USB: at91_udc: Cleanup variables after failure in usb_gadget_register_driver()
This patch zeroes some variables when usb_gadget_register_driver()
fails. gadgetfs does a dummy registration to get the name of the USB
driver and then waits for user-land driver. If someone plugs the cable
in the meantime, bad things happen, because at91_udc has been left in
inconsistent state.
Signed-off-by: Wojtek Kaniewski <wojtekka@toxygen.net> Acked-by: David Brownell <dbrownell@users.sourceforge.net> Cc: Andrew Victor <andrew@sanpeople.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Tony Olech [Wed, 6 Dec 2006 13:16:22 +0000 (13:16 +0000)]
USB: u132-hcd/ftdi-elan: add support for Option GT 3G Quad card
ELAN's U132 is a USB to CardBus OHCI controller adapter,
designed specifically for CardBus 3G data cards to
function in machines without a CardBus slot.
The "ftdi-elan" module is a USB client driver, that detects
a supported CardBus OHCI controller plugged into the
U132 adapter and thereafter provides the conduit for
for access by the "u132-hcd" module.
The "u132-hcd" module is a (cut-down OHCI) host controller
that supports a single OHCI function of the CardBus
card inserted into the U132 adapter.
The problem with the initial implementation is that when
the CardBus card inserted into the U132 adapter has multiple
functions (and a CardBus card can support up to 4 functions),
it was the first function that was arbitrarily choosen.
The first batch of 3G cards tested, like the Merlin Qualcomm
V620, have two functions each supporting a seperate USB OHCI
host controller, of which it was that first function that is
wired up to the 3G modem.
Then along comes the Vodafone Mobile Connect 3G/GPRS data card,
aka "Option GT 3G Quad" as printed on it's rear or "Option N.V.
GlobeTrotter Fusion Quad Lite" as read with "lspci -v". And it
has the meaningful functionality in the second CardBus function.
That presents a problem because it was the "ftdi-elan" module
alone that knows how to communicate to the embedded CardBus slot
and the "u132-hcd" module alone that knows how to access the
pcmcia configuration and CardBus accessible memory space. And
of course, the information about attached (internally hardwired)
devices is contained within USB configuration embedded somewhere
within the CardBus card.
If only the "u132-hcd" module probe() interface could return a
result code that propagated back to the instigating function
platform_device_register() then the "ftdi-elan" module could
try an alternative CardBus function. However in spite of
the recent changes to the drivers/base/ routines that moved
device_attach() from bus_add_device() to bus_attach_device()
both of those routines lose the "failed to attach" 0 result
code and thus the calling routine, namely device_add() is
incapable of propaging the "failed to attach" condition back
to platform_device_add() and consequently back to the caller
of platform_device_register()
Experiments show that patching bus_attach_device() to return
ENODEV fails with the kernel locking up very early during
boot. But, however, if the patch is restricted to calls from
platform_device_add() then it does seem to work.
Unfortunately, until the kernel's drivers/base is properly
modified to propagate -ENODEV back to the caller of
platform_device_register(), it is necessary to "fix" the
"ftdi-elan" module by importing knowledge from the
"u132-hcd" module. This is the reason for the duplicated
functionality introduced in this patch.
Signed-off-by: Tony Olech <tony.olech@elandigitalsystems.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Victor [Fri, 8 Dec 2006 06:44:38 +0000 (22:44 -0800)]
USB: at91_udc, misc fixes
This is an update to the AT91 USB Device (Gadget) driver.
Adds support for the Atmel AT91SAM9260 and AT91SAM9261 processors. The
only difference is how they handle the pullup pin.
[Patch from Patrice Vilchez]
Need to clear any pending USB Device interrupts before registering the
interrupt handler. The bootloader might have been using the USB Device
port. [Patch from Peer Georgi]
VBUS detection is handled by a GPIO interrupt which only triggers on a
change. Is is therefore necessary to read the current VBUS state
explicitly at startup. [Patch from Peer Georgi]
Signed-off-by: Andrew Victor <andrew@sanpeople.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Andrew Victor [Fri, 8 Dec 2006 06:44:33 +0000 (22:44 -0800)]
USB: at91 udc, support at91sam926x addresses
This is an update to the AT91 USB Device (Gadget) driver.
The base I/O address provided in the platform_device resources is now
ioremap()'ed instead of using a statically mapped memory area. This
helps portability to the newer AT91sam926x processors.
The major change is that we now have to pass a 'struct at91_udc'
parameter to at91_udp_read() and at91_udp_write().
Signed-off-by: Andrew Victor <andrew@sanpeople.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Takamasa Ohtake [Thu, 7 Dec 2006 01:04:15 +0000 (17:04 -0800)]
USB: ohci handles hardware faults during root port resets
I have found a problem where the root_port_reset() goes into an infinite
loop and stalls the kernel.
This happens when a hardware fault inside the machine occurs during a small
timing window. In case of USB device connection, if a USB device responds to
hcd_submit_urb(), and later the controller fails before root_port_reset(),
root_port_reset() will loop infinitely because ohci_readl() will always
return "-1". Such a failure can include ejecting a CardBus OHCI controller.
The probability of this problem is low, but it will increase if PnP type
usage is frequent. The attached patch can solve this problem and I believe
that it is better to fix this problem.
Andrew Victor [Tue, 5 Dec 2006 11:20:31 +0000 (03:20 -0800)]
USB: ohci at91 warning fix
Remove a warning about an unused variable in the OHCI bus glue for at91.
Signed-off-by: Andrew Victor <andrew@sanpeople.com> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Tue, 5 Dec 2006 11:18:31 +0000 (03:18 -0800)]
USB: ohci whitespace/comment fixups
This is an OHCI cleanup patch ... it removes a lot of erroneous whitespace
(space before tab, at end of line) as well as the obsolete inline changelog.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Tue, 5 Dec 2006 11:10:08 +0000 (03:10 -0800)]
USB: MAINTAINERS update, EHCI and OHCI
Update maintainer records for two USB host controller drivers. I'm the
main point of contact for both EHCI and OHCI, although I don't have much
time for them any more. Roman hasn't submitted OHCI patches for years.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Tue, 5 Dec 2006 11:15:33 +0000 (03:15 -0800)]
USB: gadget driver unbind() is optional; section fixes; misc
Allow gadget drivers to omit the unbind() method. When they're
statically linked, that's an appropriate memory saving tweak.
Similarly, provide consistent/simpler handling for a should-not-happen
error case: removing a peripheral controller driver when a gadget
driver is still loaded. Such code dates back to early versions of the
first implementation of the gadget API, and has never been triggered.
Includes relevant section annotation fixs for gmidi.c, file_storage.c,
and serial.c; we don't yet have an "init or exit" annotation. Also
some whitespace fixes in gmidi.c (space at EOL, before tabs, etc).
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Tue, 5 Dec 2006 21:29:55 +0000 (16:29 -0500)]
UHCI: module parameter to ignore overcurrent changes
Certain boards seem to like to issue false overcurrent notifications,
for example on ports that don't have anything connected to them. This
looks like a hardware error, at the level of noise to those ports'
overcurrent input signals (or non-debounced VBUS comparators). This
surfaces to users as truly massive amounts of syslog spam from khubd
(which is appropriate for real hardware problems, except for the
volume from multiple ports).
Using this new "ignore_oc" flag helps such systems work more sanely,
by preventing such indications from getting to khubd (and spamming
syslog). The downside is of course that true overcurrent errors will
be masked; they'll appear as spontaneous disconnects, without the
diagnostics that will let users troubleshoot issues like
short-circuited cables. In addition, controllers with no devices
attached will be forced to poll for new devices rather than relying on
interrupts, since each overcurrent event would generate a new
interrupt.
This patch (as826) is essentially a copy of David Brownell's ignore_oc
patch for ehci-hcd, ported to uhci-hcd.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Jeff Garzik [Mon, 4 Dec 2006 01:53:58 +0000 (20:53 -0500)]
USB: fix ohci.h over-use warnings
When u132-hcd is built, it includes local header ohci.h, which appears
to have been intended only for use by ohci-hcd.
This throws warnings about functions which are defined and not used.
The warnings thrown are because three small functions are implemented in
the header, but not declared 'inline', a rather strange affair.
Since these functions are small, let's go ahead and define them as
'inline', just like the inline functions surrounding them. This makes
things more consistent, and kills the warnings.
Signed-off-by: Jeff Garzik <jeff@garzik.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Clare [Sat, 2 Dec 2006 02:24:38 +0000 (18:24 -0800)]
USB: Prevent the funsoft serial device from entering raw mode
Added a device specific ioctl function to prevent the disabling of canonical
mode. EINVAL is returned for any TCSETSF ioctl that doesn't have ICANON set.
This patch is for 2.6.17 or later kernels.
When "hwinfo --modem" is executed it opens the funsoft USB serial device and
disables canonical mode. The device is kept this way until hwininfo has
finished probing any modems on a system. The funsoft device expects to be
running in canonical mode. Switching the device to raw mode can cause
incomplete data packets and device timeouts.
Signed-off-by: David Clare <david@funsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan [Mon, 4 Dec 2006 16:43:01 +0000 (16:43 +0000)]
usb serial: Eliminate bogus ioctl code
Several drivers have bogus ioctl code that tries unneccessarily to
override the standard processing. In the three cases here the actual code
is not only wrong but also not required as they implement the proper
set_termios method as well.
Remove the junk.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Oliver Neukum [Mon, 27 Nov 2006 17:41:30 +0000 (18:41 +0100)]
USB: removing ifdefed code from gl620a
as David has objected to the patch against the gl620a driver,
here's a patch implementing David' suggestion of removing the incomplete
ifdefed code from the gl620a driver.
Signed-off-by: Oliver Neukum <oliver@neukum.name> Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Linus Torvalds [Tue, 19 Dec 2006 18:32:40 +0000 (10:32 -0800)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
[PATCH] Generic HID layer - update MAINTAINERS
input/hid: Supporting more keys from the HUT Consumer Page
[PATCH] Generic HID layer - build: USB_HID should select HID
Leigh Brown [Mon, 18 Dec 2006 01:12:30 +0000 (17:12 -0800)]
[TCP]: Fix oops caused by tcp_v4_md5_do_del
md5sig_info.alloced4 must be set to zero when freeing keys4, otherwise
it will not be alloc'd again when another key is added to the same
socket by tcp_v4_md5_do_add.
Signed-off-by: Leigh Brown <leigh@solinno.co.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Mon, 18 Dec 2006 01:07:29 +0000 (17:07 -0800)]
[TG3]: Fix race condition when calling register_netdev().
Hot-plug scripts can call tg3_open() as soon as register_netdev() is
called in tg3_init_one(). We need to call pci_set_drvdata() before
register_netdev(), and netif_carrier_off() needs to be moved to
tg3_open() to avoid race conditions.
Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Mon, 18 Dec 2006 01:06:46 +0000 (17:06 -0800)]
[TG3]: Assign tp->link_config.orig_* values.
tp->link_config.orig_* values must be assigned during
tg3_set_settings() because these values will be used to setup the
link speed during tg3_open(). Without these assignments, the link
speed settings will be all messed by if tg3_set_settings() is called
when the device is down.
Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 17 Dec 2006 22:37:23 +0000 (14:37 -0800)]
[NETFILTER] IPV6: Fix dependencies.
Although the menu dependencies in net/ipv6/netfilter/Kconfig
guard the entries in that file from the Kconfig GUI, this does
not prevent them from being selected still via "make oldconfig"
when IPV6 etc. is disabled.
So add explicit dependencies.
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael Chan [Thu, 14 Dec 2006 23:56:32 +0000 (15:56 -0800)]
[BNX2]: Fix panic in bnx2_tx_int().
There was an off-by-one bug in bnx2_tx_avail(). If the tx ring is
completely full, the producer and consumer indices may be apart by
256 even though the ring size is only 255. One entry in the ring is
unused and must be properly accounted for when calculating the number
of available entries. The bug caused the tx ring entries to be
reused by mistake, overwriting active entries, and ultimately causing
it to crash.
This bug rarely occurs because the tx ring is rarely completely full.
We always stop when there is less than MAX_SKB_FRAGS entries available
in the ring.
Thanks to Corey Kovacs <cjk@techma.com> and Andy Gospodarek
<agospoda@redhat.com> for reporting the problem and helping to collect
debug information.
Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
rose_add_loopback_neigh uses kmalloc and the callers were ignoring the
error value. Rewrite to let the caller deal with the allocation. This
allows the use of static allocation of kmalloc use entirely.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: David S. Miller <davem@davemloft.net>
ax25_linkfail_register uses kmalloc and the callers were ignoring the
error value. Rewrite to let the caller deal with the allocation. This
allows the use of static allocation of kmalloc use entirely.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Replace ax25_protocol_register by ax25_register_pid which assumes the
caller has done the memory allocation. This allows replacing the
kmalloc allocations entirely by static allocations.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ralf Baechle [Thu, 14 Dec 2006 23:49:28 +0000 (15:49 -0800)]
[AX.25]: Mark all kmalloc users __must_check
The recent fix 0506d4068bad834aab1141b5dc5e748eb175c6b3 made obvious that
error values were not being propagated through the AX.25 stack. To help
with that this patch marks all kmalloc users in the AX.25, NETROM and
ROSE stacks as __must_check.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq:
[CPUFREQ] longhaul compile fix.
[CPUFREQ] Advise not to use longhaul on VIA C7.
[CPUFREQ] set policy->curfreq on initialization
[CPUFREQ] Trivial cleanup for acpi read/write port in acpi-cpufreq.c
[CPUFREQ] fixes typo in cpufreq.c
Linus Torvalds [Sun, 17 Dec 2006 00:01:50 +0000 (16:01 -0800)]
Fix up mm/mincore.c error value cases
Hugh Dickins correctly points out that mincore() is actually _supposed_
to fail on an unmapped hole in the user address space, rather than
return valid ("empty") information about the hole. This just simplifies
the problem further (I had been misled by our previous confusing and
complicated way of doing mincore()).
Also, in the unlikely situation that we can't allocate a temporary
kernel buffer, we should actually return EAGAIN, not ENOMEM, to keep the
"unmapped hole" and "allocation failure" error cases separate.
Finally, add a comment about our stupid historical lack of support for
anonymous mappings. I'll fix that if somebody reminds me after 2.6.20
is out.
Linus Torvalds [Sat, 16 Dec 2006 17:54:23 +0000 (09:54 -0800)]
Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
[PATCH] pata_via: Cable detect error
[PATCH] Fix help text for CONFIG_ATA_PIIX
[PATCH] initializer entry defined twice in pata_rz1000
[PATCH] ata: fix platform_device_register_simple() error check
[PATCH] ahci: do not mangle saved HOST_CAP while resetting controller
[PATCH] libata: don't initialize sg in ata_exec_internal() if DMA_NONE (take #2)
[libata] sata_svw: Disable ATAPI DMA on current boards (errata workaround)
[libata] use kmap_atomic(KM_IRQ0) in SCSI simulator
[PATCH] ata_piix: use piix_host_stop() in ich_pata_ops
[PATCH] ata_piix: IDE mode SATA patch for Intel ICH9
Linus Torvalds [Sat, 16 Dec 2006 17:53:50 +0000 (09:53 -0800)]
Make workqueue bit operations work on "atomic_long_t"
On architectures where the atomicity of the bit operations is handled by
external means (ie a separate spinlock to protect concurrent accesses),
just doing a direct assignment on the workqueue data field (as done by
commit 4594bf159f1962cec3b727954b7c598b07e2e737) can cause the
assignment to be lost due to lack of serialization with the bitops on
the same word.
So we need to serialize the assignment with the locks on those
architectures (notably older ARM chips, PA-RISC and sparc32).
So rather than using an "unsigned long", let's use "atomic_long_t",
which already has a safe assignment operation (atomic_long_set()) on
such architectures.
This requires that the atomic operations use the same atomicity locks as
the bit operations do, but that is largely the case anyway. Sparc32
will probably need fixing.
Architectures (including modern ARM with LL/SC) that implement sane
atomic operations for SMP won't see any of this matter.
Cc: Russell King <rmk+lkml@arm.linux.org.uk> Cc: David Howells <dhowells@redhat.com> Cc: David Miller <davem@davemloft.com> Cc: Matthew Wilcox <matthew@wil.cx> Cc: Linux Arch Maintainers <linux-arch@vger.kernel.org> Cc: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Linus Torvalds [Sat, 16 Dec 2006 17:44:32 +0000 (09:44 -0800)]
Fix incorrect user space access locking in mincore()
Doug Chapman noticed that mincore() will doa "copy_to_user()" of the
result while holding the mmap semaphore for reading, which is a big
no-no. While a recursive read-lock on a semaphore in the case of a page
fault happens to work, we don't actually allow them due to deadlock
schenarios with writers due to fairness issues.
Doug and Marcel sent in a patch to fix it, but I decided to just rewrite
the mess instead - not just fixing the locking problem, but making the
code smaller and (imho) much easier to understand.
Alan [Sat, 16 Dec 2006 14:32:21 +0000 (14:32 +0000)]
[PATCH] pata_via: Cable detect error
The UDMA66 VIA hardware has no controller side cable detect bits we can
use. This patch minimally fixes the problem by reporting unknown in this
case and using drive side detection.
The old drivers/ide code does some additional tricks but those aren't
appropriate now we are in -rc.
Without this update UDMA66 via controllers run slowly. They don't fail so
it's a borderline call whether this is -rc material or not.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Alan [Sat, 16 Dec 2006 12:54:29 +0000 (12:54 +0000)]
[PATCH] Fix help text for CONFIG_ATA_PIIX
> Thanks for clarifying Bill, and sorry Alan. ata_piix does indeed work
> correctly. The help text is a bit confusing:
>
> config ATA_PIIX
> tristate "Intel PIIX/ICH SATA support"
> depends on PCI
> help
> This option enables support for ICH5/6/7/8 Serial ATA.
> If PATA support was enabled previously, this enables
> support for select Intel PIIX/ICH PATA host controllers.
New help text
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Ira Snyder [Fri, 15 Dec 2006 21:08:52 +0000 (13:08 -0800)]
[PATCH] initializer entry defined twice in pata_rz1000
This removes the extra definition of the .error_handler member
in the pata_rz1000 driver.
Signed-off-by: Ira W. Snyder <kernel@irasnyder.com> Cc: Jeff Garzik <jeff@garzik.org> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Jeff Garzik <jeff@garzik.org>
The return value of platform_device_register_simple() should be checked
by IS_ERR().
Cc: Jeff Garzik <jgarzik@pobox.com> Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Tejun Heo [Tue, 12 Dec 2006 11:17:32 +0000 (20:17 +0900)]
[PATCH] ahci: do not mangle saved HOST_CAP while resetting controller
Do not mangle with HOST_CAP while resetting controller. The code is
there for a historical reason. The mangling breaks controller feature
detection and 0 PORTS_IMPL workaround code.
Tejun Heo [Mon, 11 Dec 2006 17:15:31 +0000 (02:15 +0900)]
[PATCH] libata: don't initialize sg in ata_exec_internal() if DMA_NONE (take #2)
Calling sg_init_one() with NULL buf causes oops on certain
configurations. Don't initialize sg in ata_exec_internal() if
DMA_NONE and make the function complain if @buf is NULL when dma_dir
isn't DMA_NONE. While at it, fix comment.
The problem is discovered and initial patch was submitted by Arnd
Bergmann.
Jeff Garzik [Thu, 14 Dec 2006 22:04:33 +0000 (17:04 -0500)]
[libata] sata_svw: Disable ATAPI DMA on current boards (errata workaround)
Current Broadcom/Serverworks SATA boards (including Apple K2 SATA)
have problems with ATAPI DMA, so it is disabled. ATAPI PIO, ATA PIO,
and ATA DMA continue to work just fine.
Acked-by: Anantha Subramanyam <ananth@broadcom.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Tejun Heo [Mon, 11 Dec 2006 13:26:25 +0000 (22:26 +0900)]
[PATCH] ata_piix: use piix_host_stop() in ich_pata_ops
piix_init_one() allocates host private data which should be freed by
piix_host_stop(). ich_pata_ops wasn't converted to piix_host_stop()
while merging, leaking 4 bytes on driver detach. Fix it.
This was spotted using Kmemleak by Catalin Marinas.
Jason Gaston [Thu, 7 Dec 2006 16:57:32 +0000 (08:57 -0800)]
[PATCH] ata_piix: IDE mode SATA patch for Intel ICH9
This updated patch adds the Intel ICH9 IDE mode SATA controller DID's.
Signed-off-by: Jason Gaston <jason.d.gaston@intel.com> Acked-by: Tejun Heo <htejun@gmail.com> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Roland Dreier [Sat, 16 Dec 2006 04:55:28 +0000 (20:55 -0800)]
IB/mthca: Use DEFINE_MUTEX() instead of mutex_init()
mthca_device_mutex() can be initialized automatically with
DEFINE_MUTEX() rather than explicitly calling mutex_init(). This
saves a bit of text and shrinks the source by a line, so we may as
well do it....
Linus Torvalds [Fri, 15 Dec 2006 22:13:51 +0000 (14:13 -0800)]
Fix "delayed_work_pending()" macro expansion
Nobody uses it, but it was still wrong. Using the macro argument name
'work' meant that when we used 'work' as a member name, that would also
get replaced by the macro argument.
Roland Dreier [Fri, 15 Dec 2006 22:01:49 +0000 (14:01 -0800)]
IB/srp: Fix FMR mapping for 32-bit kernels and addresses above 4G
struct srp_device.fmr_page_mask was unsigned long, which means that
the top part of addresses above 4G was being chopped off on 32-bit
architectures. Of course nothing good happens when data from SRP
targets is DMAed to the wrong place.
Fix this by changing fmr_page_mask to u64, to match the addresses
actually used by IB devices.
Thanks to Brian Cain <Brian.Cain@ge.com> and David McMillen
<davem@systemfabricworks.com> for help diagnosing the bug and testing
the fix.
Roland Dreier [Fri, 15 Dec 2006 21:57:26 +0000 (13:57 -0800)]
IB: Fix ib_dma_alloc_coherent() wrapper
The ib_dma_alloc_coherent() wrapper uses a u64* for the dma_handle
parameter, unlike dma_alloc_coherent, which uses dma_addr_t*. This
means that we need a temporary variable to handle the case when
ib_dma_alloc_coherent() just falls through directly to
dma_alloc_coherent() on architectures where sizeof u64 != sizeof
dma_addr_t.
Linus Torvalds [Fri, 15 Dec 2006 16:43:13 +0000 (08:43 -0800)]
Remove stack unwinder for now
It has caused more problems than it ever really solved, and is
apparently not getting cleaned up and fixed. We can put it back when
it's stable and isn't likely to make warning or bug events worse.
In the meantime, enable frame pointers for more readable stack traces.
Stefan Bader [Fri, 15 Dec 2006 16:18:30 +0000 (17:18 +0100)]
[S390] cio: css_register_subchannel race.
Asynchronous probe can release memory of a subchannel before
css_get_ssd_info is called. To fix this call css_get_ssd_info
before registering with driver core.
Signed-off-by: Stefan Bader <shbader@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Michael Holzheu [Fri, 15 Dec 2006 16:18:27 +0000 (17:18 +0100)]
[S390] Save prefix register for dump on panic
The dump tools expect that the saved prefix register points to the
lowcore of the dump cpu. Since we set the prefix register to 0 during
reipl/dump, we have to save the original prefix register. Before we
start the dump program, we copy the original prefix register to the
designated location in the lowcore.
Signed-off-by: Michael Holzheu <holzheu@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Michael Holzheu [Fri, 15 Dec 2006 16:18:22 +0000 (17:18 +0100)]
[S390] Fix reboot hang on LPARs
Reboot hangs on LPARs without diag308 support. The reason for this is,
that before the reboot is done, the channel subsystem is shut down.
During the reset on each possible subchannel a "store subchannel" is
done. This operation can end in a program check interruption, if the
specified subchannel set is not implemented by the hardware. During
the reset, currently we do not have a program check handler, which
leads to the described kernel bug. We install now a new program check
handler for the reboot code to fix this problem.
Signed-off-by: Michael Holzheu <holzheu@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Ursula Braun [Fri, 15 Dec 2006 16:18:14 +0000 (17:18 +0100)]
[S390] Hipersocket multicast queue: make sure outbound handler is called
A HiperSocket multicast queue works asynchronously. When sending
buffers, the buffer state change from PRIMED to EMPTY may happen
delayed. Reschedule the checking for changes in the outbound queue,
if there are still PRIMED buffers.
Signed-off-by: Ursula Braun <braunu@de.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Florian Festi [Thu, 14 Dec 2006 10:59:11 +0000 (11:59 +0100)]
input/hid: Supporting more keys from the HUT Consumer Page
On USB keyboards lots of hot/internet keys are not working. This patch
adds support for a number of keys from the USB HID Usage Table
(http://www.usb.org/developers/devclass_docs/Hut1_12.pdf).
It also adds several new key codes. Most of them are used on real world
keyboards I know. I added some others (KEY_+ EDITOR, GRAPHICSEDITOR, DATABASE,
NEWS, VOICEMAIL, VIDEOPHONE) to avoid "holes".
I also added KEY_ZOOMRESET as it is possible to have a inet keyboard and a
remote control in parallel and it makes sense to have them behave differently.
projects / linux-beck.git / log