openwrt/staging/blogic.git
19 years ago[PATCH] pegasus.h
Petko Manolov [Mon, 3 Oct 2005 14:24:11 +0000 (17:24 +0300)]
[PATCH] pegasus.h

This one is a tiny patch adding one more device to the list.  Please
apply. :)

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/net/pegasus.h |    2 ++
 1 file changed, 2 insertions(+)

19 years ago[PATCH] drivers/usb/serial/ftdi_sio: add PID/VID
Thomas Riewe [Thu, 29 Sep 2005 12:57:29 +0000 (14:57 +0200)]
[PATCH] drivers/usb/serial/ftdi_sio: add PID/VID

We would like to add a PID for the Pyramid Appliance Display, which works
on USB via FTDI_SIO.

Signed-off-by: Thomas Riewe <thomasr@pyramid.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/serial/ftdi_sio.c |    1 +
 drivers/usb/serial/ftdi_sio.h |    3 +++
 2 files changed, 4 insertions(+)

19 years ago[PATCH] USB: make wHubCharacteristics __le16 to match other usb descriptor fields
Greg Kroah-Hartman [Tue, 21 Jun 2005 04:15:16 +0000 (21:15 -0700)]
[PATCH] USB: make wHubCharacteristics __le16 to match other usb descriptor fields

Also has the nice benefit of making sparc alignment issues go away.

Thanks to David Miller for pointing out the problems here.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hub.c |   22 ++++++++++++----------
 drivers/usb/core/hub.h |    2 +-
 2 files changed, 13 insertions(+), 11 deletions(-)

19 years ago[PATCH] USB: File-Storage gadget: use the kthread API
Alan Stern [Mon, 26 Sep 2005 20:12:02 +0000 (16:12 -0400)]
[PATCH] USB: File-Storage gadget: use the kthread API

This patch (as566) converts the File-Storage gadget over to the kthread
API.  The new code doesn't use kthread_stop because the control thread
needs to terminate asynchronously when it receives a signal.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/gadget/file_storage.c |   32 +++++++++++++-------------------
 1 file changed, 13 insertions(+), 19 deletions(-)

19 years ago[PATCH] USB: sisusb warning fix
Andrew Morton [Fri, 23 Sep 2005 06:55:22 +0000 (23:55 -0700)]
[PATCH] USB: sisusb warning fix

drivers/usb/misc/sisusbvga/sisusb.c: In function `sisusb_reset_text_mode':
drivers/usb/misc/sisusbvga/sisusb.c:2461: warning: assignment discards qualifiers from pointer target type

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/misc/sisusbvga/sisusb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

19 years ago[PATCH] USB: UHCI: Spruce up some comments
Alan Stern [Mon, 26 Sep 2005 20:31:15 +0000 (16:31 -0400)]
[PATCH] USB: UHCI: Spruce up some comments

This patch (as570) changes some comments in the uhci-hcd header file and
removes an unused declaration (something I forgot to erase in an earlier
patch).

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/uhci-hcd.h |   91 +++++++++++++++++++++++---------------------
 1 file changed, 49 insertions(+), 42 deletions(-)

19 years ago[PATCH] USB: usb_bulk_message() handles interrupts endpoints
Alan Stern [Mon, 26 Sep 2005 20:22:45 +0000 (16:22 -0400)]
[PATCH] USB: usb_bulk_message() handles interrupts endpoints

Because there is no bulk_interrupt_message() routine and no
USBDEVFS_INTERRUPT ioctl, people have been forced to abuse the
usb_bulk_message() routine and USBDEVFS_BULK by using them for interrupt
transfers as well as bulk transfers.

This patch (as567) formalizes this practice and adds code to
usb_bulk_message() for detecting when the target is really an interrupt
endpoint.  If it is, the routine submits an interrupt URB (using the
default interval) instead of a bulk URB.  In theory this should help HCDs
that don't like it when people try to mix transfer types, queuing both
periodic and non-periodic types for the same endpoint.

Not fully tested -- I don't have any programs that use USBDEVFS_BULK for
interrupt transfers -- but it compiles okay and normal bulk messages work
as well as before.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/message.c |   24 ++++++++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

19 years ago[PATCH] updates for "controller suspended" handling
David Brownell [Tue, 27 Sep 2005 17:38:54 +0000 (10:38 -0700)]
[PATCH] updates for "controller suspended" handling

Reject URBs to _all_ devices when their host controllers are suspended;
even root hub registers will be unavailable.  Also, don't reject urbs
to root hubs in other cases; the only upstream link is through that
controller (on PCI or whatever SOC bus is in use).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hcd.c |   28 ++++++++++++----------------
 drivers/usb/core/urb.c |    3 ++-
 2 files changed, 14 insertions(+), 17 deletions(-)

19 years ago[PATCH] stop exporting two functions
David Brownell [Fri, 23 Sep 2005 05:45:26 +0000 (22:45 -0700)]
[PATCH] stop exporting two functions

The way we're looking at USB suspend lately doesn't expect drivers to
call usb_suspend_device() or usb_resume_device() directly; that'll
be implicit when no interfaces are in use.

This patch removes those APIs from visibility outside usbcore.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
 drivers/usb/core/hub.c |   12 ++++--------
 drivers/usb/core/usb.h |    4 ++++
 include/linux/usb.h    |    5 -----
 3 files changed, 8 insertions(+), 13 deletions(-)

19 years ago[PATCH] ISP116x PM updates
David Brownell [Fri, 23 Sep 2005 05:45:13 +0000 (22:45 -0700)]
[PATCH] ISP116x PM updates

This makes the isp116x driver stop using usb_suspend_device() and
usb_resume_device() ... usbcore now calls to the root hub methods,
removing the need for this.  It also switches from keventd to khubd
for remote wakeup.  (Compile tested.)

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/isp116x-hcd.c |   29 ++++-------------------------
 drivers/usb/host/isp116x.h     |    1 -
 2 files changed, 4 insertions(+), 26 deletions(-)

19 years ago[PATCH] UHCI PM updates
David Brownell [Fri, 23 Sep 2005 05:44:56 +0000 (22:44 -0700)]
[PATCH] UHCI PM updates

This removes suspend and resume path recursion in UHCI.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/uhci-hcd.c |    9 ---------
 1 file changed, 9 deletions(-)

19 years ago[PATCH] update PCI early-handoff handling for OHCI
David Brownell [Fri, 23 Sep 2005 05:43:30 +0000 (22:43 -0700)]
[PATCH] update PCI early-handoff handling for OHCI

The PCI "early usb handoff" quirk logic didn't work like "ohci-hcd" ...
This patch makes it do so by:

  - Resetting the controller after kicking BIOS off, matching the
    normal "chip in hardware reset" startup mode;

  - Reporting any BIOS that borks this simple handoff; it's likely
    got a few other surprises for us too.

  - Ignoring that handoff on HPPA;

The diagnostic string is mostly shared with EHCI, saving a few bytes.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/pci-quirks.c |   22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)

19 years ago[PATCH] OHCI PM updates
David Brownell [Fri, 23 Sep 2005 05:42:53 +0000 (22:42 -0700)]
[PATCH] OHCI PM updates

This simplifies the OHCI root hub suspend logic:

 - Uses new usbcore root hub calls to make autosuspend work again:
* Uses a newish usbcore root hub wakeup mechanism,
  making requests to khubd not keventd.
* Uses an even newer sibling suspend hook.

 - Expect someone always made usbcore call ohci_hub_suspend() before bus
   glue fires; and that ohci_hub_resume() is only called after that bus
   glue ran.  Previously, only CONFIG_USB_SUSPEND promised those things.
   (Includes updates to PCI and OMAP bus glue.)

 - Handle a not-noticed-before special case during resume from one of
   the swsusp snapshots when using "usb-handoff":  the controller isn't
   left in RESET state.  (A bug to fix in the usb-handoff code...)

Also cleans up a minor debug printk glitch, and switches an mdelay over
to an msleep (how did that stick around for so long?).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/ohci-dbg.c  |    4 ----
 drivers/usb/host/ohci-hcd.c  |    2 +-
 drivers/usb/host/ohci-hub.c  |   42 ++++++++++++------------------------------
 drivers/usb/host/ohci-mem.c  |    1 -
 drivers/usb/host/ohci-omap.c |   36 ++++++++++++------------------------
 drivers/usb/host/ohci-pci.c  |   40 ++++++++--------------------------------
 drivers/usb/host/ohci.h      |    1 -
 7 files changed, 33 insertions(+), 93 deletions(-)

19 years ago[PATCH] usbcore PCI glue updates for PM
David Brownell [Fri, 23 Sep 2005 05:38:16 +0000 (22:38 -0700)]
[PATCH] usbcore PCI glue updates for PM

This updates the PCI glue to address the new and simplified usbcore suspend
semantics, where CONFIG_USB_SUSPEND becomes irrelevant to HCDs because
hcd->hub_suspend() will always be called.

  - Removes now-unneeded recursion support

  - Go back to ignoring faults reported by the wakeup calls; we expect them
    to fail sometimes, and that's just fine.

The PCI HCDs will need simple changes to catch up to this, like being able
to ignore the setting of CONFIG_USB_SUSPEND.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hcd-pci.c |  106 +++++++++++++++++++++------------------------
 drivers/usb/core/hcd.h     |    6 +-
 2 files changed, 53 insertions(+), 59 deletions(-)

19 years ago[PATCH] root hub updates (greater half)
David Brownell [Fri, 23 Sep 2005 05:37:29 +0000 (22:37 -0700)]
[PATCH] root hub updates (greater half)

This patch associates hub suspend and resume logic (including for root hubs)
with CONFIG_PM -- instead of CONFIG_USB_SUSPEND as before -- thereby unifying
two troublesome versions of suspend logic into just one.  It'll be easier to
keep things right from now on.

  - Now usbcore _always_ calls hcd->hub_suspend as needed, instead of
    only when USB_SUSPEND is enabled:
     * Those root hub methods are now called from hub suspend/resume;
       no more skipping between layers during device suspend/resume;
     * It now handles cases allowed by sysfs or autosuspended root hubs,
       by forcing the hub interface to resume too.

  - All devices, including virtual root hubs, now get the same treatment
    on their resume paths ... including re-activating all their interfaces.

Plus it gets rid of those stub copies of usb_{suspend,resume}_device(), and
updates the Kconfig to match the new definition of USB_SUSPEND:  it provides
(a) selective suspend, downstream from hubs; and (b) remote wakeup, upstream
from any device configuration which supports it.

This calls for minor followup patches for most HCDs (and their PCI glue).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/Kconfig |   11 ++-
 drivers/usb/core/hub.c   |  163 +++++++++++++++++++++++++----------------------
 2 files changed, 97 insertions(+), 77 deletions(-)

19 years ago[PATCH] root hub changes (lesser half)
David Brownell [Fri, 23 Sep 2005 05:32:24 +0000 (22:32 -0700)]
[PATCH] root hub changes (lesser half)

This patch collects various small updates related to root hubs, to shrink
later patches which build on them.

  - For root hub suspend/resume support:
     * Make the existing usb_hcd_resume_root_hub() routine respect pmcore
       locking, exporting and using the dpm_runtime_resume() method.
     * Add a new usb_hcd_suspend_root_hub() to pair with that routine.
       (Essential to make OHCI autosuspend behave again...)
     * HC_SUSPENDED by itself only refers to the root hub's downstream ports.
       So let HCDs see root hub URBs unless the parent device is suspended.

  - Remove an assertion we no longer need (and now, also don't want).

  - Generic suspend/resume updates to work better with swsusp.
     * Ignore the FREEZE vs SUSPEND distinction for hardware; trying to
       use it breaks the swsusp snapshots it's supposed to help (sigh).
     * On resume, mark devices as resumed right away, but then
       do nothing else if the device is marked NOTATTACHED.

These changes shouldn't be very noticable by themselves.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/base/power/runtime.c |    1
 drivers/usb/core/hcd.c       |   64 ++++++++++++++++++++++++++++++++++++++-----
 drivers/usb/core/hcd.h       |    1
 drivers/usb/core/hub.c       |   45 ++++++++++++++++++++++++------
 drivers/usb/core/usb.c       |   20 +++++++++----
 drivers/usb/core/usb.h       |    1
 6 files changed, 111 insertions(+), 21 deletions(-)

19 years ago[PATCH] all HCDs provide root hub suspend/resume methods
David Brownell [Fri, 23 Sep 2005 05:32:11 +0000 (22:32 -0700)]
[PATCH] all HCDs provide root hub suspend/resume methods

This cleans up a small recent FIXME, ensuring that all the HCDs provide
root hub suspend/resume methods.  It also wraps the calls to those root
suspend routines just like on the PCI "USB_SUSPEND not defined" cases,
so non-PCI bus glue won't be as tempted to behave very differently.

Several of the SOC based OHCI drivers forgot to list those methods;
the patch also adds those missing declarations.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hcd.c          |   42 +++++++++++++++++++++++++---------------
 drivers/usb/host/ohci-au1xxx.c  |    5 ++++
 drivers/usb/host/ohci-lh7a404.c |    5 ++++
 drivers/usb/host/ohci-pxa27x.c  |    1
 drivers/usb/host/ohci-s3c2410.c |    1
 drivers/usb/host/ohci-sa1111.c  |    1
 6 files changed, 40 insertions(+), 15 deletions(-)

19 years ago[PATCH] EHCI, split out PCI glue
Matt Porter [Fri, 23 Sep 2005 05:31:15 +0000 (22:31 -0700)]
[PATCH] EHCI, split out PCI glue

This splits BIOS and PCI specific support out of ehci-hcd.c into
ehci-pci.c.  It follows the model already used in the OHCI driver
so support for non-PCI EHCI controllers can be more easily added.

Signed-off-by: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/ehci-hcd.c |  543 ++++++--------------------------------------
 drivers/usb/host/ehci-pci.c |  414 +++++++++++++++++++++++++++++++++
 drivers/usb/host/ehci.h     |    1
 3 files changed, 492 insertions(+), 466 deletions(-)

19 years ago[PATCH] one less word in struct device
David Brownell [Fri, 23 Sep 2005 05:30:48 +0000 (22:30 -0700)]
[PATCH] one less word in struct device

This saves a word from "struct device" ... there's a refcounting mechanism
stub that's rather ineffective (the values are never even tested!), which
can safely be deleted.  With this patch it uses normal device refcounting,
so any potential users of the pm_parent mechanism will be more correct.
(That mechanism is actually unusable for now though; it does nothing.)

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/base/power/main.c |   26 +++-----------------------
 include/linux/pm.h        |    1 -
 2 files changed, 3 insertions(+), 24 deletions(-)

19 years ago[PATCH] USB: UHCI: Split apart the physical and logical framelist arrays
Alan Stern [Fri, 16 Sep 2005 18:22:51 +0000 (14:22 -0400)]
[PATCH] USB: UHCI: Split apart the physical and logical framelist arrays

This patch (as563) splits the physical and logical framelist arrays in
uhci-hcd into two separate pieces.  This will allow slightly better memory
utilization, since each piece is no larger than a single page whereas
before the whole thing was a little bigger than two pages.  It also allows
the logical array to be allocated in non-DMA-coherent memory.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] USB: UHCI: Remove unused fields and unneeded tests for NULL
Alan Stern [Fri, 16 Sep 2005 18:17:45 +0000 (14:17 -0400)]
[PATCH] USB: UHCI: Remove unused fields and unneeded tests for NULL

This patch (as562) removes from the uhci-hcd driver a few unused fields
and some unnecessary tests against NULL and assignments to NULL.  In fact
it wasn't until fairly recently that the tests became unnecessary.
Before last winter it was possible that the driver's stop() routine would
get called even if the start() routine returned an error, but now that
can't happen.  Hence there's no longer any need to check for partial
initialization.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] USB: Fix usb hub build
Matt Porter [Mon, 19 Sep 2005 16:40:24 +0000 (09:40 -0700)]
[PATCH] USB: Fix usb hub build

Signed-off-by: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hub.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

19 years ago[PATCH] remove some USB_SUSPEND dependencies
David Brownell [Wed, 14 Sep 2005 02:59:11 +0000 (19:59 -0700)]
[PATCH] remove some USB_SUSPEND dependencies

This simplifies some of the PM-related #ifdeffing by recognizing
that USB_SUSPEND depends on PM.  Also, OHCI drivers were often
testing for USB_SUSPEND when they should have tested just PM.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hcd.c          |    2 ++
 drivers/usb/host/ohci-hcd.c     |    2 +-
 drivers/usb/host/ohci-hub.c     |    4 ++--
 drivers/usb/host/ohci-omap.c    |    2 +-
 drivers/usb/host/ohci-pci.c     |    2 +-
 drivers/usb/host/ohci-ppc-soc.c |    4 ++--
 drivers/usb/host/ohci-pxa27x.c  |    2 +-
 drivers/usb/host/ohci-s3c2410.c |    3 +--
 drivers/usb/host/ohci-sa1111.c  |    2 +-
 9 files changed, 12 insertions(+), 11 deletions(-)

19 years ago[PATCH] remove duplicated resume path code
David Brownell [Wed, 14 Sep 2005 02:57:36 +0000 (19:57 -0700)]
[PATCH] remove duplicated resume path code

This gets rid of some inconsistently duplicated logic to resume interfaces.
Similar code was in both finish_port_resume() and in usb_generic_resume().
Now there is just one copy of that code, accessed regardless of whether
CONFIG_USB_SUSPEND is enabled.  Fault handling is also more consistent.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] remove usb_suspend_device() parameter
David Brownell [Wed, 14 Sep 2005 02:57:27 +0000 (19:57 -0700)]
[PATCH] remove usb_suspend_device() parameter

This patch removes the extra usb_suspend_device() parameter.  The original
reason to pass that parameter was so that this routine could suspend any
active children.  A previous patch removed that functionality ... leaving
no reason to pass the parameter.  A close analogy is pci_set_power_state,
which doesn't need a pm_message_t either.

On the internal code path that comes through the driver model, the parameter
is now used to distinguish cases where USB devices need to "freeze" but not
suspend.   It also checks for an error case that's accessible through sysfs:
attempting to suspend a device before its interfaces (or for hubs, ports).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hub.c         |   34 +++++++++++++++++++++-------------
 drivers/usb/core/usb.c         |   23 +++++++++++++++++++++--
 drivers/usb/host/ehci-hcd.c    |    2 +-
 drivers/usb/host/isp116x-hcd.c |    2 +-
 drivers/usb/host/ohci-pci.c    |    2 +-
 include/linux/usb.h            |    2 +-
 6 files changed, 46 insertions(+), 19 deletions(-)

19 years ago[PATCH] remove suspend-path recursion
David Brownell [Wed, 14 Sep 2005 02:57:04 +0000 (19:57 -0700)]
[PATCH] remove suspend-path recursion

This patch removes some recursion in the CONFIG_USB_SUSPEND logic, which
suspended children (of devices or hubs) that weren't already suspended.
When it sees such cases, suspend now just fails cleanly.

That logic was not needed during system-wide sleep state transitions; and
given the current notions of how to manage selective suspend transitions,
we don't want it there either.  Where it was particularly handy was coping
with various limitations of the sysfs "echo -n N > power/state" support.
(These include assuming that "N" is always meaningful to the driver; and
that drivers can only transition to state N from state zero.)

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] usb_interface power state
David Brownell [Wed, 14 Sep 2005 02:56:33 +0000 (19:56 -0700)]
[PATCH] usb_interface power state

This updates the handling of power state for USB interfaces.

  - Formalizes an existing invariant:  interface "power state" is a boolean:
    ON when I/O is allowed, and FREEZE otherwise.  It does so by defining
    some inlined helpers, then using them.

  - Adds a useful invariant:  the only interfaces marked active are those
    bound to non-suspended drivers.  Later patches build on this invariant.

  - Simplifies the interface driver API (and removes some error paths) by
    removing the requirement that they record power state changes during
    suspend and resume callbacks.  Now usbcore does that.

A few drivers were simplified to address that last change.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/hub.c       |   33 +++++++++------------
 drivers/usb/core/message.c   |    1
 drivers/usb/core/usb.c       |   65 +++++++++++++++++++++++++++++++++----------
 drivers/usb/core/usb.h       |   18 +++++++++++
 drivers/usb/input/hid-core.c |    2 -
 drivers/usb/misc/usbtest.c   |   10 ------
 drivers/usb/net/pegasus.c    |    2 -
 drivers/usb/net/usbnet.c     |    2 -
 8 files changed, 85 insertions(+), 48 deletions(-)

19 years ago[PATCH] USB: move handoff code
David Brownell [Sat, 24 Sep 2005 00:14:37 +0000 (17:14 -0700)]
[PATCH] USB: move handoff code

This moves the PCI quirk handling for USB host controllers from the
PCI directory to the USB directory.  Follow-on patches will need to:

(a) merge these copies with the originals in the HCD reset methods.
they don't wholly agree, despite doing the very same thing; and

(b) eventually change it so "usb-handoff" is the default, to help
get more robust USB/BIOS/input/... interactions.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/Makefile              |    2
 drivers/pci/quirks.c          |  253 ---------------------------------------
 drivers/usb/Makefile          |    1
 drivers/usb/host/Makefile     |    5
 drivers/usb/host/pci-quirks.c |  272 ++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 280 insertions(+), 253 deletions(-)

19 years ago[PATCH] ub: suppress gcc warnings for pointer casts
Pete Zaitcev [Sat, 24 Sep 2005 20:10:16 +0000 (13:10 -0700)]
[PATCH] ub: suppress gcc warnings for pointer casts

When building on a 64-bit platform, gcc produces a warning
"cast of a pointer to an integer of a different size".
The scatterlist.offset on the LHS is unsigned int, so I used
that originally.

Signed-off-by: Pete Zaitcev <zaitcev@yahoo.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/block/ub.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

19 years ago[PATCH] USB: add more snooping hooks in devio.c
Greg Kroah-Hartman [Tue, 21 Jun 2005 04:15:16 +0000 (21:15 -0700)]
[PATCH] USB: add more snooping hooks in devio.c

Now we can log the urbs travelling through usbfs

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] devfs: Remove the mode field from usb_class_driver as it's no longer needed
Greg Kroah-Hartman [Tue, 21 Jun 2005 04:15:16 +0000 (21:15 -0700)]
[PATCH] devfs: Remove the mode field from usb_class_driver as it's no longer needed

Also fixes all drivers that set this field, and removes some other devfs
specfic USB logic.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/class/usblp.c           |    3 +--
 drivers/usb/core/file.c             |   19 ++++---------------
 drivers/usb/image/mdc800.c          |    3 +--
 drivers/usb/input/aiptek.c          |    2 +-
 drivers/usb/input/hiddev.c          |    3 +--
 drivers/usb/media/dabusb.c          |    3 +--
 drivers/usb/misc/auerswald.c        |    3 +--
 drivers/usb/misc/idmouse.c          |    5 ++---
 drivers/usb/misc/legousbtower.c     |    5 ++---
 drivers/usb/misc/rio500.c           |    3 +--
 drivers/usb/misc/sisusbvga/sisusb.c |    5 -----
 drivers/usb/misc/usblcd.c           |    9 ++++-----
 drivers/usb/usb-skeleton.c          |    3 +--
 include/linux/usb.h                 |    7 ++-----
 14 files changed, 22 insertions(+), 51 deletions(-)

19 years ago[PATCH] USB: add endpoint information to sysfs
Greg Kroah-Hartman [Tue, 21 Jun 2005 04:15:16 +0000 (21:15 -0700)]
[PATCH] USB: add endpoint information to sysfs

This patch adds endpoint information for both devices and interfaces to
sysfs.  Previously it was only possible to get the endpoint information
from usbfs, and never possible to get any information on endpoint 0.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/core/sysfs.c |  195 ++++++++++++++++++++++++++++++++++++++++++++++-
 include/linux/usb.h      |    4
 2 files changed, 197 insertions(+), 2 deletions(-)

19 years ago[PATCH] USB: disable tasklet if rtl8150 device is removed while active.
Greg Kroah-Hartman [Tue, 21 Jun 2005 04:15:16 +0000 (21:15 -0700)]
[PATCH] USB: disable tasklet if rtl8150 device is removed while active.

This fixes a oops seen in the wild.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] USB: ftdi: Artemis and ATIK based USB astronomical CCD cameras
Rui Santos [Fri, 23 Sep 2005 19:06:50 +0000 (20:06 +0100)]
[PATCH] USB: ftdi: Artemis and ATIK based USB astronomical CCD cameras

This patch enables direct kernel support for the Artemis
and ATIK astronomical based USB CCD cameras.
Since all communications with this camera are done via an
FTDI 245BM chip, it was only needed to specify the
ProductID and VendorID of all three devices.
In what tests are concerned, data was transfered from and
to the FTDI at the chips Top speed (360KB/s).

Signed-off-by: Rui Santos <rsantos@grupopie.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/serial/ftdi_sio.c |    3 +++
 drivers/usb/serial/ftdi_sio.h |   13 +++++++++++++
 2 files changed, 16 insertions(+)

19 years ago[PATCH] USB: ehci.patch (earlier irq disable)
David Brownell [Tue, 27 Sep 2005 17:19:39 +0000 (10:19 -0700)]
[PATCH] USB: ehci.patch (earlier irq disable)

This tweaks the EHCI reboot notifier to also halt the EHCI controller, and
makes that halt code force IRQs off.  Both should always have been done.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/usb/host/ehci-hcd.c |    8 ++++++++
 1 file changed, 8 insertions(+)

19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Fri, 28 Oct 2005 22:57:32 +0000 (15:57 -0700)]
Merge /pub/scm/linux/kernel/git/gregkh/pci-2.6

19 years agoMerge master.kernel.org:/home/rmk/linux-2.6-mmc
Linus Torvalds [Fri, 28 Oct 2005 22:54:06 +0000 (15:54 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-mmc

19 years ago[PATCH] PCI fixup for Toshiba laptops and ohci1394
Jesse Barnes [Tue, 25 Oct 2005 17:28:42 +0000 (10:28 -0700)]
[PATCH] PCI fixup for Toshiba laptops and ohci1394

This is a fix for a bug I see on my Toshiba laptop, where the ohci1394
controller gets initialized improperly.  The patch adds two PCI fixups
to arch/i386/pci/fixup.c, one that happens early on to cache the value
of the PCI_CACHE_LINE_SIZE config register, and another that later
restores the value, along with a valid IRQ number and some BAR values.
I've tested it on my laptop, and it prevents me from running into what I
consider to be a major bug: IRQ 11 is disabled by the IRQ debug code,
causing my wireless to break.

Thanks to Rob for the original patch to ohci1394.c and Stefan for lots
of proofreading (and a last minute bug caught in review!) and additional
information collection.  I think the DMI system list is correct, but we
may need to add some more PCI IDs to the PCI_FIXUP macros over time.

Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] pci_ids: cleanup comments
Grant Coady [Wed, 19 Oct 2005 23:55:32 +0000 (09:55 +1000)]
[PATCH] pci_ids: cleanup comments

pci_ids.h cleanup: convert // comment to /* comment */

Signed-off-by: Grant Coady <gcoady@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] kernel-doc: PCI fixes
Randy Dunlap [Sun, 23 Oct 2005 18:57:38 +0000 (11:57 -0700)]
[PATCH] kernel-doc: PCI fixes

PCI: add descriptions for missing function parameters.
Eliminate all kernel-doc warnings here.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] kernel-doc: fix PCI hotplug
Randy Dunlap [Sun, 23 Oct 2005 18:58:19 +0000 (11:58 -0700)]
[PATCH] kernel-doc: fix PCI hotplug

PCI hotplug.c: does not contain kernel-doc, so don't process it for now.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] acpiphp: allocate resources for adapters with bridges
Kristen Accardi [Wed, 19 Oct 2005 00:21:40 +0000 (17:21 -0700)]
[PATCH] acpiphp: allocate resources for adapters with bridges

Allocate resources for adapters with bridges on them.

Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] PCI: Convert megaraid to use pci_driver shutdown method
Russell King [Sun, 16 Oct 2005 20:32:46 +0000 (21:32 +0100)]
[PATCH] PCI: Convert megaraid to use pci_driver shutdown method

Convert megaraid to use pci_driver's shutdown method rather than
the generic device_driver shutdown method.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] PCI: Fixup PCI driver shutdown
Russell King [Sun, 16 Oct 2005 20:31:36 +0000 (21:31 +0100)]
[PATCH] PCI: Fixup PCI driver shutdown

Add a warning to pci driver registration code so that we know
whether we have drivers using the obsolete driver shutdown
method.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] PCI: fix edac drivers for radisys 82600 borkage
Andrew Morton [Sun, 16 Oct 2005 05:02:19 +0000 (22:02 -0700)]
[PATCH] PCI: fix edac drivers for radisys 82600 borkage

I told you that the pci_ids.h cleanup was a bad idea ;)

Cc: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: fix oops at driver unload
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:44 +0000 (12:05 -0700)]
[PATCH] shpchp: fix oops at driver unload

At unload time, the shpchp driver does not remove sysfs files
it had created in the driver's probe entry point. This patch
fixes this problem.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: reduce debug message verbosity
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:43 +0000 (12:05 -0700)]
[PATCH] shpchp: reduce debug message verbosity

Reduce the number of debug messages generated if shpchp debug is
enabled. I tried to restrict this to removing debug messages that
are either early-driver-debug type messages, or print information
that can be inferred through other debug prints.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: miscellaneous cleanups
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:42 +0000 (12:05 -0700)]
[PATCH] shpchp: miscellaneous cleanups

Remove un-necessary header includes, remove dead code, remove
some type casts, receive function return in the correct data
type...

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: remove redundant data structures
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:41 +0000 (12:05 -0700)]
[PATCH] shpchp: remove redundant data structures

State information is currently stored in per-slot as well as
per-pci-function data structures in shpchp. There's a lot of
overlap in the information kept, and some of it is never used.
This patch consolidates the state information to per-slot and
eliminates unused data structures. The biggest change is to
eliminate the pci_func structure and the code around managing
its lists.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: dont save PCI config for hotplug slots/devices
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:40 +0000 (12:05 -0700)]
[PATCH] shpchp: dont save PCI config for hotplug slots/devices

This patch eliminates saving the PCI config header for devices
in hotplug capable slots. We now use the PCI core to get the
specific parts of the config header as required.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: detect SHPC capability before doing a lot of work
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:39 +0000 (12:05 -0700)]
[PATCH] shpchp: detect SHPC capability before doing a lot of work

The shpc driver registers its probe function for all pci-pci
bridges in the system. Not all of them will be shpc capable, so
look for this capability early in the probe function and return
if there's no work to do on this bridge. The old shpc driver
did some initialization work on all bridges before detecting
that shpc is not supported and unwinds the work it's already done
in that case.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: reduce dependence on ACPI
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:38 +0000 (12:05 -0700)]
[PATCH] shpchp: reduce dependence on ACPI

Reduce the SHPC hotplug driver's dependence on ACPI. We don't
walk the acpi namespace anymore to build a list of bridges and
devices. The remaining interaction with ACPI is to run the
_OSHP method to transition control of hotplug hardware from
system BIOS to the shpc hotplug driver, and to run the _HPP
method to get hotplug device parameters like cache line size,
latency timer and SERR/PERR enable from BIOS.

Note that one of the side effects of this patch is that shpchp
does not enable the hot-added device or its DMA bus mastering
automatically now. It expects the device driver to do that.
This may break some drivers and we will have to fix them as
they are reported.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: remove redundant display of PCI device resources
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:37 +0000 (12:05 -0700)]
[PATCH] shpchp: remove redundant display of PCI device resources

Info about resources assigned to PCI devices is already available
through sysfs and pci utilities. There's no need for shpchp to
create another sysfs file to display the same information.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] shpchp: use the PCI core for hotplug resource management
rajesh.shah@intel.com [Thu, 13 Oct 2005 19:05:36 +0000 (12:05 -0700)]
[PATCH] shpchp: use the PCI core for hotplug resource management

This patch converts the standard hotplug controller driver to use
the PCI core for resource management. This eliminates a whole lot
of duplicated code, and integrates shpchp in the system's normal
PCI handling code.

Signed-off-by: Rajesh Shah <rajesh.shah@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] PCI: Add quirk for SMBus on HP D530
Jean Delvare [Wed, 21 Sep 2005 20:26:31 +0000 (22:26 +0200)]
[PATCH] PCI: Add quirk for SMBus on HP D530

One more SMBus unhiding quirk, this time for the HP D530. Requested and
successfully tested by Ben Cranston.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] ppc64 PCI Hotplug: cleanup unsymmetric API routines
linas [Thu, 29 Sep 2005 00:33:38 +0000 (19:33 -0500)]
[PATCH] ppc64 PCI Hotplug: cleanup unsymmetric API routines

This is a minor patch to the ppc64 PCI hotplug code; it makes the call to
rpaphp_unconfig_pci_adapter() symmetric with respect to the call to
rpaphp_config_pci_adapter().  I discussed this with John Rose, who
had provided the last round of changes for these functions; he
appearently had this patch but somehow failed to mail it out.

Tested. (added/removed device).

Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
 drivers/pci/hotplug/rpaphp.h      |    3 ++-
 drivers/pci/hotplug/rpaphp_core.c |    5 ++++-
 drivers/pci/hotplug/rpaphp_pci.c  |   11 +++--------
 3 files changed, 9 insertions(+), 10 deletions(-)

19 years ago[PATCH] pci_ids: remove non-referenced symbols from pci_ids.h
Grant Coady [Thu, 29 Sep 2005 01:06:40 +0000 (11:06 +1000)]
[PATCH] pci_ids: remove non-referenced symbols from pci_ids.h

pci_ids.h cleanup: removed non-referenced symbols, compile tested
with 'make allmodconfig'

Signed-off-by: Grant Coady <gcoady@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 include/linux/pci_ids.h |  540 ------------------------------------------------
 1 file changed, 540 deletions(-)

19 years ago[PATCH] pci_ids: macros: replace partial with whole symbols
Grant Coady [Thu, 29 Sep 2005 00:40:52 +0000 (10:40 +1000)]
[PATCH] pci_ids: macros: replace partial with whole symbols

pci_ids cleanup: replace symbols built by macros with whole symbols to
aid grep searches.

Signed-off-by: Grant Coady <gcoady@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/video/cirrusfb.c |   24 ++++++++++++------------
 sound/oss/ymfpci.c       |   17 +++++++++--------
 sound/pci/bt87x.c        |   11 +++++++----
 3 files changed, 28 insertions(+), 24 deletions(-)

19 years ago[PATCH] pci_ids: remove duplicates from pci_ids.h
Grant Coady [Thu, 29 Sep 2005 00:39:46 +0000 (10:39 +1000)]
[PATCH] pci_ids: remove duplicates from pci_ids.h

pci_ids.h cleanup: remove duplicated entries and change some defines to
explicit value rather than in terms of another constant, preparation for
removing unused symbols

Signed-off-by: Grant Coady <gcoady@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 include/linux/pci_ids.h |   28 +++++++++-------------------
 1 file changed, 9 insertions(+), 19 deletions(-)

19 years ago[PATCH] unhide ICH6 SMBus - take 2
R.Marek@sh.cvut.cz [Thu, 29 Sep 2005 08:35:41 +0000 (08:35 +0000)]
[PATCH] unhide ICH6 SMBus - take 2

This patch unhides hidden SMBus on ICH6 chipset installed in
Asus M6V notebook. I would like to thank Michal Mleczko for
testing and help.

Signed-Off-By: Rudolf Marek <r.marek@sh.cvut.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/pci/quirks.c |   24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

19 years ago[PATCH] cpqphp: add pci_enable_device()
Bjorn Helgaas [Wed, 28 Sep 2005 20:03:08 +0000 (14:03 -0600)]
[PATCH] cpqphp: add pci_enable_device()

Add pci_{enable,disable}_device() calls.  Without pci_enable_device(),
dev->irq is garbage, and cpqphp relies on it.

This fixes a problem reported by Bruno Redondi.  He reported a flood
of ACPI interrupts, that caused kacpid to run 100% of the time:
    http://bugzilla.kernel.org/show_bug.cgi?id=5312

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/pci/hotplug/cpqphp_core.c |   24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

19 years ago[PATCH] cpcihp_zt5550: add pci_enable_device()
Bjorn Helgaas [Wed, 28 Sep 2005 21:15:16 +0000 (15:15 -0600)]
[PATCH] cpcihp_zt5550: add pci_enable_device()

Add pci_{enable,disable}_device() calls.  Without pci_enable_device(),
dev->irq is garbage, and cpcihp_zt5550 relies on it.

Compiled but untested, since I don't have the hardware.

Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Scott Murray <scottm@somanetworks.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/pci/hotplug/cpcihp_zt5550.c |   25 +++++++++++++++++++++----
 1 file changed, 21 insertions(+), 4 deletions(-)

19 years ago[PATCH] pci: cleanup need_restore switch statement
John W. Linville [Wed, 28 Sep 2005 21:50:51 +0000 (17:50 -0400)]
[PATCH] pci: cleanup need_restore switch statement

Cleanup the need_restore switch statement in
pci_set_power_state(). This makes it more safe by explicitly handling
all the PCI power states instead of handling them as the default
case. It also reads a little better IMHO.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] PCI: ipr: Block config access during BIST
Brian King [Tue, 27 Sep 2005 08:21:56 +0000 (01:21 -0700)]
[PATCH] PCI: ipr: Block config access during BIST

IPR scsi adapter have an exposure today in that they issue BIST to the adapter
to reset the card.  If, during the time it takes to complete BIST, userspace
attempts to access PCI config space, the host bus bridge will master abort the
access since the ipr adapter does not respond on the PCI bus for a brief
period of time when running BIST.  On PPC64 hardware, this master abort
results in the host PCI bridge isolating that PCI device from the rest of the
system, making the device unusable until Linux is rebooted.  This patch makes
use of some newly added PCI layer APIs that allow for protection from
userspace accessing config space of a device in scenarios such as this.

Signed-off-by: Brian King <brking@us.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/scsi/ipr.c |    2 ++
 1 file changed, 2 insertions(+)

19 years ago[PATCH] PCI: Block config access during BIST
Brian King [Tue, 27 Sep 2005 08:21:55 +0000 (01:21 -0700)]
[PATCH] PCI: Block config access during BIST

Some PCI adapters (eg.  ipr scsi adapters) have an exposure today in that they
issue BIST to the adapter to reset the card.  If, during the time it takes to
complete BIST, userspace attempts to access PCI config space, the host bus
bridge will master abort the access since the ipr adapter does not respond on
the PCI bus for a brief period of time when running BIST.  On PPC64 hardware,
this master abort results in the host PCI bridge isolating that PCI device
from the rest of the system, making the device unusable until Linux is
rebooted.  This patch is an attempt to close that exposure by introducing some
blocking code in the PCI code.  When blocked, writes will be humored and reads
will return the cached value.  Ben Herrenschmidt has also mentioned that he
plans to use this in PPC power management.

Signed-off-by: Brian King <brking@us.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/pci/access.c    |   89 ++++++++++++++++++++++++++++++++++++++++++++++++
 drivers/pci/pci-sysfs.c |   20 +++++-----
 drivers/pci/pci.h       |    7 +++
 drivers/pci/proc.c      |   28 +++++++--------
 drivers/pci/syscall.c   |   14 +++----
 include/linux/pci.h     |    7 +++
 6 files changed, 134 insertions(+), 31 deletions(-)

19 years ago[PATCH] PCI: ICH6 ACPI and GPIO quirk
R.Marek@sh.cvut.cz [Tue, 27 Sep 2005 21:54:51 +0000 (21:54 +0000)]
[PATCH] PCI: ICH6 ACPI and GPIO quirk

This patch just adds ACPI and GPIO regions to its LPC bridge, similar
way as ICH4 did. I would like to thank Michal Mleczko for testing.

Signed-off-by: Rudolf Marek <r.marek@sh.cvut.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
 drivers/pci/quirks.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/i2c-2.6
Linus Torvalds [Fri, 28 Oct 2005 21:14:57 +0000 (14:14 -0700)]
Merge /pub/scm/linux/kernel/git/gregkh/i2c-2.6

19 years ago[PATCH] I2C: add i2c module alias for i2c drivers to use
Greg Kroah-Hartman [Fri, 21 Oct 2005 22:23:27 +0000 (00:23 +0200)]
[PATCH] I2C: add i2c module alias for i2c drivers to use

This is the start of adding hotplug-like support for i2c devices.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] I2C: remove devfs support from i2c-dev driver
Greg Kroah-Hartman [Fri, 21 Oct 2005 22:23:27 +0000 (00:23 +0200)]
[PATCH] I2C: remove devfs support from i2c-dev driver

as devfs is gone, this isn't needed anymore.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: smsc47m1 documentation update
Jean Delvare [Wed, 26 Oct 2005 20:21:24 +0000 (22:21 +0200)]
[PATCH] hwmon: smsc47m1 documentation update

The SMSC LPC47M997 Super-I/O chip seems to be compatible with the
LPC47M192, so it is supported by the smsc47m1 driver.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: lm90 documentation update
Jean Delvare [Wed, 26 Oct 2005 20:20:21 +0000 (22:20 +0200)]
[PATCH] hwmon: lm90 documentation update

Update the I2C addresses for the ADM1032 and ADT7461 chips.
Also update the links to the Analog Devices web site.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: Add PEC support to the lm90 driver
Jean Delvare [Wed, 26 Oct 2005 19:39:40 +0000 (21:39 +0200)]
[PATCH] hwmon: Add PEC support to the lm90 driver

Add PEC support to the lm90 driver. Only the ADM1032 chip supports it,
and in a rather tricky way, which is why this patch comes with
documentation reinforcements. At least, this demonstrates that the new
PEC support logic in i2c-core can properly deal with chips with partial
PEC support.

As enabling PEC causes a significant performance drop, it can be
disabled through a sysfs file (unsurprisingly named "pec").

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: Separate the lm90 register read function
Jean Delvare [Wed, 26 Oct 2005 19:37:52 +0000 (21:37 +0200)]
[PATCH] hwmon: Separate the lm90 register read function

Preparatory patch to add PEC support to the lm90 driver. We need a
centralized function to read register values, where the PEC code will
be later inserted. A positive side effect is that read errors are now
handled properly.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: i2c-i801 PEC code cleanups
Jean Delvare [Wed, 26 Oct 2005 19:34:42 +0000 (21:34 +0200)]
[PATCH] i2c: i2c-i801 PEC code cleanups

The tests leading to the use of hardware PEC in the i2c-i801 driver
can be simplified.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: SMBus PEC support rewrite, 3 of 3
Jean Delvare [Wed, 26 Oct 2005 19:31:15 +0000 (21:31 +0200)]
[PATCH] i2c: SMBus PEC support rewrite, 3 of 3

The new SMBus PEC implementation doesn't support PEC emulation on
non-PEC non-I2C SMBus masters, so we can drop all related code.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: SMBus PEC support rewrite, 2 of 3
Jean Delvare [Wed, 26 Oct 2005 19:28:55 +0000 (21:28 +0200)]
[PATCH] i2c: SMBus PEC support rewrite, 2 of 3

This is my rewrite of the SMBus PEC support. The original
implementation was known to have bugs (credits go to Hideki Iwamoto
for reporting many of them recently), and was incomplete due to a
conceptual limitation.

The rewrite affects only software PEC. Hardware PEC needs very little
code and is mostly untouched.

Technically, both implementations differ in that the original one
was emulating PEC in software by modifying the contents of an
i2c_smbus_data union (changing the transaction to a different type),
while the new one works one level lower, on i2c_msg structures (working
on message contents). Due to the definition of the i2c_smbus_data union,
not all SMBus transactions could be handled (at least not without
changing the definition of this union, which would break user-space
compatibility), and those which could had to be implemented
individually. At the opposite, adding PEC to an i2c_msg structure
can be done on any SMBus transaction with common code.

Advantages of the new implementation:

* It's about twice as small (from ~136 lines before to ~70 now, only
  counting i2c-core, including blank and comment lines). The memory
  used by i2c-core is down by ~640 bytes (~3.5%).

* Easier to validate, less tricky code. The code being common to all
  transactions by design, the risk that a bug can stay uncovered is
  lower.

* All SMBus transactions have PEC support in I2C emulation mode
  (providing the non-PEC transaction is also implemented). Transactions
  which have no emulation code right now will get PEC support for free
  when they finally get implemented.

* Allows for code simplifications in header files and bus drivers
  (patch follows).

Drawbacks (I guess there had to be at least one):

* PEC emulation for non-PEC capable non-I2C SMBus masters was dropped.
  It was based on SMBus tricks and doesn't quite fit in the new design.
  I don't think it's really a problem, as the benefit was certainly
  not worth the additional complexity, but it's only fair that I at
  least mention it.

Lastly, let's note that the new implementation does slightly affect
compatibility (both in kernel and user-space), but doesn't actually
break it. Some defines will be dropped, but the code can always be
changed in a way that will work with both the old and the new
implementations. It shouldn't be a problem as there doesn't seem to be
many users of SMBus PEC to date anyway.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: SMBus PEC support rewrite, 1 of 3
Jean Delvare [Wed, 26 Oct 2005 19:25:04 +0000 (21:25 +0200)]
[PATCH] i2c: SMBus PEC support rewrite, 1 of 3

Discard I2C_FUNC_SMBUS_*_PEC defines. i2c clients are not supposed to
check for PEC support of i2c bus drivers on individual SMBus
transactions, and i2c bus drivers are not supposed to advertise them.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: i2c-i810 documentation update
Jean Delvare [Wed, 26 Oct 2005 19:21:50 +0000 (21:21 +0200)]
[PATCH] i2c: i2c-i810 documentation update

Update the documented list of devices supported by the i2c-i810
driver.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Drop unused parport i2c IDs
Jean Delvare [Wed, 26 Oct 2005 19:20:17 +0000 (21:20 +0200)]
[PATCH] i2c: Drop unused parport i2c IDs

Drop unused i2c-over-parallel-port i2c IDs:
* I2C_HW_B_LPC was never actually used as far as I could search.
* I2C_HW_B_ELV and I2C_HW_B_VELLE are no more used since the
  introduction of the unified i2c-parport driver in Linux 2.6.2.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: ID redefinition cleanups
Jean Delvare [Wed, 26 Oct 2005 19:14:16 +0000 (21:14 +0200)]
[PATCH] i2c: ID redefinition cleanups

Fix several redefinitions of i2c IDs. i2c IDs must not be defined
outside of i2c-id.h.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Henk Vergonet <henk@god.dyndns.org>
Acked-by: Mark McClelland <mark@alpha.dyndns.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Static function fixes, 4 of 4
Ben Dooks [Wed, 26 Oct 2005 19:09:41 +0000 (21:09 +0200)]
[PATCH] i2c: Static function fixes, 4 of 4

Fix functions declared static and then implemented
without the static in drivers/i2c/chips.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: Static function fixes, 3 of 4
Ben Dooks [Wed, 26 Oct 2005 19:07:25 +0000 (21:07 +0200)]
[PATCH] hwmon: Static function fixes, 3 of 4

Fixup functions that have been declared static
and then actually defined without the static on.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: Static function fixes, 2 of 4
Ben Dooks [Wed, 26 Oct 2005 19:05:46 +0000 (21:05 +0200)]
[PATCH] hwmon: Static function fixes, 2 of 4

lm78.c and lm85.c have a number of items declared static
then implemented without the static on them. The following
patch fixes these sparse errors.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Static function fixes, 1 of 4
Ben Dooks [Wed, 26 Oct 2005 19:04:12 +0000 (21:04 +0200)]
[PATCH] i2c: Static function fixes, 1 of 4

eeprom_detect is first declared static and then when
the function is actually implemented, there is no static.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Documentation update
Jean Delvare [Mon, 17 Oct 2005 21:16:25 +0000 (23:16 +0200)]
[PATCH] i2c: Documentation update

Update the i2c documentation: kzalloc should be used instead of
kmalloc.

I also fixed a couple other things nearby in writing-clients, as several
past changes had never been reported there.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: kzalloc cleanups, 2 of 2
Jean Delvare [Mon, 17 Oct 2005 21:14:31 +0000 (23:14 +0200)]
[PATCH] i2c: kzalloc cleanups, 2 of 2

Use kzalloc instead of kmalloc in the S4882 SMBus multiplexing driver.
I guess it's safer that way.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: kzalloc cleanups, 1 of 2
Jean Delvare [Mon, 17 Oct 2005 21:12:36 +0000 (23:12 +0200)]
[PATCH] i2c: kzalloc cleanups, 1 of 2

Drop useless casts on kzalloc returned values, as suggested by
Jiri Slaby.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: kzalloc conversion, other drivers
Deepak Saxena [Mon, 17 Oct 2005 21:09:43 +0000 (23:09 +0200)]
[PATCH] i2c: kzalloc conversion, other drivers

Use kzalloc instead of kmalloc+memset in all remaining i2c bus and
chip drivers.

Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: kzalloc conversion
Deepak Saxena [Mon, 17 Oct 2005 21:08:32 +0000 (23:08 +0200)]
[PATCH] hwmon: kzalloc conversion

Use kzalloc instead of kmalloc+memset in all hardware monitoring
drivers.

Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: kzalloc conversion, ixp bus drivers
Deepak Saxena [Mon, 17 Oct 2005 21:07:05 +0000 (23:07 +0200)]
[PATCH] i2c: kzalloc conversion, ixp bus drivers

Use kzalloc instead of kmalloc+memzero in the ixp2000 and ixp4xx
I2C bus drivers.

Signed-off-by: Deepak Saxena <dsaxena@mvista.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: New Xicor X1205 RTC driver
Alessandro Zummo [Mon, 17 Oct 2005 21:04:42 +0000 (23:04 +0200)]
[PATCH] i2c: New Xicor X1205 RTC driver

New driver for the Xicor X1205 RTC chip.

Signed-off-by: Alessandro Zummo <alessandro.zummo@towertech.it>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: Missing class check in two hwmon drivers
Jean Delvare [Mon, 17 Oct 2005 21:02:42 +0000 (23:02 +0200)]
[PATCH] hwmon: Missing class check in two hwmon drivers

The atxp1 and ds1621 drivers should make sure they do not probe
non-hwmon i2c adapters.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] hwmon: New device ID for the smsc47b397 driver
Mark M. Hoffman [Mon, 17 Oct 2005 21:01:45 +0000 (23:01 +0200)]
[PATCH] hwmon: New device ID for the smsc47b397 driver

This patch adds a new ID to the SMSC LPC47B397-NC hardware
monitoring driver - for a chip that is claimed to be 100%
compatible otherwise.

Signed-off-by: Bryan Young (Utilitek Systems, Inc.)
Signed-off-by: Mark M. Hoffman <mhoffman@lightlink.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] Owner field additions to many i2c drivers, 5 of 5
Laurent Riffard [Mon, 17 Oct 2005 20:58:35 +0000 (22:58 +0200)]
[PATCH] Owner field additions to many i2c drivers, 5 of 5

In function i2c_isa_add_driver, copied driver should inherit the owner
field as well as the name field.

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] Owner field additions to many i2c drivers, 4 of 5
Laurent Riffard [Mon, 17 Oct 2005 20:54:45 +0000 (22:54 +0200)]
[PATCH] Owner field additions to many i2c drivers, 4 of 5

This patch updates the .owner field for the i2c core struct xxxx_driver
variables.

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] Owner field additions to many i2c drivers, 3 of 5
Laurent Riffard [Mon, 17 Oct 2005 20:53:52 +0000 (22:53 +0200)]
[PATCH] Owner field additions to many i2c drivers, 3 of 5

This patch updates the .owner field for various struct xxxx_driver
variables which are available on PPC_MAC arch.

This one was _not_ even compile-tested...

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] Owner field additions to many i2c drivers, 2 of 5
Laurent Riffard [Mon, 17 Oct 2005 20:51:37 +0000 (22:51 +0200)]
[PATCH] Owner field additions to many i2c drivers, 2 of 5

This patch updates the .owner field for various struct xxxx_driver variables,
other than pci_driver.

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Signed-off-by: Jean Delvare <khali@linux-fr>
Acked-by: Mark A. Greer <mgreer@mvista.com>
Acked-by: Ben Dooks <ben-linux@fluff.org>
Acked-by: Deepak Saxena <dsaxena@mvista.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] Owner field additions to many i2c drivers, 1 of 5
Laurent Riffard [Mon, 17 Oct 2005 20:47:48 +0000 (22:47 +0200)]
[PATCH] Owner field additions to many i2c drivers, 1 of 5

This patch updates .owner field for various struct pci_driver variables.

Signed-off-by: Laurent Riffard <laurent.riffard@free.fr>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Big i2c-elektor cleanup
Stig Telfer [Fri, 7 Oct 2005 22:23:27 +0000 (00:23 +0200)]
[PATCH] i2c: Big i2c-elektor cleanup

Cleanups to the i2c-elektor driver:

* Set the i2c_adapter name field to "i2c-elektor" and use this string
  in all resource requests and printks.
* Change space-padding for tab indentation, kill trailing white space,
  remove space before comma.
* Use dev_info, pr_info and pr_debug instead of printk.
* Lines chopped to 80 columns.

Signed-off-by: Stig Telfer <stig@lizardlogic.co.uk>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
19 years ago[PATCH] i2c: Fix i2c-elektor on Alpha
Stig Telfer [Fri, 7 Oct 2005 22:21:48 +0000 (00:21 +0200)]
[PATCH] i2c: Fix i2c-elektor on Alpha

This patch updates the i2c-elektor driver, enabling it to compile
cleanly, load and run.  The key change is that it uses the new
__iomem/iowrite8/ioread8 functions to abstract the direct or
memory-mapped variants of register access.  Also, the original driver
would crash on module load on the Alpha because the PCI memory region
was not remapped into kernel memory.

I have managed the following testing:

* compiled and tested it on my Alpha UP2000+ system.
* compiles cleanly for x86 but I don't have the hardware to test.

Signed-off-by: Stig Telfer <stig@lizardlogic.co.uk>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>