09 May, 2007
1 commit
-
Remove includes of where it is not used/needed.
Suggested by Al Viro.Builds cleanly on x86_64, i386, alpha, ia64, powerpc, sparc,
sparc64, and arm (all 59 defconfigs).Signed-off-by: Randy Dunlap
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
07 May, 2007
1 commit
-
* 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm: (82 commits)
[ARM] Add comments marking in-use ptrace numbers
[ARM] Move syscall saving out of the way of utrace
[ARM] 4360/1: S3C24XX: regs-udc.h remove unused macro
[ARM] 4358/1: S3C24XX: mach-qt2410.c: remove linux/mmc/protocol.h header
[ARM] mm 10: allow memory type to be specified with ioremap
[ARM] mm 9: add additional device memory types
[ARM] mm 8: define mem_types table L1 bit 4 to be for ARMv6
[ARM] iop: add missing parens in macro
[ARM] mm 7: remove duplicated __ioremap() prototypes
ARM: OMAP: fix OMAP1 mpuio suspend/resume oops
ARM: OMAP: MPUIO wake updates
ARM: OMAP: speed up gpio irq handling
ARM: OMAP: plat-omap changes for 2430 SDP
ARM: OMAP: gpio object shrinkage, cleanup
ARM: OMAP: /sys/kernel/debug/omap_gpio
ARM: OMAP: Implement workaround for GPIO wakeup bug in OMAP2420 silicon
ARM: OMAP: Enable 24xx GPIO autoidling
[ARM] 4318/2: DSM-G600 Board Support
[ARM] 4227/1: minor head.S fixups
[ARM] 4328/1: Move i.MX UART regs to driver
...
28 Apr, 2007
7 commits
-
Update gadget_chip.c, ether.c for newly added Freescale Highspeed USB
device driver.Signed-off-by: Li Yang
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Freescale high-speed USB SOC can be found on some Freescale processors
among different architectures. It supports both host and device functions.
This driver adds its device support for Linux USB Gadget layer.
It is tested on MPC8349 and MPC8313, but should work on other platforms
with minor tweaks. The driver passed USBCV 1.3 compliance tests. Note
that this driver doesn't yet include OTG support.Signed-off-by: Li Yang
Signed-off-by: Jiang Bo
Signed-off-by: Bruce Schmid
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Another workaround for the glitch in the network layer, whereby one call
ignores the (otherwise kernel-wide) convention that free() calls should
not oops when passed nulls. This code already handles that API glitch in
most other paths.From: Erik Hovland
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
skb_push function may return a pointer which is not aligned as required
by struct rndis_packet_msg_type. Using attribute trick to fix this bug.Signed-off-by: Roy Huang
Signed-off-by: Jie Zhang
Signed-off-by: Bryan Wu
Cc: David Brownell
Signed-off-by: Andrew Morton
Signed-off-by: Greg Kroah-Hartman -
With CONFIG_BLOCK=n, this build error happens:
WARNING: "bdev_read_only" [drivers/usb/gadget/g_file_storage.ko] undefined!Signed-off-by: Randy Dunlap
Acked-by: Alan Stern
Signed-off-by: Greg Kroah-Hartman -
This patch changes last use of hardcoded number of irq to
use platfrom_get_irq.Signed-off-by: Milan Svoboda
Acked-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Make the pxa2xx_udc driver fetch its IRQ from platform resources
rather than using compile-time constants, so that it works properly
on IXP4xx systems not just PXA21x/25x/26x.Other updates:
- Do that using platform_get_irq()
- Switch to platform_driver_probe()
- Handle device_add() errors
- Remove "function" sysfs attribute and its potential errors
- Whitespace cleanupsSigned-off-by: Milan Svoboda
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman
26 Apr, 2007
1 commit
-
One less thing for drivers writers to worry about.
Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller
22 Apr, 2007
1 commit
-
This patch removes the unnecessary bit number from CKENnn_XXXX
definitions for PXA, so thatCKEN0_PWM0 --> CKEN_PWM0
CKEN1_PWM1 --> CKEN_PWM1
...
CKEN24_CAMERA --> CKEN_CAMERAThe reasons for the change of these defitions are:
1. they do not scale - they are currently valid for pxa2xx, but
definitely not valid for pxa3xx, e.g., pxa3xx has bit 3 for camera
instead of bit 242. they are unnecessary - the peripheral name within the definition
has already announced its usage, we don't need those bit numbers
to know which peripheral we are going to enable/disable clock for3. they are inconvenient - think about this: a driver programmer
for pxa has to remember which bit in the CKEN register to turn
on/offAnother change in the patch is to make the definitions equal to its
clock bit index, so that#define CKEN_CAMERA (24)
instead of
#define CKEN_CAMERA (1 << 24)
this change, however, will add a run-time bit shift operation in
pxa_set_cken(), but the benefit of this change is that it scales
when bit index exceeds 32, e.g., pxa3xx has two registers CKENA
and CKENB, totally 64 bit for this, suppose CAMERA clock enabling
bit is CKENB:10, one can simply define CKEN_CAMERA to be (32 + 10)
and so that pxa_set_cken() need minimum change to adapt to that.Signed-off-by: eric miao
Signed-off-by: Russell King
27 Mar, 2007
1 commit
-
Various fixes to omap_udc, noted with some recent testing:
- Cope with some SMP-induced braindamage in ARM's dma_{alloc,free}_coherent()
implementation: alloc() can be called with IRQs blocked, but since late
last year that's no longer true for free(). This resolves really NASTY
problems with logspamming via WARN_ON(), indicating N-page leaks.- Be more correct in handling GET_STATUS request for RECIP_ENDPOINT ... the
previous code only handled RECIP_INTERFACE, this version should be correct
except for (sigh) bulk/interrupt endpoints.- Provide a better name for the function reporting whether the board has
vbus sensing wired up.GET_STATUS requests for endpoint status still acts strangely though, at least
given one flakey host doesn't always ack the first DATA packet, then the packet
that gets retransmitted doesn't have data!Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman
11 Mar, 2007
1 commit
-
This reverts commit d2487cb4257dafb686f682285854fe7f02ca29d8.
Russell King points out that it's obviously bogus, and I have to agree.
Not only does "irq" not even exist in that scope, but we obviously need
to free the irq that we actually requested, and that's IRQ_USB.Reported-by: Russell King
Cc: Andrew Morton
Cc: Greg KH
Cc: David Brownell ,
Cc: Milan Svoboda
Signed-off-by: Linus Torvalds
10 Mar, 2007
3 commits
-
This patch changes last use of hardcoded number of irq to
use platfrom_get_irq.Signed-off-by: Milan Svoboda
Acked-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This is deep architecture specific magic and does should not to exist
in a driver.Signed-off-by: Ralf Baechle
Cc: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Rename the driver struct used with at91_udc to prevent yet another
bogus warning from "modpost".Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman
24 Feb, 2007
1 commit
-
I don't think the current code works with multiple iovecs.
The original would just copy the first part of priv->buf
over and over into multiple iovecs.Signed-off-by: Sarah Bailey
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman
20 Feb, 2007
1 commit
-
* 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm: (117 commits)
[ARM] 4058/2: iop32x: set ->broken_parity_status on n2100 onboard r8169 ports
[ARM] 4140/1: AACI stability add ac97 timeout and retries
[ARM] 4139/1: AACI record support
[ARM] 4138/1: AACI: multiple channel support for IRQ handling
[ARM] 4211/1: Provide a defconfig for ns9xxx
[ARM] 4210/1: base for new machine type "NetSilicon NS9360"
[ARM] 4222/1: S3C2443: Remove reference to missing S3C2443_PM
[ARM] 4221/1: S3C2443: DMA support
[ARM] 4220/1: S3C24XX: DMA system initialised from sysdev
[ARM] 4219/1: S3C2443: DMA source definitions
[ARM] 4218/1: S3C2412: fix CONFIG_CPU_S3C2412_ONLY wrt to S3C2443
[ARM] 4217/1: S3C24XX: remove the dma channel show at startup
[ARM] 4090/2: avoid clash between PXA and SA1111 defines
[ARM] 4216/1: add .gitignore entries for ARM specific files
[ARM] 4214/2: S3C2410: Add Armzone QT2410
[ARM] 4215/1: s3c2410 usb device: per-platform vbus_draw
[ARM] 4213/1: S3C2410 - Update definition of ADCTSC_XY_PST
[ARM] 4098/1: ARM: rtc_lock only used with rtc_cmos
[ARM] 4137/1: Add kexec support
[ARM] 4201/1: SMP barriers pair needed for the secondary boot process
...Fix up conflict due to typedef removal in sound/arm/aaci.h
18 Feb, 2007
1 commit
17 Feb, 2007
2 commits
-
This is a runtime codespace shrink: in most cases, platform devices should
put probe() should in the init section, and remove() in the exit section.
And I have no idea why the module init/exit routines were mismarked.
It also moves one function table into read-only data.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
A SET_LINE_CODING control request should return a zero length packet
as an ACK to the host, during the status phase of a USB transaction.The return value of gs_setup_class() is treated as the number of
bytes to write in the status phase of the control request, by
gs_setup(). For this case, the value returned by gs_setup_class should
be zero for SET_LINE_CODING but, right now, appears to be
sizeof(struct usb_cdc_line_coding).However, if after doing the memcpy of the line coding descriptor we
set the variable "ret" to be zero, we should return the appropiate ZLP
to the host as an ACK in the status phase of the control request.
I've tested this out using Linux as both host and slave and confirmed
that the following small change fixes the spurious return of
sizeof(struct usb_cdc_line_coding)/wLength bytes in the status phase
of a USB_CDC_REQ_SET_LINE_CODING request. It's not a huge bug but, it
is worth fixing.Signed-off-by: Bryan O'Donoghue
Signed-off-by: Greg Kroah-Hartman
15 Feb, 2007
2 commits
-
The obsolete SA_xxx interrupt flags have been used despite the scheduled
removal. Fixup the remaining users.Signed-off-by: Thomas Gleixner
Acked-by: Ingo Molnar
Cc: "Luck, Tony"
Cc: Roman Zippel
Cc: Geert Uytterhoeven
Cc: Jeff Garzik
Cc: Wim Van Sebroeck
Cc: Roland Dreier
Cc: Alessandro Zummo
Cc: James Bottomley
Cc: Greg KH
Cc: Dave Airlie
Cc: James Simmons
Cc: "Antonino A. Daplas"
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
After Al Viro (finally) succeeded in removing the sched.h #include in module.h
recently, it makes sense again to remove other superfluous sched.h includes.
There are quite a lot of files which include it but don't actually need
anything defined in there. Presumably these includes were once needed for
macros that used to live in sched.h, but moved to other header files in the
course of cleaning it up.To ease the pain, this time I did not fiddle with any header files and only
removed #includes from .c-files, which tend to cause less trouble.Compile tested against 2.6.20-rc2 and 2.6.20-rc2-mm2 (with offsets) on alpha,
arm, i386, ia64, mips, powerpc, and x86_64 with allnoconfig, defconfig,
allmodconfig, and allyesconfig as well as a few randconfigs on x86_64 and all
configs in arch/arm/configs on arm. I also checked that no new warnings were
introduced by the patch (actually, some warnings are removed that were emitted
by unnecessarily included header files).Signed-off-by: Tim Schmielau
Acked-by: Russell King
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
12 Feb, 2007
2 commits
-
Bug: pnx8550 code creates directory but resets ->nlink to 1.
create_proc_entry() et al will correctly set ->nlink for you.
Signed-off-by: Alexey Dobriyan
Cc: Ralf Baechle
Cc: Benjamin Herrenschmidt
Cc: Paul Mackerras
Cc: Jeff Dike
Cc: Corey Minyard
Cc: Alan Cox
Cc: Kyle McMartin
Cc: Martin Schwidefsky
Cc: Greg KH
Cc: Ingo Molnar
Cc: Thomas Gleixner
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Convert all calls to invalidate_inode_pages() into open-coded calls to
invalidate_mapping_pages().Leave the invalidate_inode_pages() wrapper in place for now, marked as
deprecated.Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
08 Feb, 2007
15 commits
-
This patch renames pxa_gpio_set/get functions defined in drivers/usb/gadget/pxa2xx_udc.h to udc_gpio_set/get.
These functions are moved from drivers/usb/gadget/pxa2xx_udc.h to include/asm-arm/arch-pxa2xx/udc.h
Creates new functions: udc_gpio_to_irq, udc_gpio_init_vbus, udc_gpio_init_pullup in include/asm-arm/arch-pxa2xx/udc.h. These functions are used in drivers/usb/gadget/pxa2xx_udc.c instead of direct low-level (pxa2xx only) functions.
Creates all these udc_gpio_* functions in include/asm-arm/arch-ixp4xx/udc.h. This implementation has no real code because ixp4xx doesn't use vbus - only vbus uses all these gpio functions (and because ixp4xx misses any function which converts number of gpio pin into it's irq).
This is next step to make pxa2xx_udc fully work on ixp4xx platform.
Signed-off-by: Milan Svoboda
Signed-off-by: Russell King -
Add support for the Atmel AT91SAM9263 processor. It is similar to the
AT91SAM9260 but with more integrated peripherals, 5 GPIO banks, etc.Original patch from Nicolas Ferre.
Signed-off-by: Andrew Victor
Signed-off-by: Russell King -
Define DEV_CONFIG_CDC when compiling for HUSB2DEV.
From: Håvard Skinnemoen
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This identifies the driver for the Atmel HUSB2 Device Controller,
as integrated into the first AVR32 chip, the AT32AP700.From: Håvard Skinnemoen
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This patch (as837) fixes several mistakes in the AIO interface of the
gadgetfs driver:The ki_retry method is not supposed to do a put on the kiocb.
The extra call to aio_put_req() causes memory corruption.
(Note: This call was removed before, by patch as691, and then
mysteriously re-introduced later.)Even if a read transfer is cancelled, we can and should send
to the user all the data that did manage to get transferred.Testing for AIO cancellation in the I/O completion handler
is both racy and (now) unnecessary. aio_complete() does its
own checking, in a safe manner.Signed-off-by: Alan Stern
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Resolve an initizlization issue that could come up if the userspace
driver wrote invalid descriptors to a dual-speed device.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This resolves a race in gadgetfs associated with changing device/ep0
when processing control requests. The fix is to change that state
earlier, when the control response is issued, so there's no window
in which userspace could see the wrong state; and enlarge the scope
of the spinlock during the ep0 request completion handler.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This simplifies event reading by eliminating arithmetic and being
more direct/obvious, and tweaks some debug messages slightly.
The math elimination will change timings, sometimes enough to
allow a race to appear.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Minor gadgetfs cleanups:
- EP0 state constants become consistently STATE_DEV_* rather than
sometimes omitting the "DEV_"; STATE_EP_* were already consistent.- Comment that ep0 state is protected by the spinlock, and update
code that was neglecting that rule.None of this is expected to change behavior.
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This updates the AT91 UDC driver's handling of wakeup events:
- Fix a bug in the original scheme, which was never updated after
the {enable,disable}_irq_wake() semantics were updated to address
refcounting issues (i.e. behave for shared irqs).- Couple handling of both type of wakeup events, to be more direct. The
controller can be source of wakeup events for cases like bus reset
and USB resume. On some boards, VBUS sensing is also IRQ driven.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Gadgetfs had a mode in which endpoint descriptors were written by the user
program before connection. This mode had some bugs, and hasn't seen much
(if any) use. This patch removes that mode, leaving the mode of operation
where the user program waits for endpoint 0 to report a SET_CONFIGURATION,
and only then configures the endpoints.From: "Phil Endecott"
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Remove some whitespace bugs in gadgetfs (mostly from someone's
patch updating the AIO support).Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
Went looking through some usb stuff and found some unnecessary casts in
file_storage.c This is part of the KernelJanitors TODO list.Signed-off-by: John Daiker
Acked-by: Alan Stern
Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
It turns out that minor tweaks to the "CDC Subset" support in the Ethernet
gadget driver, just updating a config descriptor, let it be automagically
recognized by a Windows driver supported by MCCI.This patch adds those descriptors, so systems using PXA 255 processors
(like Gumstix etc) can interop with those commercial MS-Windows drivers.
This is a Good Thing since Microsoft's RNDIS code has bugginess issues,
which are unfortunately compounded by "won't fix" issues as well as "the
published specs are incomplete and wrong" issues. Being able to talk to
the MCCI driver gives Windows users another connectivity option. (MCCI
also has CDC Ethernet drivers, which can help most non-PXA processors.)Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman -
This moves to to reduce some of the
clutter of usb header files.Signed-off-by: David Brownell
Signed-off-by: Greg Kroah-Hartman
