Kiyoshi Ueda [Fri, 17 Jun 2005 14:15:10 +0000 (16:15 +0200)]
When cfq I/O scheduler is selected, get_request() in __make_request() calls
__cfq_get_queue(). __cfq_get_queue() finds an existing queue (struct
cfq_queue) of the current process for the device and returns it. If it's not
found, __cfq_get_queue() creates and returns a new one if __cfq_get_queue() is
called with __GFP_WAIT flag, or __cfq_get_queue() returns NULL (this means that
get_request() fails) if no __GFP_WAIT flag.
On the other hand, in __make_request(), get_request() is called without
__GFP_WAIT flag at the first time. Thus, the get_request() fails when there is
no existing queue, typically when it's called for the first I/O request of the
process to the device.
Though it will be followed by get_request_wait() for general case,
__make_request() will just end the I/O with an error (EWOULDBLOCK) when the
request was for read-ahead.
Catalin Marinas [Thu, 16 Jun 2005 17:01:11 +0000 (18:01 +0100)]
[PATCH] ARM: 2712/1: Fix the RGB order for the Versatile CLCD
Patch from Catalin Marinas
The current red and blue colours on the Versatile CLCD are
reversed when the 5:6:5 mode is used. The patch sets the proper
bit in the SYS_CLCD register value.
Signed-off-by: Catalin Marinas Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The ELF core dump code has one use of off_t when writing out segments.
Some of the segments may be passed the 2GB limit of an off_t, even on a
32-bit system, so it's important to use loff_t instead. This fixes a
corrupted core dump in the bigcore test in GDB's testsuite.
Signed-off-by: Daniel Jacobowitz <dan@codesourcery.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Alexandre Oliva [Thu, 16 Jun 2005 05:26:31 +0000 (22:26 -0700)]
[PATCH] sbp2 slab corruption fix
This fixed a problem that showed up in the Fedora development tree a few
weeks before the Fedora Core 4 release, initially as slab corruption, later
as hard crashes on boot up, when slab debugging was disabled for the
release. More details on the history at
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=158424
The problem is caused by sbp2's use of scsi_host->hostdata[0] to hold a
scsi_id, without explicitly requesting space for it. Since hostdata is
declared as a zero-sized array, we don't get any such space by default, so
it must be explicitly requested. The patch below implements just that.
Tejun Heo [Thu, 16 Jun 2005 10:57:31 +0000 (12:57 +0200)]
This patch fixes q->unplug_thresh condition check in
__elv_add_request(). rq.count[READ] + rq.count[WRITE] can increase
more than one if another thread has allocated a request after the
current request is allocated or in_flight could have changed resulting
in larger-than-one change of nrq, thus breaking the threshold
mechanism.
Olaf Hering [Tue, 14 Jun 2005 20:52:19 +0000 (13:52 -0700)]
[PATCH] update ppc64 defconfig
enable cpusets
enable new lpfc and jsm drivers
enable new dm-multipath
leave new agp disabled
disable rivafb, it does not handle the cards in G5 models (FX5200 as example)
the new nvidiafb doesnt work on bigendian, yet
Signed-off-by: Olaf Hering <olh@suse.de> Acked-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Karsten Wiese [Tue, 14 Jun 2005 16:56:20 +0000 (09:56 -0700)]
[PATCH] usbusx2y: prevent oops & dead keyboard on usb unplugging while the device is being used
Without this patch, some usb kobjects, which are parents to the usx2y's
kobjects can be freed before the usx2y's. This led to an oops in
get_kobj_path_length() and a dead keyboard, when the usx2y's kobjects
were freed. The patch ensures the correct sequence. Tested ok on
kernel 2.6.12-rc2.
Karsten Wiese [Tue, 14 Jun 2005 16:54:55 +0000 (09:54 -0700)]
[PATCH] usbaudio: prevent oops & dead keyboard on usb unplugging while the device is being used
Without this patch, some usb kobjects, which are parents to the usx2y's
kobjects can be freed before the usx2y's. This led to an oops in
get_kobj_path_length() and a dead keyboard, when the usx2y's kobjects
were freed. The patch ensures the correct sequence. Tested ok on
kernel 2.6.12-rc2.
Thomas Hood [Tue, 14 Jun 2005 05:58:04 +0000 (22:58 -0700)]
[PATCH] apm.c: ignore_normal_resume is set a bit too late
This patch causes the ignore_normal_resume flag to be set slightly earlier,
before there is a chance that the apm driver will receive the normal resume
event from the BIOS. (Addresses Debian bug #310865)
Signed-off-by: Thomas Hood <jdthood@yahoo.co.uk> Acked-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Olof Johansson [Mon, 13 Jun 2005 22:52:27 +0000 (15:52 -0700)]
[PATCH] Fix PCI BAR size interpretation on 64-bit arches
On 64-bit machines, PCI_BASE_ADDRESS_MEM_MASK and other mask constants
passed to pci_size() are 64-bit (for example ~0x0fUL). However, pci_size
does comparisons between the u32 arguments and the mask, which will fail
even though any result from pci_size is still just 32-bit.
Changing the mask argument to u32 seems the obvious thing to do, since all
arithmetic in the function is 32-bit and having a larger mask makes no
sense.
This triggered on a PPC64 system here where an adapter (VGA, as it
happened) had a memory region base of 0xfe000000 and a sz of the same,
matching the if (max == maxbase ...) test at the bottom of pci_size but
failing the mask comparison. Quite a corner case which I guess explains
why we haven't seen it until now.
Signed-off-by: Olof Johansson <olof@lixom.net> Acked-by: Greg KH <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Mon, 13 Jun 2005 22:52:14 +0000 (15:52 -0700)]
[PATCH] uml: use fork instead of clone
Convert the boot-time host ptrace testing from clone to fork. They were
essentially doing fork anyway. This cleans up the code a bit, and makes
valgrind a bit happier about grinding 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>
Linus Torvalds [Tue, 14 Jun 2005 00:51:55 +0000 (17:51 -0700)]
Update DCO ("signoff") rules to 1.1
This adds a clause that notes explicitly that the person doing the
sign-off knows that the project (and his sign-off) is public and will
possibly get archived and re-distributed.
This patch alows you to change the source address of icmp error
messages. It applies cleanly to 2.6.11.11 and retains the default
behaviour.
In the old (default) behaviour icmp error messages are sent with the ip
of the exiting interface.
The new behaviour (when the sysctl variable is toggled on), it will send
the message with the ip of the interface that received the packet that
caused the icmp error. This is the behaviour network administrators will
expect from a router. It makes debugging complicated network layouts
much easier. Also, all 'vendor routers' I know of have the later
behaviour.
Signed-off-by: David S. Miller <davem@davemloft.net>
[SCTP] Extend the info exported via /proc/net/sctp to support netstat for SCTP.
Signed-off-by: Vladislav Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Sridhar Samudrala <sri@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[SCTP]: Fix bug in restart of peeled-off associations.
Signed-off-by: Vladislav Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Sridhar Samudrala <sri@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Userland layer-2 tunneling devices allocated through the TUNTAP driver
(drivers/net/tun.c) have a type of ARPHRD_NONE, and have no link-layer
address. The kernel complains at regular interval when IPv6 Privacy
extension are enabled because it can't find an hardware address :
Dec 29 11:02:04 auguste kernel: __ipv6_regen_rndid(idev=cb3e0c00):
cannot get EUI64 identifier; use random bytes.
IPv6 Privacy extensions should probably be disabled on that sort of
device. They won't work anyway. If userland wants a more usual
Ethernet-ish interface with usual IPv6 autoconfiguration, it will use a
TAP device with an emulated link-layer and a random hardware address
rather than a TUN device.
As far as I could fine, TUN virtual device from TUNTAP is the very only
sort of device using ARPHRD_NONE as kernel device type.
David Brownell [Mon, 13 Jun 2005 14:15:28 +0000 (07:15 -0700)]
[PATCH] spin longer for ehci port reset completion
This makes the EHCI driver spin a bit longer before concluding that the
port reset failed. "Obviously safe."
It allows some devices to enumerate that previously didn't. We've seen
a bunch of these problem reports recently, this will make some go away.
As reported by Michael Zapf <Michael.Zapf@uni-kassel.de>, some EHCI
controllers seem to take forever to finish port resets and produce
"port N reset error -110" type errors. Spinning a bit longer helps.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Alan Cox [Sat, 11 Jun 2005 17:00:52 +0000 (18:00 +0100)]
[PATCH] pwc bug fix
The pwc chainsaw session left some setups not working. There is a
sanity check on compression buffers that simply isn't right any more as
we never allocate one.
This doesn't address the email and other changes. I'll do those
tomorrow if I get time, but it is the minimal fix for the code and basic
feature set.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] radeonfb: don't blow up VGA console on load
The current radeonfb memset's the framebuffer to 0 when loaded. This
removes occasional artifacts but has the nasty side effect that if you
load radeonfb without framebuffer console, you destroy the VGA text
buffer, font, etc... radeon must not touch the framebuffer content when
it doesn't "own" it.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
M68k: Mark Sun-3 NCR5380 SCSI broken until NCR5380_abort() and
NCR5380_bus_reset() are replaced with real new-style EH routines (the old EH
SCSI constants were removed in 2.6.12-rc3).
Now m68k no longer sets HAVE_ARCH_GET_SIGNAL_TO_DELIVER, can it be removed
completely? Or may ARM26 still need it? Note that its usage was removed from
kernel/signal.c about 2 months ago.
David Brownell [Sun, 12 Jun 2005 22:26:05 +0000 (23:26 +0100)]
[PATCH] ARM: 2709/1: Systems with PCMCIA should also see IDE options (for CompactFlash memories)
Patch from David Brownell
The ARM generic Kconfig filters out IDE options ... except for
an error prone ARMload of special cases.
This adds one general case to the systems that will offer IDE options:
kernels with PCMCIA support, which probably want to use IDE to access
CompactFlash cards. This might allow many (most?) of the other cases
to disappear, for systems that only see IDE hardware through CF cards.
Right now this one patch is used to gate access to CF cards, including
MicroDrives, for both omap_cf and at91_cf drivers.
Signed-off-by: David Brownell Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Dave Airlie [Fri, 10 Jun 2005 09:27:51 +0000 (19:27 +1000)]
[PATCH] remove bogus hack from radeon IRQ handler
This removes a bogus hack from the radeon IRQ handler.
There is a better fix from myself and benh in DRM CVS but I'll wait
until 2.6.13-rc so it gets more testing.
Despite all the care lately in making the powermac sleep/wakeup as
robust as possible, there is still a nasty related to the use of cpufreq
on PMU based machines. Unfortunately, it affects paulus old powerbook
so I have to fix it :)
We didn't manage to understand what is precisely going on, it leads to
memory corruption and might have to do with RAM not beeing properly
refreshed when a cpufreq transition is done right before the sleep.
The best workaround (and less intrusive at this point) we could come up
with is included in this patch. We basically do _not_ force a switch to
high speed on suspend anymore (that is what is causing the problem) on
those machines. We still force a speed switch on wakeup (since we don't
know what speed we are coming back from sleep at, and that seems to work
fine).
Since, during this short interval, the actual CPU speed might be
incorrect, we also hack around by multiplying loops_per_jiffy by 2 (max
speed factor on those machines) during early wakeup stage to make sure
udelay's during that time aren't too short.
For after 2.6.12, we'll change udelay implementation to use the CPU
timebase (which is always constant) instead like we do on ppc64 and thus
get rid of all those problems.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] iseries_veth: Supress spurious WARN_ON() at module unload
My patch from a few weeks back (now in mainline), called "Cleanup skbs to
prevent unregister_netdevice() hanging", can cause our TX timeout code to
fire on machines with lots of VLANs (because it takes > 2 seconds between
when we stop the queues and when we're finished stopping the connections).
When that happens the TX timeout code freaks out and does a WARN_ON()
because as far as it's concerned there shouldn't be a TX timeout happening,
which is fair enough.
I have a "proper" fix for this, which is to a) do refcounting on
connections and b) implement a proper ack timer so we don't keep unacked
skbs lying around for ever. But for 2.6.12 I propose just supressing the
WARN_ON(). Users will still see the "NETDEV WATCHDOG" warning, but that's
not nearly as bad as a WARN_ON() which users interpret as an Oops.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Narendra Sankar [Fri, 6 May 2005 19:00:05 +0000 (12:00 -0700)]
[PATCH] PCI: MSI functionality broken on Serverworks GC chipset
MSI functionality is broken on the GC_LE x86 chipset that Serverworks
developed and that is being used in various platforms today. Broadcom is
going to push out to the kernel MSI enabled Gigabit drivers (in the very
near future), and we would like to make sure that MSI does not get
enabled on any platforms using the GC_LE chipset (device id 0x17).
Following the AMD 8131 example, I am including a patch to disable MSI
functionality when a GCNB_LE is detected. Please let me know if there
are any issues with this. This is a permanent fix for this chipset, as
the hardware will not be updated.
[IA64] Fix race condition in the rt_sigprocmask fastcall
current->blocked will be set to the value of current->thread_info->flags if the
cmpxchg to update thread_info->flags fails. For performance reasons the store into
current->blocked was placed in the cmpxchg loop. However, the cmpxchg overwrites the
register holding the value to be stored. In the rare case of a retry the value of
thread_info->flags will be written into current->blocked.
The fix is to use another register so that the register containing the current->blocked
value is not overwritten.
Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
Ian Abbott [Thu, 2 Jun 2005 09:34:11 +0000 (10:34 +0100)]
[PATCH] USB: ftdi_sio: avoid losing received data in tty-ldisc
ftdi_sio: Avoid losing bytes at tty-ldisc.
This patch was originally developed by Daniel Smertnig. I
(Ian Abbott) made a few changes. It has been tested by both
Daniel and I, at least for raw, non-canonical receive data
processing.
Here is Daniel's original description of the patch:
===
During a project in which I was using a FTDI 232BM to
transmit data at relative high speeds (625kBit/s), I
noticed a problem where data was lost even if flow
control was enabled: The FTDI-Driver receives 512 Bytes
of data over USB at a time, which consists of 8 64-Byte
packets. Subtracting the 2 bytes of status information
included in each packet this gives 496 "real" data
bytes per read.
This data is passed (indirectly, via the flip buffers)
to the tty line discipline which takes care of
throttling when there the free buffer space reaches
TTY_THRESHOLD_THROTTLE (128). Because the FTDI driver
processes up to 496 bytes at a time, throttling won't
happen in time and the line discipline will discard the
remaining bytes.
To avoid this the patch passes data in 62-byte blocks
to the tty layer and checks the available space in the
ldisc-buffers. If there isn't enough free space,
processing the rest of the data is delayed using a
workqueue.
Note: The original problem should be easily
reproducible with a userspace program which does slow &
small reads.
===
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Daniel Smertnig <daniel.smertnig@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Pete Zaitcev [Mon, 6 Jun 2005 20:54:59 +0000 (13:54 -0700)]
[PATCH] USB: fix ub issues
This smoothes two imperfections:
- Increase number of LUNs per device from 4 to 9. The best solution
would be to remove this limit altogether, but that has to wait until
the time when more than 26 hosts are allowed.
- Replace mdelay with msleep in a probing routine.
Signed-off-by: Pete Zaitcev <zaitcev@yahoo.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Albert Lee [Mon, 6 Jun 2005 07:56:03 +0000 (15:56 +0800)]
[PATCH] sg traverse fix for __atapi_pio_bytes()
Problem:
Incorrect md5sum when using ATAPI PIO mode to verify a distro CD.
Root cause: sg traverse problem.
In __atapi_pio_bytes(), if qc->cursg++ is increased and "goto
next_page" is executed, then sg is not updated to the new qc->cursg
and the old sg is overwritten with the new data.
Changes:
- Replace "goto next_page" with "goto next_sg" to make sg updated.
Paul Mackerras [Wed, 8 Jun 2005 11:59:15 +0000 (21:59 +1000)]
[PATCH] ppc64: Fix PER_LINUX32 behaviour
This patch fixes some bugs in the ppc64 PER_LINUX32 implementation,
noted by Juergen Kreileder:
* uname(2) doesn't respect PER_LINUX32, it returns 'ppc64' instead of 'ppc'
* Child processes of a PER_LINUX32 process don't inherit PER_LINUX32
Along the way I took the opportunity to move things around so that
sys_ppc32.c only has 32-bit syscall emulation functions and to remove
the obsolete "fakeppc" command line option.
Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
READA errors failing with EWOULDBLOCK/EAGAIN do not constitute a valid
reason for failing the path; this lead to erratic errors on DM multipath
devices. This error can be safely propagated upwards without failing the
path.
Acked-by: Kevin Corry <kevcorry@us.ibm.com> Acked-by: Jens Axboe <axboe@suse.de> Signed-off-by: Lars Marowsky-Bree <lmb@suse.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Peter Chubb [Wed, 8 Jun 2005 22:50:20 +0000 (15:50 -0700)]
[PATCH] ia64: fix floating-point preemption problem
There've been reports of problems with CONFIG_PREEMPT=y and the high
floating point partition. This is caused by the possibility of preemption
and rescheduling on a different processor while saving or restioirng the
high partition.
The only places where the FPU state is touched are in ptrace, in
switch_to(), and where handling a floating-point exception. In switch_to()
preemption is off. So it's only in trap.c and ptrace.c that we need to
prevent preemption.
Here is a patch that adds commentary to make the conditions clear, and adds
appropriate preempt_{en,dis}able() calls to make it so. In trap.c I use
preempt_enable_no_resched(), as we're about to return to user space where
the preemption flag will be checked anyway.
Signed-off-by: Peter Chubb <peterc@gelato.unsw.edu.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove spurious MSR_SE reset during kprobe processing.
single_step_exception() already does it for us. Reset it to be safe when
executing the fault_handler.
Signed-off-by: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Keith Owens [Wed, 8 Jun 2005 22:49:07 +0000 (15:49 -0700)]
[PATCH] Stop arch/i386/kernel/vsyscall-note.o being rebuilt every time
arch/i386/kernel/vsyscall-note.o is not listed as a target so its .cmd file
is neither considered as a target nor is it read on the next build. This
causes vsyscall-note.o to be rebuilt every time that you run make, which
causes vmlinux to be rebuilt every time.
Signed-off-by: Keith Owens <kaos@ocs.com.au> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Wed, 8 Jun 2005 22:48:27 +0000 (15:48 -0700)]
[PATCH] uml: clean up error path
This cleans an error path which used to leak file descriptors by returning
without trying to tidy up.
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Wed, 8 Jun 2005 22:48:13 +0000 (15:48 -0700)]
[PATCH] uml: fix strace -f
It turns out that we need to check for pending signals when a newly forked
process is run for the first time. With strace -f, strace needs to know about
the forked process before it gets going. If it doesn't, then it ptraces some
bogus values into its registers, and the process segfaults. So, I added calls
to interrupt_end, which does that, plus checks for reschedules. There
shouldn't be any of those, but x86 does the same thing, so I'm copying that
behavior to be safe.
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Wed, 8 Jun 2005 22:48:01 +0000 (15:48 -0700)]
[PATCH] uml: compile fixes for gcc 4
This is a bunch of compile fixes provoked by building UML with gcc 4. There
are a bunch of signedness mismatches, a couple of uninitialized references,
and a botched C99 structure initialization which had somehow gone unnoticed.
Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
projects / linux-stable / log