Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
Thanks to Andrew for the original patch for this.
I need to upgrade my version of gcc to catch these things...
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as790b) adds "autostop" support to ohci-hcd: the driver
will automatically stop the host controller when no devices have been
connected for at least one second. This feature is useful when the
USB autosuspend facility isn't available, such as when
CONFIG_USB_SUSPEND hasn't been set.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The autosuspend technique used by ohci-hcd doesn't mesh well with the
newer USB core autosuspend code. This patch (as789) removes ohci-hcd's
autosuspend support. Now the driver will be usable, but it won't
automatically go into a low-power state when no devices are connected.
That's for a later patch.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as740) removes the existing support for autosuspend of
root hubs. That support fit in rather awkwardly with the rest of
usbcore and it was used only by ohci-hcd. It won't be needed any more
since the hub driver will take care of autosuspending all hubs, root
or external.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This teaches OHCI to use the root hub status change (RHSC) IRQ, bypassing
root hub timers most of the time and switching over to the "new" root hub
polling scheme. It's complicated by the fact that implementations of OHCI
trigger and ack that IRQ differently (the spec is vague there).
Avoiding root hub timers helps mechanisms like "dynamic tick" leave the
CPU in lowpower modes for longer intervals.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
I noticed this while debugging something unrelated on
sparc64.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: David Brownell <david-b@pacbell.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This makes OHCI use the driver model wakeup control bits for its root hub
(e.g. disable on amd756, because of chip erratum) and for the controller
itself. It no longer uses the hcd glue bits with those roles, and depends
on the previous patch making the root hub available earlier.
Note that on most platforms (boot code properly setting the RWC bit) this
gives a partial workaround for the way PCI isn't currently flagging devices
that support PME# signals. (Because of odd PCI init sequencing on PPC.)
That's because many OHCI controllers support "legacy PCI PM" ... without
involving any PCI PM capability.
USB wakeup from STR, if it works on your system, may still involve
tweaking things by hand in /proc/acpi/wakeup.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as605) removes the private udev->serialize semaphore,
relying instead on the locking provided by the embedded struct device's
semaphore. The changes are confined to the core, except that the
usb_trylock_device routine now uses the return convention of
down_trylock rather than down_read_trylock (they return opposite values
for no good reason).
A couple of other associated changes are included as well:
Now that we aren't concerned about HCDs that avoid using the
hcd glue layer, usb_disconnect no longer needs to acquire the
usb_bus_lock -- that can be done by usb_remove_hcd where it
belongs.
Devices aren't locked over the same scope of code in
usb_new_device and hub_port_connect_change as they used to be.
This shouldn't cause any trouble.
Along with the preceding driver core patch, this needs a lot of testing.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch closes the IRQ race and makes various other OHCI & EHCI code
path safer vs. suspend/resume.
I've been able to (finally !) successfully suspend and resume various
Mac models, with or without USB mouse plugged, or plugging while asleep,
or unplugging while asleep etc... all without a crash.
Alan, please verify the UHCI bit I did, I only verified that it builds.
It's very simple so I wouldn't expect any issue there. If you aren't
confident, then just drop the hunks that change uhci-hcd.c
I also made the patch a little bit more "safer" by making sure the store
to the interrupt register that disables interrupts is not posted before
I set the flag and drop the spinlock.
Without this patch, you cannot reliably sleep/wakeup any recent Mac, and
I suspect PCs have some more sneaky issues too (they don't frankly crash
with machine checks because x86 tend to silently swallow PCI errors but
that won't last afaik, at least PCI Express will blow up in those
situations, but the USB code may still misbehave).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as580) is perhaps the only result from the long discussion I
had with David about his changes to the root-hub suspend/resume code. It
renames the hub_suspend and hub_resume methods in struct usb_hcd to
bus_suspend and bus_resume. These are more descriptive names, since the
methods really do suspend or resume an entire USB bus, and less likely to
be confused with the hub_suspend and hub_resume routines in hub.c.
It also takes David's advice about removing the layer of bus glue, where
those methods are called. And it implements a related change that David
made to the other HCDs but forgot to put into dummy_hcd.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
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(-)
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(-)
Some OHCI implementations have differences in the way the NDP register
(in roothub_a) reports the number of ports present. This patch allows the
platform specific code to optionally supply the number of ports. The
driver just reads the value at init (if not supplied) instead of reading
it every time its needed (except for an AMD756 bug workaround).
It also sets the value correctly for the ARM pxa27x architecture.
Signed-Off-By: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
The length of the array desc->bitmap is 3, and not 4:
Definitions involved:
In drivers/usb/core/hcd.h
464 #define bitmap DeviceRemovable
In drivers/usb/host/ohci-hub.c
395 struct usb_hub_descriptor *desc
In drivers/usb/core/hub.h
130 struct usb_hub_descriptor {
131 __u8 bDescLength;
132 __u8 bDescriptorType;
133 __u8 bNbrPorts;
134 __u16 wHubCharacteristics;
135 __u8 bPwrOn2PwrGood;
136 __u8 bHubContrCurrent;
137 /* add 1 bit for hub status change; round to bytes */
138 __u8 DeviceRemovable[(USB_MAXCHILDREN + 1 + 7) / 8];
139 __u8 PortPwrCtrlMask[(USB_MAXCHILDREN + 1 + 7) / 8];
140 } __attribute__ ((packed));
In include/linux/usb.h
306 #define USB_MAXCHILDREN (16)
This defect was found automatically by Coverity Prevent, a static analysis
tool.
(akpm: this code should be shot. Field `bitmap' doesn't exist in struct
usb_hub_descriptor. And this .c file is #included in
drivers/usb/host/ohci-hcd.c, and someone somewhere #defines `bitmap' to
`DeviceRemovable'.
>From a maintainability POV it would be better to memset the whole array
beforehand - I changed the patch to do that)
Signed-off-by: Zaur Kambarov <zkambarov@coverity.com>
Cc: <linux-usb-devel@lists.sourceforge.net?
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
