will jump pending writelocks if
there's a readlock currently granted on a file. This makes the behaviour
similar to Linux's VFS locking.
Paul E. McKenney [Tue, 31 Jul 2007 07:38:48 +0000 (00:38 -0700)]
bpqether: fix rcu usage
The rcu_dereference() primitive needs to be applied to an l-value in order to
ensure that compiler writers don't get an opportunity to apply reordering
optimizations that could result in multiple fetches or in other misbehavior.
This patch pulls the rcu_dereference() calls in bpq_seq_next() up to the point
at which the fetched pointers are still l-values, rather than after
list_entry() has transformed them into r-values.
Pavel Emelianov [Tue, 31 Jul 2007 07:38:48 +0000 (00:38 -0700)]
Fix user struct leakage with locked IPC shem segment
When user locks an ipc shmem segmant with SHM_LOCK ctl and the segment is
already locked the shmem_lock() function returns 0. After this the
subsequent code leaks the existing user struct:
Other results of this are:
1. the new shp->mlock_user is not get-ed and will point to freed
memory when the task dies.
2. the RLIMIT_MEMLOCK is screwed on both user structs.
kbuild: whitelist references from __dbe_table to .init
This is needed on MIPS where the same mechanism as get_user() is used to
intercept bus error exceptions for some hardware probes. Without this
patch modpost will throw spurious warnings:
In file included from kernel/notifier.c:1:
include/linux/kdebug.h:14: warning: 'struct notifier_block' declared inside parameter list
include/linux/kdebug.h:14: warning: its scope is only this definition or declaration, which is probably not what you want
include/linux/kdebug.h:15: warning: 'struct notifier_block' declared inside parameter list
kernel/notifier.c:529: error: conflicting types for 'register_die_notifier'
include/linux/kdebug.h:14: error: previous declaration of 'register_die_notifier' was here
kernel/notifier.c:533: error: conflicting types for 'register_die_notifier'
include/linux/kdebug.h:14: error: previous declaration of 'register_die_notifier' was here
kernel/notifier.c:536: error: conflicting types for 'unregister_die_notifier'
include/linux/kdebug.h:15: error: previous declaration of 'unregister_die_notifier' was here
kernel/notifier.c:539: error: conflicting types for 'unregister_die_notifier'
include/linux/kdebug.h:15: error: previous declaration of 'unregister_die_notifier' was here
Anton Vorontsov [Tue, 31 Jul 2007 07:38:44 +0000 (00:38 -0700)]
spidev_test utility
This is a simple utility used to test SPI functionality. It could stand
growing options to support using other test data patterns; this initial
version only issues full duplex transfers, which rules out 3WIRE or
Microwire links.
Anton Vorontsov [Tue, 31 Jul 2007 07:38:43 +0000 (00:38 -0700)]
spidev supports more communications modes
The spidev driver doesn't currently expose all SPI communications modes to
userspace. This passes them all through to the driver.
Two of them are potentially troublesome, in the sense that they could cause
hardware conflicts on shared busses. It might be appropriate to add some
privilege checks for for those modes.
Is there a reason why the "online" file in the subdirectories for the CPUs
in /sys/devices/system isn't world-readable? I cannot imagine it to be
security relevant especially now that a getcpu() syscall can be used to
determine what CPUa thread runs on.
The file is useful to correctly implement the sysconf() function to return
the number of online CPUs. In the presence of hotplug we currently cannot
provide this information. The patch below should to it.
arch/i386/kernel/apm.c: In function 'apm_init':
arch/i386/kernel/apm.c:2240: warning: format '%lx' expects type 'long
unsigned int', but argument 3 has type 'u32'
Sam Ravnborg [Tue, 31 Jul 2007 07:38:13 +0000 (00:38 -0700)]
lib: move kasprintf to a separate file
kasprintf pulls in kmalloc which proved to be fatal for at least
bootimage target on alpha.
Move it to a separate file so only users of kasprintf are exposed
to the dependency on kmalloc.
Meelis Roos [Tue, 31 Jul 2007 07:38:13 +0000 (00:38 -0700)]
alpha: fix boot/main.c warning
In current 2.6.23-rc1+git, make bootimage gives the following warning while
compiling arch/alpha/boot/main.c. The patch below fixes the warning by
casting callback argument explicitly to void*. The original value comes from
START_ADDR macro and is clearly numeric so only cast it for the callback.
CC arch/alpha/boot/main.o
arch/alpha/boot/main.c: In function 'load':
arch/alpha/boot/main.c:135: warning: passing argument 3 of 'callback_read' makes pointer from integer without a cast
Meelis Roos [Tue, 31 Jul 2007 07:38:12 +0000 (00:38 -0700)]
alpha: fix objstrip.c compilation warnings
In current 2.6.23-rc1+git, make bootimage gives the following warnings while
compiling objstrip.c. The patch below fixes these warnings by casting strncmp
argument to char * - it does not seem feasible to change its type in struct
elfhdr.
HOSTCC arch/alpha/boot/tools/objstrip
arch/alpha/boot/tools/objstrip.c: In function 'main':
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of 'strlen' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of 'strlen' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of '__builtin_strcmp' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of 'strlen' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of '__builtin_strcmp' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of '__builtin_strcmp' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of '__builtin_strcmp' differ in signedness
arch/alpha/boot/tools/objstrip.c:147: warning: pointer targets in passing argument 1 of 'strncmp' differ in signedness
Meelis Roos [Tue, 31 Jul 2007 07:38:11 +0000 (00:38 -0700)]
alpha: fix mkbb compilation warnings
In current 2.6.23-rc1+git, make bootimage gives the following warnings while
compiling mkbb.c. The patch below fixes these warnings by using the proper
include for exit() and using appropriate printf format.
HOSTCC arch/alpha/boot/tools/mkbb
arch/alpha/boot/tools/mkbb.c: In function 'main':
arch/alpha/boot/tools/mkbb.c:95: warning: implicit declaration of function 'exit'
arch/alpha/boot/tools/mkbb.c:95: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:102: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:110: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:117: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int'
arch/alpha/boot/tools/mkbb.c:118: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:125: warning: format '%d' expects type 'int', but argument 3 has type 'long unsigned int'
arch/alpha/boot/tools/mkbb.c:126: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:143: warning: incompatible implicit declaration of built-in function 'exit'
arch/alpha/boot/tools/mkbb.c:148: warning: incompatible implicit declaration of built-in function 'exit'
This seems like it will break decades-long-working stuff, in favor of
breaking new ground in our favorite area, "trusting the BIOS."
It's just not worth it for serial ports, IMO. Serial ports are something
that just shouldn't break at this late stage in the game. My new Intel
platform boxes don't even have serial ports, so I question the value of
messing with serial port probing even more... because... just wait a year,
and your box won't have a serial port either! :)
I certainly don't object to the use of platform devices (or isa_driver),
but the probe change seems questionable. That's sorta analagous to
rewriting the floppy driver probe routine. Sure you could do it... but why
risk all that damage and go through debugging all over again?
It seems clear from this report that we cannot, should not, trust BIOS for
something (a) so simple and (b) that has been working for over a decade.
Much discussion ensued and we've decided to have another go at all of this.
Remove unused TIF_NOTIFY_RESUME flag for all processor architectures. The
flag was not used excecpt on IA-64 where the patch replaces it with
TIF_PERFMON_WORK.
Arne Redlich [Tue, 31 Jul 2007 07:37:57 +0000 (00:37 -0700)]
md: handle writes to broken raid10 arrays gracefully
When writing to a broken array, raid10 currently happily emits empty bio
lists. IOW, the master bio will never be completed, sending writers to
UNINTERRUPTIBLE_SLEEP forever.
J. Bruce Fields [Tue, 31 Jul 2007 07:37:53 +0000 (00:37 -0700)]
knfsd: eliminate unnecessary -ENOENT returns on export downcalls
A succesful downcall with a negative result (which indicates that the given
filesystem is not exported to the given user) should not return an error.
Currently mountd is depending on stdio to write these downcalls. With some
versions of libc this appears to cause subsequent writes to attempt to write
all accumulated data (for which writes previously failed) along with any new
data. This can prevent the kernel from seeing responses to later downcalls.
Symptoms will be that nfsd fails to respond to certain requests.
Jeff Layton [Tue, 31 Jul 2007 07:37:51 +0000 (00:37 -0700)]
knfsd: set the response bitmask for NFS4_CREATE_EXCLUSIVE
RFC 3530 says:
If the server uses an attribute to store the exclusive create verifier, it
will signify which attribute by setting the appropriate bit in the attribute
mask that is returned in the results.
Linux uses the atime and mtime to store the verifier, but sends a zeroed out
bitmask back to the client. This patch makes sure that we set the correct
bits in the bitmask in this situation.
Randy Dunlap [Tue, 31 Jul 2007 07:37:50 +0000 (00:37 -0700)]
sched: fix kernel-doc warnings
Fix kernel-doc warnings in sched.c:
Warning(linux-2623-rc1g4//kernel/sched.c:1685): No description found for parameter 'notifier'
Warning(linux-2623-rc1g4//kernel/sched.c:1696): No description found for parameter 'notifier'
Warning(linux-2623-rc1g4//kernel/sched.c:1750): No description found for parameter 'prev'
Randy Dunlap [Tue, 31 Jul 2007 07:37:50 +0000 (00:37 -0700)]
pnp: fix kernel-doc warnings
Fix PNP docbook warnings:
Warning(linux-2623-rc1g4//drivers/pnp/core.c): no structured comments found
Warning(linux-2623-rc1g4//drivers/pnp/driver.c): no structured comments found
Mingming Cao [Tue, 31 Jul 2007 07:37:46 +0000 (00:37 -0700)]
"ext4_ext_put_in_cache" uses __u32 to receive physical block number
Yan Zheng wrote:
> I think I found a bug in ext4/extents.c, "ext4_ext_put_in_cache" uses
> "__u32" to receive physical block number. "ext4_ext_put_in_cache" is
> used in "ext4_ext_get_blocks", it sets ext4 inode's extent cache
> according most recently tree lookup (higher 16 bits of saved physical
> block number are always zero). when serving a mapping request,
> "ext4_ext_get_blocks" first check whether the logical block is in
> inode's extent cache. if the logical block is in the cache and the
> cached region isn't a gap, "ext4_ext_get_blocks" gets physical block
> number by using cached region's physical block number and offset in
> the cached region. as described above, "ext4_ext_get_blocks" may
> return wrong result when there are physical block numbers bigger than
> 0xffffffff.
>
Andrew Morton [Tue, 31 Jul 2007 07:37:43 +0000 (00:37 -0700)]
parport_pc locking fix
http://bugzilla.kernel.org/show_bug.cgi?id=8821 reports a might_sleep()
warning due to parport_pc_exit() running platform_device_unregister() while
holding ports_lock.
Just remove the locking: nobody else can access ports_list during module_exit.
Randy Dunlap [Tue, 31 Jul 2007 07:37:40 +0000 (00:37 -0700)]
various doc/kernel-parameters fixes
- tell what APIC (by request), MTD, & PARIDE mean
- correct some source file names
- remove IA64 "llsc*=" (seems to have been removed from source tree)
- removel SCSI "53c7xx=" (driver already removed)
projects / linux.git / log