- 01 Nov, 2004 40 commits
-
-
Alan Stern authored
This patch fixes a tiny SMP-type hole in root-hub synchronization. Although the HCD glue layer properly unlinks root-hub status URBs synchronously, it doesn't do so for URBs sent to endpoint 0. This patch copies some code from usb_kill_urb, to make such unlinks wait until the host controller driver has finished handling the URB. This behavior is required for hcd_endpoint_disable to work correctly. The patch also renames usb_rh_status_dequeue to usb_rh_urb_dequeue (to better describe its updated function) and declares the routine static. Signed-off-by:
Alan Stern <stern@rowland.harvard.edu> Signed-off-by:
Greg Kroah-Hartman <greg@kroah.com>
-
Alan Stern authored
This patch, suggested by Karsten Wiese, converts a few remainder ('%') operations in the UHCI driver to bitwise-and ('&'). It's not a huge change, but this is a common idiom in C and it will save a few bytes with some compilers. Also one of the changes is in an inner loop, so it might help a little bit. From: Karsten Wiese <annabellesgarden@yahoo.de> Signed-off-by: -
David Brownell authored
Minor updates for better diagnostics from usbcore: - Only print ep0 maxpacket when it's not what we first tried - have khubd say when it's looking at a hub - When warning about synchronous control or bulk timeouts, report the program that triggered the message. Signed-off-by:
David Brownell <dbrownell@users.sourceforge.net> Signed-off-by:
-
David Brownell authored
This marks the controller as running at the end of a "resume" path used when the controller lost power, or was otherwise reset (not a real resume). Signed-off-by:
-
David Brownell authored
This patch much improves suspend/resume behavior of EHCI, as tested with "echo suspend > /sys/power/state" and "echo disk > /sys/power/state". - Quiescing the controller has to handle a case where the software has finished shutting down periodic and async schedules, but the hardware hasn't yet finished its part. - When suspending the root hub, be more by-the-book so that resumes behave better. - When resuming after suspend-to-disk, or in general when no port is suspended, the driver re-uses the initialization logic. That logic needed to understand that resources don't need re-allocation, and that a few things need to be reset to the default state. - More code paths needed to notice a suspended controller, and stop right away. - To shut down the controller, start with the root hub ports and work up from there. Tested by writing "suspend" or "disk" to /sys/power/state; several systems resumed OK, at least given the associated usbcore patches. Signed-off-by:
-
David Brownell authored
Recent PCI changes are killing USB Power Management, and maybe more. The root cause is a change to the PCI suspend code from back in May, saving a copy of pci config space after the PCI driver did. That was OK until a recent change made it overwrite the copy that the PCI device driver would later restore ... net result for USB is that DMA would always be disabled after resume. This patch restores the long-documented behavior of PCI driver suspend/resume calls (drivers are responsible for saving and restoring config space state), and also removes a superfluous line from the USB suspend logic. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by:
-
David Brownell authored
This patch makes usbcore behave better with system sleep states, fixing some PM/PCI integration problems as well as some PM/USB ones. - PCI HCDs should behave better, even if they don't support PCI PM; tested with both ACPI and APM suspend cycles, and swsusp. - Systems that advertise ACPI S1 support don't necessarily support the PCI D1 or D2 states for their USB controllers. (Examples include Centrino laptops.) This patch makes them suspend into D3hot, which always works (and won't take much longer to resume, either). - Handle the "go into deeper suspend" transition the way PCI drivers are supposed to. - Understand that USB only has one suspend state: if the device is suspended, it can't do any better. - Export some symbols to HCDs, so they can reuse more of the existing usbcore framework. Plus minor related cleanups. The PCI D1/D2 state issue is a variant of an earlier problem, with the same root cause: the PM core thinking there's a one-to-one mapping between system states (roughly: ACPI S0/S1/S3/S4) and PCI states that a given device can support (only D0 is required, D3hot is common). That same issue comes up again with USB. Signed-off-by:
-
Arjan van de Ven authored
Hi, Patch below removes somem dead, never used code from the usbvideo driver, including a function that uses interruptible_sleep_on_timeout() oddly. Signed-off-by:
Arjan van de Ven <arjan@fenrus.demon.nl> Signed-off-by:
-
Alan Stern authored
This patch adds a workaround for host controllers that don't properly support remote wakeup. That includes Intel controllers (on certain badly designed motherboards) and a recent Genesys Logic controller. Although the UHCI driver avoids suspending the Intel-based controllers, this workaround will handle things about as well. In the not-too-distant future I will have to revamp most of the root-hub suspend/resume code in the UHCI driver. For now, please apply the patch. Signed-off-by:
-
Lonnie Mendez authored
This patch adds the Delorme Earthmate usb gps and the Cypress hid->com rs232 adapter to hid-core's device blacklist. Signed-off-by:
Lonnie Mendez <lmendez19@austin.rr.com> Signed-off-by:
-
Nishanth Aravamudan authored
Description: Use msleep() instead of schedule_timeout() to guarantee the task delays as expected. Signed-off-by:
Nishanth Aravamudan <nacc@us.ibm.com> Signed-off-by:
-
Christoph Lameter authored
- fix broken IBM cyclone time interpolator support - add support for cyclic timers through an addition of a mask in the timer interpolator structure - Allow time_interpolator_update() and time_interpolator_get_offset() to be invoked without an active time interpolator (necessary since the cyclone clock is initialized late in ACPI processing) - remove obsolete function time_interpolator_resolution() - add a mask to all struct time_interpolator setups in the kernel - Make time interpolators work on 32bit platforms Signed-off-by:
Christoph Lameter <clameter@sgi.com> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Kay Sievers authored
Move hotplug_path[] out of kmod.[ch] to kobject_uevent.[ch] where it belongs now. At some time in the future we should fix the remaining bad hotplug calls (no SEQNUM, no netlink uevent): ./drivers/input/input.c (no DEVPATH on some hotplug events!) ./drivers/pnp/pnpbios/core.c ./drivers/s390/crypto/z90main.c Signed-off-by:
Kay Sievers <kay.sievers@vrfy.org> Signed-off-by:
-
Dmitry Torokhov authored
Driver core: when binding device to a driver create "driver" symlink in device's directory. Rename serio's "driver" attribute to "drvctl" (write-only) Signed-off-by:Dmitry Torokhov <dtor@mail.ru> Signed-off-by:
-
Kay Sievers authored
On Sat, Oct 30, 2004 at 04:54:29AM +0200, Kay Sievers wrote: > On Sat, Oct 30, 2004 at 02:25:23AM +0200, Kay Sievers wrote: > > On Sat, Oct 30, 2004 at 02:00:45AM +0200, Kay Sievers wrote: > > > On Fri, Oct 29, 2004 at 06:13:19PM -0500, Greg KH wrote: > > > > On Fri, Oct 29, 2004 at 11:28:56PM +0200, Kay Sievers wrote: > > > > > > But there might still be a problem. With this change, the sequence > > > > > > number is not sent out the kevent message. Kay, do you think this is an > > > > > > issue? I don't think we can get netlink messages out of order, right? > > > > > > > > > > Right, especially not the events with the same DEVPATH, like "remove" > > > > > beating an "add". But I'm not sure if the number isn't useful. Whatever > > > > > we may do with the hotplug over netlink in the future, we will only have > > > > > /sbin/hotplug for the early boot and it may be nice to know, what events > > > > > we have already handled... > > > > > > > > > > > I'll hold off on applying this patch until we figure this out... > > > > > > > > > > How about just reserving 20 bytes for the number (u64 will never be > > > > > more than that), save the pointer to that field, and fill the number in > > > > > later? > > > > > > > > Ah, something like this instead? I like it, it's even smaller than the > > > > previous patch. Compile tested only... > > > > > > I like that. How about the following. It will keep the buffer clean from > > > random chars, cause the kevent does not have the vector and relies on > > > the '\0' to separate the strings from each other. > > > I've tested it. The netlink-hotplug message looks like this: > > > > > > recv(3, "remove@/class/input/mouse2\0ACTION=remove\0DEVPATH=/class/input/mouse2\0SUBSYSTEM=input\0SEQNUM=961 \0", 1024, 0) = 113 > > > > Hmm, these trailing spaces are just bad, sorry. I'll better pass the > > envp array over to send_uevent() and clean up the keys while copying > > the env values into the skb buffer. This will make the event payload > > more safe too. So your first version looks better. > > How about this? We copy over key by key into the skb buffer and the > netlink message can get the envp array without depending on a single > continuous buffer. > > The netlink message looks nice like this now: > > recv(3, " > add@/devices/pci0000:00/0000:00:1d.1/usb3/3-2/3-2:1.0\0 > HOME=/\0 > PATH=/sbin:/bin:/usr/sbin:/usr/bin\0 > ACTION=add\0 > DEVPATH=/devices/pci0000:00/0000:00:1d.1/usb3/3-2/3-2:1.0\0 > SUBSYSTEM=usb\0 > SEQNUM=991\0 > DEVICE=/proc/bus/usb/003/008\0 > PRODUCT=46d/c03e/2000\0 > TYPE=0/0/0\0 > INTERFACE=3/1/2\0 > ", 1024, 0) = 268 Here is an improved version that uses skb_put() to fill the skb buffer, instead of trimming the buffer to the final size after we've copied over all keys. Signed-off-by:
-
Dmitry Torokhov authored
Driver core: rename bus_match into driver_probe_device and export it so subsystems can bind an individual device to a specific driver without getting involved with driver core internals. Signed-off-by: -
Dmitry Torokhov authored
Driver core: make device_attach() global and export it and driver_attach() so subsystems can have finer control over binding process. Signed-off-by: -
Andrew Morton authored
From: Paul Mackerras <paulus@samba.org> Currently the device_pm_foo() functions are rather prone to deadlocks during suspend/resume. This is because the dpm_sem is held for the duration of device_suspend() and device_resume() as well as device_pm_add() and device_pm_remove(). If for any reason you get a device addition or removal triggered by a device's suspend or resume code, you get a deadlock. (The classic example is a USB host adaptor resuming and discovering that the mouse you used to have plugged in has gone away.) This patch fixes the problem by using a separate semaphore, called dpm_list_sem, to cover the places where we need the device pm lists to be stable, and by being careful about how we traverse the lists on suspend and resume. I have analysed the various cases that can occur and I am confident that I have handled them all correctly. I posted this patch together with a detailed analysis 10 days ago. Signed-off-by Paul Mackerras <paulus@samba.org> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
-
Andrew Morton authored
From: Simon Derr <Simon.Derr@bull.net> Add a `needs_read_fill' field in sysfs_buffer so that reading after a write in a sysfs file returns valid data. (instead of the data that have been written, that may be invalid or at the wrong offset) Signed-off-by:
Simon Derr <simon.derr@bull.net> Signed-off-by:
-
Andrew Morton authored
From: Simon Derr <Simon.Derr@bull.net> - fixes the race between threads by adding a semaphore in sysfs_buffer - allocates the buffer upon call to pread(). We still call again fill_read_buffer() if the file is "rewinded" to offset zero. - fixes the comparison in flush_read_buffer(). Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> o This patch stops the pinning of non-directory or leaf dentries and inodes. The leaf dentries and inodes are created during lookup based on the entries on sysfs_dirent tree. These leaves are removed from the dcache through the VFS dentry ageing process during shrink dcache operations. Thus reducing about 80% of sysfs lowmem needs. o This implments the ->lookup() for sysfs directory inodes and allocates dentry and inode if the lookup is successful and avoids the need of allocating and pinning of dentry and inodes during the creation of corresponding sysfs leaf entry. As of now the implementation has not required negative dentry creation on failed lookup. As sysfs is still a RAM based filesystem, negative dentries are not of any use IMO. o The leaf dentry allocated after successful lookup is connected to the existing corresponding sysfs_dirent through the d_fsdata field. Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> o This patch implements the sysfs_dir_operations file_operations strucutre for sysfs directories. It uses the sysfs_dirent based tree for ->readdir() and ->lseek() methods instead of simple_dir_operations which use dentry based tree. Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> o This patch uses the sysfs_dirent based tree while removing sysfs files and directories. This avoids holding dcache_lock by not using dentry based vfs tree. Thus simplyfying the removal logic in sysfs. o It uses two helper routines sysfs_get_name(), to get the name for sysfs element and sysfs_drop_dentry() to delete the dentry given a sysfs_dirent. Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> o This patch introduces the new sysfs_dirent data structure. The sysfs_dirent is added to the dentry corresponding to each of the element which can be represented in sysfs like, kobject (directory), text or binary attributes (files), attribute groups (directory) and symlinks. o It uses dentry's d_fsdata field to attach the corresponding sysfs_dirent. o The sysfs_dirents are maintained in a tree of all the sysfs entries using the s_children and s_sibling list pointers. o This patch also changes how we access attributes and kobjects in file_operations from a given dentry, basically introducing one more level of indirection. o The sysfs_dirents are freed and the sysfs_dirent tree is updated accordingly upon the deletion of corresponding dentry. The sysfs dirents are kept alive as long as there is corresponding dentry around. The are freed when the dentry is finally out of dcache using the ->d_iput() method. o This also fixes the dentry leaks in case of error paths after sysfs has got a newly alocated (and hashed) dentry from sysfs_get_dentry() by d_drop()'ing the dentry. Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> My fault, a bad typo in fs/sysfs/bin.c. Signed-off-by:
-
Andrew Morton authored
From: Maneesh Soni <maneesh@in.ibm.com> o The following patch provides dumb helpers to access the corresponding kobject, attribute or binary attribute given a dentry and prepare the sysfs_file_operation methods for using sysfs_dirents. Signed-off-by:
-
Matthew Wilcox authored
Add light-weight-syscall support to parisc. The first LWS is atomic CAS. Committed-by:Carlos O'Donell <carlos@parisc-linux.org>
-
Matthew Wilcox authored
Add user_space macro, to determine if we are in a user versus kernel space. Committed-by: -
Matthew Wilcox authored
Calling sys_restart_syscall without having taking a signal across a syscall leaves the restart_block uninitialized, allowing malicious user code to crash the kernel. FIX: Define INIT_THREAD_INFO such that restart_block.fn is set to the default do_not_restart_syscall. Committed-by: -
Matthew Wilcox authored
Enable SMP on A500 by default Committed-by:Paul Bame <bame@parisc-linux.org>
-
Matthew Wilcox authored
Fix N class SMP The main fixes are: - memory barriers to our space and control register moves - fix for the N class merced bus problem which can't have more than one PxTLB broadcast outstanding at once - make smp_call_function() wait until the function completes This now boots and runs on a 32MB N4000 in Fort Collins with 2 cpus and discontig mem support. Committed-by:
James Bottomley <jejb@parisc-linux.org> Committed-by:
Randolph Chung <tausq@parisc-linux.org>
-
Matthew Wilcox authored
Add a make install target Committed-by:Matthew Wilcox <willy@parisc-linux.org>
-
Matthew Wilcox authored
PA-RISC does have spinlock debugging and doesn't have the frame pointer option. USERMODE had been added twice to spinlock debugging, so remove that. Committed-by:
-
Matthew Wilcox authored
add some more BL macros and b,l in appropriate places in entry.S to expand the range of our jumps and work around the 64 bit binutils problem where it will allow the jump ranges to go beyond their permitted limits without complaining. Committed-by: -
Matthew Wilcox authored
add HP copyright Committed-by:Grant Grundler <grundler@parisc-linux.org>
-
Matthew Wilcox authored
small debug cleanup Committed-by: -
Matthew Wilcox authored
Merge head64.S and head.S Cleanup 64-bit sp manipulations Committed-by:
Kyle McMartin <kyle@parisc-linux.org> Committed-by:
-
-
Bartlomiej Zolnierkiewicz authored
CONFIG_IDEDMA_PCI_AUTO and CONFIG_IDEDMA_ICS_AUTO Host drivers should deal with broken hardware themselves. Warn if DMA support is enabled but "enable DMA by default" is not. Signed-off-by:Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
-
Bartlomiej Zolnierkiewicz authored
"hdx=autotune", "hdx=noautotune": * should be handled by host drivers needing them "idex=autotune", "idex=noautotune": * should die "idex=ata66": * should be handled by host drivers needing it "idex=dma": * works only for: cs5220.c, generic.c, hpt366.c, triflex.c * DMA should be used by default "idex=reset", "idex=serialize": * host drivers should set these settings when needed "idex=base[,ctl[,irq]]": * host drivers should auto-detect correct settings * ordering should be controlled by user-space "ide0=four": * should be handled by ide-generic driver Signed-off-by:
-
