Jason Wessel [Wed, 23 Jul 2008 16:30:15 +0000 (11:30 -0500)]
kgdb, powerpc: arch specific powerpc kgdb support
This patch removes the old kgdb reminants from ARCH=powerpc and
implements the new style arch specific stub for the common kgdb core
interface.
It is possible to have xmon and kgdb in the same kernel, but you
cannot use both at the same time because there is only one set of
debug hooks.
The arch specific kgdb implementation saves the previous state of the
debug hooks and restores them if you unconfigure the kgdb I/O driver.
Kgdb should have no impact on a kernel that has no kgdb I/O driver
configured.
Jason Wessel [Wed, 20 Feb 2008 19:33:40 +0000 (13:33 -0600)]
kgdb: support for ARCH=arm
This patch adds the ARCH=arm specific a kgdb backend, originally
written by Deepak Saxena <[email protected]> and George Davis
<[email protected]>. Geoff Levand <[email protected]>,
Nicolas Pitre, Manish Lachwani, and Jason Wessel have contributed
various fixups here as well.
The KGDB patch makes one change to the core ARM architecture such that
the traps are initialized early for use with the debugger or other
subsystems.
Henrik Rydberg [Tue, 22 Jul 2008 13:54:05 +0000 (15:54 +0200)]
HID: apple_fn_keys F5 and F6
In drivers/hid/hid-input.c, the apple_fn_keys translation table
contains no entries for the F5 and F6 keys. This patch maps the F5 and
F6 keys to KBDILLUMDOWN and KBDILLUMUP, respektively, which make them
work as keyboard backlight control keys.
This patch completely removes the dependency of usbkbd and usbmouse drivers
on usbhid code.
usbkbd/usbmouse drivers are not needed in the vast majority of cases anyway,
and they shouldn't be loaded in standard configurations at all. They are supposed
to be as trivial as possible, and searching for HID_QUIRK_IGNORE can even be
bad in some cases.
Dan Nicholson [Sat, 14 Jun 2008 17:26:24 +0000 (10:26 -0700)]
HID: Invert HWHEEL mappings for some Logitech mice
Some Logitech mice have a tilt wheel which register as HWHEEL buttons.
The events are positive for a click to the right and negative for a
click to the left. Applications expect the opposite, though.
I suspect this affects a lot more Logitech mice, but these are the only
two I have. I tested this using evtest and a GTK application. A similar
Microsoft Intellimouse I have works correctly.
HID: add fn key support on Macbook Pro 4,1 and Macbook Air
Added device ids to hid-quriks for detection of keyboards on 4th
generation Macbook Pro and Macbook Air
The naming scheme is consistent with past Apple keyboards in hid-quirks;
as defined by Apple (including device ids) in:
/System/Library/Extensions/AppleUSBTopCase.kext/Contents/PlugIns/AppleUSBTCKeyboard.kext/Co
ntents/Info.plist
Patch was originally posted and tested at:
https://bugs.launchpad.net/mactel-support/+bug/207127
Alan Cox [Mon, 26 May 2008 09:25:20 +0000 (11:25 +0200)]
HID: Switch hiddev to unlocked_ioctl
Push down the BKL. In some cases compat_ioctl already doesn't take the
BKL so we don't either. Some of the locking here seems already dubious
and object lifetimes want documenting
This is alternative implementation of sg content iterator introduced
by commit 83e7d317... from Pierre Ossman in next-20080716. As there's
already an sg iterator which iterates over sg entries themselves, name
this sg_mapping_iterator.
Slightly edited description from the original implementation follows.
Iteration over a sg list is not that trivial when you take into
account that memory pages might have to be mapped before being used.
Unfortunately, that means that some parts of the kernel restrict
themselves to directly accesible memory just to not have to deal with
the mess.
This patch adds a simple iterator system that allows any code to
easily traverse an sg list and not have to deal with all the details.
The user can decide to consume part of the iteration. Also, iteration
can be stopped and resumed later if releasing the kmap between
iteration steps is necessary. These features are useful to implement
piecemeal sg copying for interrupt drive PIO for example.
Pierre Ossman [Tue, 22 Jul 2008 12:38:35 +0000 (14:38 +0200)]
mmc_test: print message when attaching to card
Make it a bit more obvious that the card has been claimed by the
mmc_test driver so that people don't have to wonder why their block
device never shows up.
Pierre Ossman [Sun, 20 Jul 2008 22:45:15 +0000 (00:45 +0200)]
sdhci: add warnings for bad buffers in ADMA path
The ADMA code path assumes that the 3 byte alignment fix doesn't cross
a page boundary. I'm not convinced this is worth supporting, but at
least print a warning in the off chance we'll actually see such a request.
Pierre Ossman [Sun, 20 Jul 2008 22:14:52 +0000 (00:14 +0200)]
mmc_test: test oversized sg lists
Add tests that make sure the driver properly checks the blocks and
blksz fields and doesn't assume the sg list has a size that perfectly
matches the current request.
Pierre Ossman [Fri, 4 Jul 2008 16:17:13 +0000 (18:17 +0200)]
mmc_test: highmem tests
Add a couple of tests to make sure the host driver handles highmem
memory pages properly. Unfortunately there is no way to guarantee an
allocation below 4 GB in i386, so it might give you addresses that
are out of reach for the hardware (OTOH, so will any other highmem
allocation in the kernel).
Ben Dooks [Thu, 17 Jul 2008 14:32:54 +0000 (15:32 +0100)]
s3cmci: ensure host stopped on machine shutdown
Ensure that the s3cmci host controller is turned off
when the machine is shutdown, otherwise we end up
leaving the card powered and processing insertion and
removal events after the system prints "System halted."
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (82 commits)
ipw2200: Call netif_*_queue() interfaces properly.
netxen: Needs to include linux/vmalloc.h
[netdrvr] atl1d: fix !CONFIG_PM build
r6040: rework init_one error handling
r6040: bump release number to 0.18
r6040: handle RX fifo full and no descriptor interrupts
r6040: change the default waiting time
r6040: use definitions for magic values in descriptor status
r6040: completely rework the RX path
r6040: call napi_disable when puting down the interface and set lp->dev accordingly.
mv643xx_eth: fix NETPOLL build
r6040: rework the RX buffers allocation routine
r6040: fix scheduling while atomic in r6040_tx_timeout
r6040: fix null pointer access and tx timeouts
r6040: prefix all functions with r6040
rndis_host: support WM6 devices as modems
at91_ether: use netstats in net_device structure
sfc: Create one RX queue and interrupt per CPU package by default
sfc: Use a separate workqueue for resets
sfc: I2C adapter initialisation fixes
...
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
sparc32: pass -m32 when building vmlinux.lds
sparc: Fixes the DRM layer build on sparc.
ide: merge <asm-sparc/ide_64.h> with <asm-sparc/ide_32.h>
ide: <asm-sparc/ide_64.h>: use __raw_{read,write}w()
ide: <asm-sparc/ide_32.h>: use __raw_{read,write}w()
ide: <asm-sparc/ide_64.h>: use %r0 for outw_be()
sparc64: Do not define BIO_VMERGE_BOUNDARY.
netif_carrier_{on,off}() handles starting and stopping packet
flow into the driver. So there is no reason to invoke netif_stop_queue()
and netif_wake_queue() in response to link status events.
Maciej Sosnowski [Wed, 23 Jul 2008 00:30:57 +0000 (17:30 -0700)]
I/OAT: I/OAT version 3.0 support
This patch adds to ioatdma and dca modules
support for Intel I/OAT DMA engine ver.3 (aka CB3 device).
The main features of I/OAT ver.3 are:
* 8 single channel DMA devices (8 channels total)
* 8 DCA providers, each can accept 2 requesters
* 8-bit TAG values and 32-bit extended APIC IDs
Maciej Sosnowski [Wed, 23 Jul 2008 00:30:57 +0000 (17:30 -0700)]
I/OAT: tcp_dma_copybreak default value dependent on I/OAT version
I/OAT DMA performance tuning showed different optimal values of
tcp_dma_copybreak for different I/OAT versions (4096 for 1.2 and 2048
for 2.0). This patch lets ioatdma driver set tcp_dma_copybreak value
according to these results.
This patch reworks the error handling in r6040_init_one
in order not to leak resources and correcly unmap and release
PCI regions of the MAC. Also prefix printk's with the driver name
for clarity.
This patch bumps the release of the r6040 driver. There has been
quite some versions of it out there, but this one is the one
people should report bugs against.
r6040: handle RX fifo full and no descriptor interrupts
This patch allows the MAC to handle the RX FIFO full
and no descriptor available interrupts. While we are at it
replace the TX interrupt with its corresponding definition.
This patch changes the default waiting time of a packet, which
along with our previous r6040_rx path, was causing huge delays
with another host (160 to 230 ms).
This patch completely reworks the RX path in order to be
more accurate about what is going on with the MAC.
We no longer read the error from the MLSR register instead read
the descriptor status register which reflects, the error per descriptor.
We now allocate skbs on the fly in r6040_rx, and we handle allocation
failure instead of simply dropping the packet. Remove the
rx_free_desc counter of r6040_private structure since we
allocate skbs in the RX path.
r6040_rx_buf_alloc is now removed and becomes unuseless.
Joseph Fannin <[email protected]> and Takashi Iwai <[email protected]>
noticed that commit 073a345c04b01da0cc5b79ac7be0c7c8b1691ef5
("mv643xx_eth: clarify irq masking and unmasking") broke the
mv643xx_eth build when NETPOLL is enabled, due to it not renaming
one instance of INT_CAUSE_EXT in mv643xx_eth_netpoll(). This patch
takes care of that instance as well.
Rework the RX buffers allocation function so that we do not
leak memory in the case we could not allocate skbs for the
RX path. Propagate the errors to the r6040_up function
where we call the RX buffers allocation function.
Also rename the r6040_alloc_txbufs function to
r6040_init_txbufs, to reflect what it really does.
r6040: fix scheduling while atomic in r6040_tx_timeout
Add a helper function which only modifies R6040 MAC registers
use it when we timeout, and on adapter initialization. Fix
the scheduling while atomic but in the timeout routine due
to the reallocation of rx/tx buffers.
Ben Hutchings [Fri, 18 Jul 2008 18:03:10 +0000 (19:03 +0100)]
sfc: Create one RX queue and interrupt per CPU package by default
Using multiple cores in the same package to handle received traffic
does not appear to provide a performance benefit. Therefore use CPU
topology information to count CPU packages and use that as the default
number of RX queues and interrupts. We rely on interrupt balancing to
spread the interrupts across packages.
Ben Hutchings [Fri, 18 Jul 2008 17:59:12 +0000 (18:59 +0100)]
sfc: I2C adapter initialisation fixes
As recommended by Jean Delvare:
- Increase timeout to 50 ms
- Leave adapter class clear so that unwanted drivers do not probe our bus
- Use strlcpy() for name initialisation
This patch makes e1000 driver ioport-free.
This corrects behavior in probe function so as not to request ioport
resources as long as they are not really needed. This is based on the
ioport-free patch of e1000 driver from Auke Kok and Tomohiro Kusumi.
The ixgbe driver was untested with device ID 8086:10c8 but still advertises
support. Currently if this device is present in the system when the driver
is loaded, the system will panic.
Remove this device ID until full support can be tested with available
hardware. This patch is necessary for 2.6.24, 2.6.25 and 2.6.26
Anton Vorontsov [Tue, 22 Jul 2008 22:41:24 +0000 (15:41 -0700)]
netdev: bunch of drivers: avoid WARN at net/core/dev.c:1328
The drivers were touching net queue before it has been started, so
without this patch, the drivers will potentially WARN at
net/core/dev.c:1328.
I don't have the hardware for the drivers below, so this patch is
untested, and thus should be carefully peer reviewed.
tc35815.c
au1000_eth.c
bfin_mac.c
macb.c
^ The four drivers are using phylib, they're calling netif_start_queue()
in open() callback. So trivially remove netif_tx_schedule_all().
Phylib will handle netif_carrier_*().
cpmac.c
fec_mpc52xx.c
fs_enet/fs_enet-main.c
sh_eth.c
^ The same as above, but these were also needlessly calling
netif_carrier_*() functions. So removed queue calls and also remove
carrier calls, since phylib will handle it. fs_enet-main.c also didn't
call netif_start_queue() at open(), this is fixed now.
Dave Jones [Tue, 15 Jul 2008 23:54:53 +0000 (19:54 -0400)]
8139too: Make PIO/MMIO a runtime decision through a module parameter.
This is needed to support devices that only work with PIO without
penalising devices that work fine with MMIO in distro kernels.
It also allows us to eventually use PIO as a fallback when setting
up MMIO fails.
Wang Jian [Wed, 16 Jul 2008 13:46:17 +0000 (21:46 +0800)]
net/phy: Fix copper/fiber auto-selection for 88e1111
The 27.15 bit (MII_M1111_HWCFG_FIBER_COPPER_AUTO) is disable bit. When
set to 1, copper/fiber auto selection is disabled. The current code
to enable but actually disable auto selection.
NX3031 supports cut-through operation where ingress packets are
directly dma'ed into host buffers to reduce latency.
This requires larger dma buffers (2kb) and different alignemnt.
The buffer posting logic is changed a bit. The free rx buffers
are maintained in linked list, since the received reference
handles can be out of order. However rx descriptors are still
posted sequentially, indexed by producer.
MAC addr, multicast filters, mtu are set through firmware commands
in firmware v4.0.0+ because of virtualization of physical ports.
Link status is also read from registers allocated by firmware for
each virtual port.
New revision of netxen chip has 2MB PCI memory. Older chips
had 128MB addressable PCI memory. To retain compatibility,
this patch adds function pointers based on pci bar0 size.
Enable multicast address filtering capabilities in the hardware.
Upto 16 multicast addresses can be programmed for each physical
port. Support "allmulti" mode, if enabled.
o Reduce access to global arrays in data path.
o Remove duplicate/unused variables, unecessary alignment constraints.
o Use correct pci_dev instead of fallback device for consistent
allocations.
o Disable ethtool set_eeprom functionality for now, it was only used
for flashing firmware.
projects / linux.git / log