- 27 Apr, 2007 40 commits
-
-
Oliver Neukum authored
This fixes: - breaking DMA rules about buffers - usage of _global_ variables to save a single device's attributes - racy access to urb->status - smp monotonity issue with statistics - use of one buffer for many simultaneous URBs - error handling introduced - several instances of following NULL pointers - use after free - unnecessary GFP_ATOMIC - GFP_KERNEL in interrupt - various cleanups - write room granularity issue that bit cdc-acm - race in shutdown Signed-off-by:
Oliver Neukum <oneukum@suse.de> Signed-off-by:
Greg Kroah-Hartman <gregkh@suse.de>
-
Oliver Neukum authored
io_edgeport is using a global variable without locking. This is _the_ classical race condition. This patch switches to atomic_t. Signed-off-by:
-
Alan Stern authored
This patch (as870) adds a delay to ehci-hcd's bus_resume routine. Apparently there are controllers and/or BIOSes out there which need such a delay to get the ports back into their correct state. This fixes Bugzilla #8190. Signed-off-by:
Alan Stern <stern@rowland.harvard.edu> Cc: David Brownell <david-b@pacbell.net> Signed-off-by:
-
Alan Stern authored
This patch (as867) adds an entry for the new power/autosuspend attribute in Documentation/ABI/testing, and it changes the behavior of the delay value. Now a delay of 0 means to autosuspend as soon as possible, and negative values will prevent autosuspend. Signed-off-by:
-
Alan Stern authored
This patch (as866) adds new entry points for external USB device suspend and resume requests, as opposed to internally-generated autosuspend or autoresume. It also changes the existing remote-wakeup code paths to use the new routines, since remote wakeup is not the same as autoresume. As part of the change, it turns out to be necessary to do remote wakeup of root hubs from a workqueue. We had been using khubd, but it does autoresume rather than an external resume. Using the ksuspend_usb_wq workqueue for this purpose seemed a logical choice. Signed-off-by:
-
Oliver Neukum authored
this driver's help text incorrectly claims to support only single port devices. Signed-off-by:
-
Oliver Neukum authored
this driver has an interesting way of handling ENOMEM: complain and ignore. If you decide to live with allocation failures, you must 1. guard against URBs without corresponding buffers 2. complete allocation failures 3. always test entries for NULL before you follow the pointers This patch does so. Signed-off-by:
-
Christian Engelmayer authored
Correct the offsets of the SI_CTRL, PRI_CTRL registers according to the Reference Manual errata sheet in order to prevent unwanted settings regarding burst transactions and priority states. Signed-off-by:
Christian Engelmayer <Christian.Engelmayer@frequentis.com> Signed-off-by:
-
Alan Stern authored
This patch (as865) fixes a race in the HCD removal code discovered by Milan Plzik. Arrival of an interrupt after the root hub was unregistered could cause the root-hub status timer to start up, even after it was supposed to have been shut down. The problem is fixed by moving the del_timer_sync() call to after the HCD's stop() method, at which time IRQ generation should be disabled. Cc: Milan Plzik <milan.plzik@gmail.com> Signed-off-by:
-
Simon Arlott authored
I've acquired a second device for testing and plan to make some changes in the near future to export all the device stats to sysfs (based on my proposed patch to add them to the proc file ~2007-01-30). Signed-off-by:
Simon Arlott <simon@fire.lp0.eu> Acked-by:
Duncan Sands <baldrick@free.fr> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Simon Arlott authored
When the device is polled for status there is a lot of useful status information available that is ignored. This patch stores the device info array when the status is polled and adds sysfs files to the usb device to allow userspace to query it. Since the device updates its status internally once a second the poll time is changed to this, and round_jiffies_relative is used to avoid waking the cpu unnecessarily. Signed-off-by:
-
Simon Arlott authored
There is currently no path from the ATM device in /sys to the USB device's interface that the driver is using; this patch creates a "device" symlink. It is then possible to get to the cxacru ADSL statistics (http://lkml.org/lkml/2007/2/23/328): /sys/class/atm/cxacru0/device $ ls *_rate *_margin *_attenuation downstream_attenuation downstream_snr_margin upstream_rate downstream_rate upstream_attenuation upstream_snr_margin If this link is not appropriate I'd have to create device files in /sys/class/atm/cxacru0 instead - which seems less appropriate since the ADSL statistics are for the USB device not ATM (which is running over the ADSL). [akpm@linux-foundation.org: cleanups] Signed-off-by:
-
Oliver Neukum authored
GFP_KERNEL will do. Signed-off-by:
-
Alan Stern authored
This patch (as864) moves the work routine for USB autosuspend from one source file to another. This permits the removal of one whole global symbol (!) and should smooth the way for more changes in the future. Signed-off-by:
-
Pete Zaitcev authored
Someone changed the code to kthread and used his style instead of mine. The problem with the block variables is that they provoke shadowing, which is actually exactly what has happened in my other tree which has the class patch. Signed-off-by:
Pete Zaitcev <zaitcev@redhat.com> Signed-off-by:
-
Pete Zaitcev authored
The usb-storage switched to binding to first endpoint recently. Apparently, there are devices out there with extra endpoints. It is perfectly legal. Signed-off-by:
-
Oliver Neukum authored
this patch exports the attributes cdc-acm knows about a device through sysfs. Signed-off-by:
-
Oliver Neukum authored
this adds another structure for CDC devices to cdc.h. Signed-off-by:
-
Pete Zaitcev authored
This patch adds a new text API, codenamed '1u', which captures more URB fields than old '1t' interface did. Also the '1u' text API is compatible with the future "bus zero" extension. Signed-off-by:
-
Oliver Neukum authored
this patch removes usage of BKL from usblcd, which got it from the old skeleton driver. Signed-off-by:
-
Oliver Neukum authored
Iet's kill BKL where we can. This is relative to the last patch to the skeleton driver. Signed-off-by:
-
Oliver Neukum authored
compilation of the skeleton driver is currently broken. It doesn't compile. So while I am it: - fix typo - add comments to answer common questions - actually allow autosuspend in the driver struct - increase paralellism by restricting code under locks Signed-off-by:
-
Milan Svoboda authored
This patch changes last use of hardcoded number of irq to use platfrom_get_irq. Signed-off-by:
Milan Svoboda <msvoboda@ra.rockwell.com> Acked-by:
David Brownell <david-b@pacbell.net> Signed-off-by:
-
David Brownell authored
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 cleanups Signed-off-by:
David Brownell <dbrownell@users.sourceforge.net> Signed-off-by:
-
Norihiko Tomiyama authored
I write a patch adding support "SHARP EMONE(S01SH)" device for ipaq.c. EMONE is a PDA with built-in HSDPA function. From: Norihiko Tomiyama <norihiko.tomiyama@ctc-g.co.jp> Signed-off-by: -
Linus Torvalds authored
* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6: (46 commits) dev_dbg: check dev_dbg() arguments drivers/base/attribute_container.c: use mutex instead of binary semaphore mod_sysfs_setup() doesn't return errno when kobject_add_dir() failure occurs s2ram: add arch irq disable/enable hooks define platform wakeup hook, use in pci_enable_wake() security: prevent permission checking of file removal via sysfs_remove_group() device_schedule_callback() needs a module reference s390: cio: Delay uevents for subchannels sysfs: bin.c printk fix Driver core: use mutex instead of semaphore in DMA pool handler driver core: bus_add_driver should return an error if no bus debugfs: Add debugfs_create_u64() the overdue removal of the mount/umount uevents kobject: Comment and warning fixes to kobject.c Driver core: warn when userspace writes to the uevent file in a non-supported way Driver core: make uevent-environment available in uevent-file kobject core: remove rwsem from struct subsystem qeth: Remove usage of subsys.rwsem PHY: remove rwsem use from phy core IEEE1394: remove rwsem use from ieee1394 core ...
-
Dan Williams authored
Duplicate what Zach Brown did for pr_debug in commit 8b2a1fd1 [akpm@linux-foundation.org: fix a couple of things which broke] Signed-off-by:
Dan Williams <dan.j.williams@intel.com> Signed-off-by:
-
Matthias Kaehlcke authored
use mutex instead of binary semaphore in drivers/base/attribute_container.c Signed-off-by:
Matthias Kaehlcke <matthias.kaehlcke@gmail.com> Signed-off-by:
-
Akinobu Mita authored
mod_sysfs_setup() doesn't return an errno when kobject_add_dir() for module "holders" directory fails. So caller of mod_sysfs_setup() will keep going and get oops. Signed-off-by:
Akinobu Mita <akinobu.mita@gmail.com> Signed-off-by:
-
Johannes Berg authored
After some more discussion this patch replaces it: From: Johannes Berg <johannes@sipsolutions.net> Subject: suspend: add arch irq disable/enable hooks For powermac, we need to do some things between suspending devices and device_power_off, for example setting the decrementer. This patch allows architectures to define arch_s2ram_{en,dis}able_irqs in their asm/suspend.h to have control over this step. Signed-off-by:Johannes Berg <johannes@sipsolutions.net> Acked-by:
Pavel Machek <pavel@ucw.cz> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
David Brownell authored
This defines a platform hook to enable/disable a device as a wakeup event source. It's initially for use with ACPI, but more generally it could be used whenever enable_irq_wake()/disable_irq_wake() don't suffice. The hook is called -- if available -- inside pci_enable_wake(); and the semantics of that call are enhanced so that support for PCI PME# is no longer needed. It can now work for devices with "legacy PCI PM", when platform support allows it. (That support would use some board-specific signal for for the same purpose as PME#.) [akpm@linux-foundation.org: Make it compile with CONFIG_PM=n] Signed-off-by:
Zhang Rui <rui.zhang@intel.com> Cc: Len Brown <lenb@kernel.org> Signed-off-by:
-
James Morris authored
Prevent permission checking from being performed when the kernel wants to unconditionally remove a sysfs group, by introducing an kernel-only variant of lookup_one_len(), lookup_one_len_kern(). Additionally, as sysfs_remove_group() does not check the return value of the lookup before using it, a BUG_ON has been added to pinpoint the cause of any problems potentially caused by this (and as a form of annotation). Signed-off-by:
James Morris <jmorris@namei.org> Cc: Nagendra Singh Tomar <nagendra_tomar@adaptec.com> Cc: Tejun Heo <htejun@gmail.com> Cc: Stephen Smalley <sds@tycho.nsa.gov> Cc: Eric Paris <eparis@redhat.com> Signed-off-by:
-
Alan Stern authored
This patch (as896b) fixes an oversight in the design of device_schedule_callback(). It is necessary to acquire a reference to the module owning the callback routine, to prevent the module from being unloaded before the callback can run. Signed-off-by:
-
Cornelia Huck authored
We often have the situation that we register a subchannel and start device recognition, only to find out that the device is not usable after all, which triggers an unregister of the subchannel. This often happens on hundreds of subchannels on a LPAR, leading to a storm of events which aren't of any use. Therefore, use uevent_suppress to delay the KOBJ_ADD uevent for a subchannel until we know that its ccw_device is to be registered. Signed-off-by:
Cornelia Huck <cornelia.huck@de.ibm.com> Cc: Kay Sievers <kay.sievers@vrfy.org> Cc: Eric Rannaud <eric.rannaud@gmail.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
-
Andrew Morton authored
fs/sysfs/bin.c: In function 'read': fs/sysfs/bin.c:77: warning: format '%zd' expects type 'signed size_t', but argument 4 has type 'int' Signed-off-by:
-
Matthias Kaehlcke authored
the DMA pool handler uses a semaphore as mutex. use the mutex API instead of the (binary) semaphore Signed-off-by:
-
Greg Kroah-Hartman authored
As pointed out by Dave Jones. Cc: Dave Jones <davej@redhat.com> Signed-off-by: -
Michael Ellerman authored
I went to use this the other day, only to find it didn't exist. It's a straight copy of the debugfs u32 code, then s/u32/u64/. A quick test shows it seems to be working. Signed-off-by:
Michael Ellerman <michael@ellerman.id.au> Signed-off-by:
-
Adrian Bunk authored
This patch contains the overdue removal of the mount/umount uevents. Signed-off-by:
Adrian Bunk <bunk@stusta.de> Signed-off-by:
-
Eric W. Biederman authored
This dots some i's and crosses some t's after left over from when kobject_kset_add_dir was built from kobject_add_dir. Signed-off-by:
Eric W. Biederman <ebiederm@xmission.com> Cc: Kay Sievers <kay.sievers@vrfy.org> Signed-off-by:
-
