Alan Cox [Thu, 22 May 2008 21:06:02 +0000 (22:06 +0100)]
USB: iowarrior: Push down BKL
I'm pretty sure the mutex is sufficient for all locking but will come
back to that later if the USB folks don't beat me to it. For now get rid
of the old BKL ioctl method and wrap the ioctl handler
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Cox [Thu, 22 May 2008 21:07:51 +0000 (22:07 +0100)]
USB: usblcd: Push down BKL into driver
I'm pretty sure this can be eliminated however I couldn't prove (or find)
what stopped the device vanishing mid IOCTL_GET_HARD_VERSION. Perhaps a
USB wizard could double check that and see if the lock_kernel can go
entirely.
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:20:11 +0000 (18:20 -0700)]
usb ethernet gadget: use composite gadget framework
Building on the previous patches which took code from this driver and
pakaged it in more-reusable network "function" components, this patch
gets rid of the original code and uses those components instead.
As seen with the other gadget driver conversions, the resulting code
is much easier to understand and (presumably) work with. In this case
that's especially true, since the Ethernet gadget had grown to handle
three (!) different Ethernet-over-USB protocols. This modularization
should make it much easier to add a fourth option for the newish CDC
"Ethernet Emulation Model" (or EEM).
Lightly tested, primarily at full speed.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:20:26 +0000 (18:20 -0700)]
usb gadget: new "CDC Composite" gadget driver
This is a simple example of a composite gadget, combining two
Communications Class Device (CDC) functions: ECM and ACM.
This provides a clear example of how the composite gadget framework
is intended to work. It's surprising that MS-Windows (or at least,
XP and previous) won't "just work" with something this simple...
Not all USB peripheral controller hardware can support this driver.
All the highspeed-capable peripheral controllers with drivers now in
the mainline kernel seem to support this, as does omap_udc. But
many full speed controllers don't have enough endpoints, or (as with
the PXA controllers) don't support altsettings.
Lightly tested.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:19:46 +0000 (18:19 -0700)]
usb ethernet gadget: split CDC Ethernet function
This is a "CDC Ethernet" (ECM) function driver, extracted from the
all-in-one Ethernet gadget driver.
This is a good example of how to implement interface altsettings.
In fact it's currently the only such example in the gadget stack,
pending addition of OBEX support.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:19:28 +0000 (18:19 -0700)]
usb ethernet gadget: split out network core
Abstract the peripheral side Ethernet-over-USB link layer code from
the all-in-one Ethernet gadget driver into a component that can be
called by various functions, so the various flavors can be split
apart and selectively reused.
A notable difference from the approach taken with the serial link
layer code (beyond talking to NET not TTY) is that because of the
initialization requirements, this only supports one network link.
(And one set of Ethernet link addresses.)
That is, each configuration may have only one instance of a network
function. This doesn't change behavior; the current code has that
same restriction. If you want multiple logical links, that can
easily be done using network layer tools.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:19:16 +0000 (18:19 -0700)]
usb gadget: RNDIS cleanups
Some cleanup to the RNDIS code:
- Minor bugfix: rndis_unit() is supposed to put the link into the
RNDIS_UNINITIALIZED state, which does not mean "unused". There's
a separate method to stop using the link. (Bug doesn't affect
anything right now because of how the code is used.)
- Reduce coupling between RNDIS code and its user(s), in preparation
for updates in that code:
* Decouple RNDIS_RESPONSE_AVAILABLE notifications from net_device
by passing just a void* handle. (Also, remove the unused return
value of the notification callback.)
* When it needs a copy of net_device stats, just ask for it
- Remove unused/untested code backing various never-used OIDs:
* RNDIS_PM, RNDIS_WAKEUP ... "should" get implemented, but the
relevant docs were unclear, ambguous, and incomplete. Someone
with access to the Hidden Gospels (maybe in the EU?) might be
able to figure out what this should do.
* RNDIS_OPTIONAL_STATS ... as the name suggests, optional. Never
implemented in part because not all the semantics were clear.
* OID_GEN_RNDIS_CONFIG_PARAMETER, which has been #if 0 forever.
- A few small whitespace fixes
Plus switch the VERBOSE symbol over to the newer VERBOSE_DEBUG style.
There should be no functional changes because of this patch; it's a
net source code shrink (because of the dead/unused code removal) and
a small object code shrink (a couple hundred bytes on ARMv5).
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:19:03 +0000 (18:19 -0700)]
usb gadget serial: use composite gadget framework
This switches the serial gadget over to using the new "function"
versions of the serial port interfacing code. The remaining code
in the main source file is quite small...
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:18:40 +0000 (18:18 -0700)]
usb gadget serial: split out CDC ACM function
Split out CDC ACM parts of "gadget serial" to a "function driver".
Some key structural differences from the previous ACM support, shared
with with the generic serial function (next patch):
- As a function driver, it can be combined with other functions.
One gadget configuration could offer both serial and network
links, as an example.
- One serial port can be exposed in multiple configurations;
the /dev/ttyGS0 node could be exposed regardless of which
config the host selected.
- One configuration can expose multiple serial ports, such as
ttyGS0, ttyGS1, ttyGS2, and ttyGS3.
This code should be a lot easier to understand than the previous
all-in-one-big-file version of the driver.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 01:18:27 +0000 (18:18 -0700)]
usb gadget zero: use composite gadget framework
Update Gadget Zero to use the more modular versions of the loopback
and source/sink configuration drivers which build on the new gadget
framework code.
The core code is a LOT simpler, and it should be much easier now to
understand how the parts fit together. The conversion is an overall
source shrink in terms of this gadget, since it uses more midlayer
support. However, it's an overall increase in object size because
there's less sharing between the two configurations (improves code
clarity) and because the midlayer is a bit more functional than this
driver actually needs.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:55:35 +0000 (17:55 -0700)]
usb gadget zero: split out loopback config
This splits the gadget zero "loopback" configuration into a standalone
"configuration driver", building on the composite gadget framework code.
It doesn't yet pull the original code out of gadget zero or update how
that driver is built.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:55:23 +0000 (17:55 -0700)]
usb gadget zero: split out source/sink config
This splits the gadget zero "source/sink" configuration into a standalone
"configuration driver", building on the composite gadget framework code.
It doesn't yet pull the original code out of gadget zero or update how
that driver is built.
Neither this, nor its sibling "loopback" configuration, is a function
driver that can be combined with other functions. (The host "usbtest"
driver wouldn't know how to deal with that!) However the code becomes
simpler because of this conversion, so it's a net win.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:52:58 +0000 (17:52 -0700)]
usb gadget: composite gadget core
Add <linux/usb/composite.h> interfaces for composite gadget drivers, and
basic implementation support behind it:
- struct usb_function ... groups one or more interfaces into a function
managed as one unit within a configuration, to which it's added by
usb_add_function().
- struct usb_configuration ... groups one or more such functions into
a configuration managed as one unit by a driver, to which it's added
by usb_add_config(). These operate at either high or full/low speeds
and at a given bMaxPower.
- struct usb_composite_driver ... groups one or more such configurations
into a gadget driver, which may be registered or unregistered.
- struct usb_composite_dev ... a usb_composite_driver manages this; it
wraps the usb_gadget exposed by the controller driver.
This also includes some basic kerneldoc.
How to use it (the short version): provide a usb_composite_driver with a
bind() that calls usb_add_config() for each of the needed configurations.
The configurations in turn have bind() calls, which will usb_add_function()
for each function required. Each function's bind() allocates resources
needed to perform its tasks, like endpoints; sometimes configurations will
allocate resources too.
Separate patches will convert most gadget drivers to this infrastructure.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:52:25 +0000 (17:52 -0700)]
usb gadget: descriptor copying support
Define three new descriptor manipulation utilities, for use when
setting up functions that may have multiple instances:
usb_copy_descriptors() to copy a vector of descriptors
usb_free_descriptors() to free the copy
usb_find_endpoint() to find a copied version
These will be used as follows. Functions will continue to have static
tables of descriptors they update, now used as __initdata templates.
When a function creates a new instance, it patches those tables with
relevant interface and string IDs, plus endpoint assignments. Then it
copies those morphed descriptors, associates the copies with the new
function instance, and records the endpoint descriptors to use when
activating the endpoints. When initialization is done, only the copies
remain in memory. The copies are freed on driver removal.
This ensures that each instance has descriptors which hold the right
instance-specific data. Two instances in the same configuration will
obviously never share the same interface IDs or use the same endpoints.
Instances in different configurations won't do so either, which means
this is slightly less memory-efficient in some cases.
This also includes a bugfix to the epautoconf code that shows up with
this usage model. It must replace the previous endpoint number when
updating the template descriptors, not just mask in a few more bits.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:52:07 +0000 (17:52 -0700)]
usb gadget: use new serial core
Teach "gadget serial" to use the new abstracted (and bugfixed) TTY glue,
and remove all the orignal tangled-up code. Update the documentation
accordingly. This is a net object code shrink and cleanup; it should
make it a lot easier to see how the TTY glue should accomodate updates
to the TTY layer, be bugfixed, etc.
Notable behavior changes include: it can now support getty even when
there's no USB connection; it fits properly into the mdev/udev world;
and RX handling is better (throttling works, and low latency).
Configurations with scripts setting up the /dev/ttygserial device node
(with "experimental" major number) may want to change that to be a
symlink pointing to the /dev/ttyGS0 file, as a migration aid; else,
just switch entirely over to mdev/udev.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
David Brownell [Fri, 20 Jun 2008 00:51:44 +0000 (17:51 -0700)]
usb gadget: split out serial core
This abstracts the "gadget serial" driver TTY glue into a separate
component, cleaning it up and disentangling it from connection state.
It also changed some behaviors for the better:
- Stops using "experimental" major #127, and switches over to
having the TTY layer allocate the dev_t numbers.
- Provides /sys/class/tty/ttyGS* nodes, thus mdev/udev support.
(Note "mdev" hotplug bug in Busybox v1.7.2: /dev/ttyGS0 will
be a *block* device without CONFIG_SYSFS_DEPRECATED_V2.)
- The tty nodes no longer reject opens when there's no host.
Now they can support normal getty configs in /etc/inttab...
- Now implements RX throttling. When the line discipline says
it doesn't want any more data, only packets in flight will be
delivered (currently, max 1K/8K at full/high speeds) until it
unthrottles the data.
- Supports low_latency. This is a good policy for all USB serial
adapters, since it eliminates scheduler overhead on RX paths.
This also includes much cleanup including better comments, fixing
memory leaks and other bugs (including some locking fixes), messaging
cleanup, and an interface audit and tightening. This added up to a
significant object code shrinkage, on the order of 20% (!) depending
on CPU and compiler.
A separate patch actually kicks in this new code, using the functions
declared in this new header, and removes the previous glue.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Thu, 8 May 2008 15:55:59 +0000 (11:55 -0400)]
usb-storage: implement "soft" unbinding
This patch (as1092) implements "soft" unbinding for usb-storage. When
the disconnect routine is called, all commands and reset delays are
allowed to complete normally until after scsi_remove_host() returns.
This means that the commands needed for an orderly shutdown will be
sent through to the device.
Unlike before, the driver will now execute every command that it
accepts. Hence there's no need for special code to catch unexecuted
commands and fail them.
The new sequence of events when disconnect runs goes as follows:
If the device is truly unplugged, set the DISCONNECTING
flag so we won't try to access it any more.
If the SCSI-scanning thread hasn't started up yet, prevent
it from doing anything by setting the new DONT_SCAN flag.
Then wake it up and wait for it to terminate.
Remove the SCSI host. This unbinds the upper-level drivers,
doing an orderly shutdown. Commands sent to quiesce the
device will be transmitted normally, unless the device is
unplugged.
Set the DISCONNECTING flag so that we won't accept any new
commands that might get submitted (there aren't supposed to be
any) and we won't try to access the device for resets.
Tell the control thread to exit by waking it up with no
pending command, and wait for it to terminate.
Go on to do all the other normal stuff: releasing resources,
freeing memory, and so on.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Thu, 8 May 2008 15:54:37 +0000 (11:54 -0400)]
USB: implement "soft" unbinding
This patch (as1091) changes the way usbcore handles interface
unbinding. If the interface's driver supports "soft" unbinding (a new
flag in the driver structure) then in-flight URBs are not cancelled
and endpoints are not disabled. Instead the driver is allowed to
continue communicating with the device (although of course it should
stop before its disconnect routine returns).
The purpose of this change is to allow drivers to do a clean shutdown
when they get unbound from a device that is still plugged in. Killing
all the URBs and disabling the endpoints before calling the driver's
disconnect method doesn't give the driver any control over what
happens, and it can leave devices in indeterminate states. For
example, when usb-storage unbinds it doesn't want to stop while in the
middle of transmitting a SCSI command.
The soft_unbind flag is added because in the past, a number of drivers
have experienced problems related to ongoing I/O after their disconnect
routine returned. Hence "soft" unbinding is made available only to
drivers that claim to support it.
The patch also replaces "interface_to_usbdev(intf)" with "udev" in a
couple of places, a minor simplification.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Thu, 1 May 2008 19:35:18 +0000 (15:35 -0400)]
usb-storage: separate dynamic flags from fixed flags
This patch (as1089) separates out the dynamic atomic bitflags and the
static bitfields in usb-storage. Until now the two sorts of flags
have been sharing the same word; this has always been awkward.
To help prevent possible confusion, the two new fields each have a
different name from the original. us->fflags contains the fixed
bitfields (mostly taken from the USB ID table in unusual_devs.h), and
us->dflags contains the dynamic atomic bitflags (used with set_bit,
test_bit, and so on).
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Cc: Matthew Dharm <mdharm-usb@one-eyed-alien.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Cox [Tue, 29 Apr 2008 13:45:15 +0000 (14:45 +0100)]
USB: io_ti: FIrst cut at a big clean up
Sort out the insane naming like "OperationalFirmwareVersion" which seems
designed to cause formatting problems and RSI
Merge various common code together
Clean up the pointlessly complex and spread about MCR handling
This is really just the low hanging fruit.
Needs lots of testing before it goes upstream so testers and reports
appreciated
Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:07:31 +0000 (11:07 -0400)]
USB: combine hub_quiesce and hub_stop
This patch (as1083) combines hub_quiesce() and hub_stop() into a
single routine. There's no point keeping them separate since they are
usually called together.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:07:17 +0000 (11:07 -0400)]
USB: combine hub_activate and hub_restart
This patch (as1071) combines hub_activate() and hub_restart() into a
single routine. There's no point keeping them separate, since they
are always called together.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:07:07 +0000 (11:07 -0400)]
USB: optimize port debouncing during hub activation
This patch (as1082) makes a small optimization to the way the hub
driver carries out port debouncing immediately after a hub is
activated (i.e., initialized, reset, or resumed). If any port-change
statuses are observed, the code will delay for a minimal debounce
period -- thereby making a good start at debouncing all the ports at
once.
If this wasn't sufficient then khubd will debounce any port that still
requires attention. But in most cases it should suffice; it's rare
for a device to need more than a minimal debounce delay. (In the
cases of hub initialization or reset even that is most likely not
needed, since any devices plugged in at such times have probably been
attached for a while.)
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:06:55 +0000 (11:06 -0400)]
USB: try to salvage lost power sessions
This patch (as1073) adds to khubd a way to recover from power-session
interruption caused by transient connect-change or enable-change
events. After the debouncing period, khubd attempts to do a
USB-Persist-style reset or reset-resume. If it works, the connection
will remain unscathed.
The upshot is that we will be more immune to noise caused by EMI. The
grace period is on the order of 100 ms, so this won't permit recovery
from the "accidentally knocked the USB cable out of its socket" type
of event, but it's a start.
As an added bonus, if a device was suspended when the system goes to
sleep then we no longer need to check for power-session interruptions
when the system wakes up. Khubd will naturally see the status change
while processing the device's parent hub and will do the right thing.
The remote_wakeup() routine is changed; now it expects the caller to
acquire the device lock rather than acquiring the lock itself.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:06:42 +0000 (11:06 -0400)]
USB: simplify hub_restart() logic
This patch (as1081) straightens out the logic of the hub_restart()
routine. Each port of the hub is scanned and the driver makes sure
that ports which are supposed to be disabled really _are_ disabled.
Any ports with a significant change in status are flagged in
hub->change_bits, so that khubd can focus on them without the need to
scan all the ports a second time -- which means the hub->activating
flag is no longer needed.
Also, it is now recognized explicitly that the only reason for
resuming a port which was not suspended is to carry out a reset-resume
operation, which happens only in a non-CONFIG_USB_SUSPEND setting.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
USB: revert "don't use reset-resume if drivers don't support it"
This reverts Linus's previous patch that is in mainline to make it
easier for the USB hub.c patches that follow this to apply cleanly. The
functionality will be added back in a followon patch in this series.
Alan Stern [Mon, 28 Apr 2008 15:06:28 +0000 (11:06 -0400)]
USB: debounce before unregistering
This patch (as1080) makes a significant change to the way khubd
handles port connect-change and enable-change events. Both types of
event are now debounced, and the debouncing is carried out _before_ an
existing usb_device is unregistered, instead of afterward.
This means that drivers will have to deal with longer runs of errors
when a device is unplugged, but they are supposed to be prepared for
that in any case.
The advantage is that when an enable-change occurs (caused for example
by electromagnetic interference), the debouncing period will provide
time for the cause of the problem to die away. A simple port reset
(added in a forthcoming patch) will then allow us to recover from the
fault.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Mon, 28 Apr 2008 15:06:11 +0000 (11:06 -0400)]
USB: add new routine for checking port-resume type
This patch (as1070) creates a new subroutine to check whether a device
can be resumed. This code is needed even when CONFIG_USB_SUSPEND
isn't set, because devices do suspend themselves when the root hub
(and hence the entire bus) is suspended, and power sessions can get
lost during a system sleep even without individual port suspends.
The patch also fixes a loose end in USB-Persist reset-resume handling.
When a low- or full-speed device is attached to an EHCI's companion
controller, the port handoff during resume will cause the companion
port's connect-status-change feature to be set. If that flag isn't
cleared, the port-reset code will think it indicates that the device
has been unplugged and the reset-resume will fail.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
USB: revert "don't lose disconnections during suspend"
This reverts Alan's previous patch so that the recent Hub changes will
apply cleanly. The above mentioned patch was needed for 2.6.26 to work
properly.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
netfilter: nf_conntrack_sctp: fix sparse warnings
netfilter: nf_nat_sip: c= is optional for session
netfilter: xt_TCPMSS: collapse tcpmss_reverse_mtu{4,6} into one function
netfilter: nfnetlink_log: send complete hardware header
netfilter: xt_time: fix time's time_mt()'s use of do_div()
netfilter: accounting rework: ct_extend + 64bit counters (v4)
netlink: add NLA_PUT_BE64 macro
netfilter: nf_nat_core: eliminate useless find_appropriate_src for IP_NAT_RANGE_PROTO_RANDOM
hdlcdrv: Fix CRC calculation.
Revert "pkt_sched: Make default qdisc nonshared-multiqueue safe."
net: In __netif_schedule() use WARN_ON instead of BUG_ON
net: Improve simple_tx_hash().
pkt_sched: Remove unused variable skb in dev_deactivate_queue function.
sunhme: Remove stop/wake TX queue calls in set-multicast-list handler.
ucc_geth: do not touch net queue in adjust_link phylib callback
gianfar: do not touch net queue in adjust_link phylib callback
atl1: Do not wake queue before queue has been started.
Randy Dunlap [Mon, 21 Jul 2008 16:27:18 +0000 (09:27 -0700)]
pci kernel-doc fatal error
Fix kernel-doc comments so that they don't produce errors.
Also cut some extraneous copy-paste text.
Error(linhead//drivers/pci/pci.c:1133): duplicate section name 'Description'
Error(linhead//drivers/pci/pci.c:1189): duplicate section name 'Description'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge branch 'x86/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86/for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (160 commits)
x86: remove extra calling to get ext cpuid level
x86: use setup_clear_cpu_cap() when disabling the lapic
KVM: fix exception entry / build bug, on 64-bit
x86: add unknown_nmi_panic kernel parameter
x86, VisWS: turn into generic arch, eliminate leftover files
x86: add ->pre_time_init to x86_quirks
x86: extend and use x86_quirks to clean up NUMAQ code
x86: introduce x86_quirks
x86: improve debug printout: add target bootmem range in early_res_to_bootmem()
Subject: devmem, x86: fix rename of CONFIG_NONPROMISC_DEVMEM
x86: remove arch_get_ram_range
x86: Add a debugfs interface to dump PAT memtype
x86: Add a arch directory for x86 under debugfs
x86: i386: reduce boot fixmap space
i386/xen: add proper unwind annotations to xen_sysenter_target
x86: reduce force_mwait visibility
x86: reduce forbid_dac's visibility
x86: fix two modpost warnings
x86: check function status in EDD boot code
x86_64: ia32_signal.c: remove signal number conversion
...
Merge branch 'for-linus' of git://neil.brown.name/md
* 'for-linus' of git://neil.brown.name/md: (52 commits)
md: Protect access to mddev->disks list using RCU
md: only count actual openers as access which prevent a 'stop'
md: linear: Make array_size sector-based and rename it to array_sectors.
md: Make mddev->array_size sector-based.
md: Make super_type->rdev_size_change() take sector-based sizes.
md: Fix check for overlapping devices.
md: Tidy up rdev_size_store a bit:
md: Remove some unused macros.
md: Turn rdev->sb_offset into a sector-based quantity.
md: Make calc_dev_sboffset() return a sector count.
md: Replace calc_dev_size() by calc_num_sectors().
md: Make update_size() take the number of sectors.
md: Better control of when do_md_stop is allowed to stop the array.
md: get_disk_info(): Don't convert between signed and unsigned and back.
md: Simplify restart_array().
md: alloc_disk_sb(): Return proper error value.
md: Simplify sb_equal().
md: Simplify uuid_equal().
md: sb_equal(): Fix misleading printk.
md: Fix a typo in the comment to cmd_match().
...
Patrick McHardy [Mon, 21 Jul 2008 17:03:49 +0000 (10:03 -0700)]
netfilter: nf_conntrack_sctp: fix sparse warnings
Introduced by a258860e (netfilter: ctnetlink: add full support for SCTP to ctnetlink):
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: incorrect type in argument 1 (different base types)
net/netfilter/nf_conntrack_proto_sctp.c:483:2: expected unsigned int [unsigned] [usertype] x
net/netfilter/nf_conntrack_proto_sctp.c:483:2: got restricted unsigned int const <noident>
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:483:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: incorrect type in argument 1 (different base types)
net/netfilter/nf_conntrack_proto_sctp.c:487:2: expected unsigned int [unsigned] [usertype] x
net/netfilter/nf_conntrack_proto_sctp.c:487:2: got restricted unsigned int const <noident>
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:487:2: warning: cast from restricted type
net/netfilter/nf_conntrack_proto_sctp.c:532:42: warning: incorrect type in assignment (different base types)
net/netfilter/nf_conntrack_proto_sctp.c:532:42: expected restricted unsigned int <noident>
net/netfilter/nf_conntrack_proto_sctp.c:532:42: got unsigned int
net/netfilter/nf_conntrack_proto_sctp.c:534:39: warning: incorrect type in assignment (different base types)
net/netfilter/nf_conntrack_proto_sctp.c:534:39: expected restricted unsigned int <noident>
net/netfilter/nf_conntrack_proto_sctp.c:534:39: got unsigned int
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Mon, 21 Jul 2008 17:03:23 +0000 (10:03 -0700)]
netfilter: nf_nat_sip: c= is optional for session
According to RFC2327, the connection information is optional
in the session description since it can be specified in the
media description instead.
My provider does exactly that and does not provide any connection
information in the session description. As a result the new
kernel drops all invite responses.
This patch makes it optional as documented.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Jan Engelhardt [Mon, 21 Jul 2008 17:02:59 +0000 (10:02 -0700)]
netfilter: xt_TCPMSS: collapse tcpmss_reverse_mtu{4,6} into one function
Signed-off-by: Jan Engelhardt <jengelh@medozas.de> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds some fields to NFLOG to be able to send the complete
hardware header with all necessary informations.
It sends to userspace:
* the type of hardware link
* the lenght of hardware header
* the hardware header
Signed-off-by: Eric Leblond <eric@inl.fr> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Initially netfilter has had 64bit counters for conntrack-based accounting, but
it was changed in 2.6.14 to save memory. Unfortunately in-kernel 64bit counters are
still required, for example for "connbytes" extension. However, 64bit counters
waste a lot of memory and it was not possible to enable/disable it runtime.
This patch:
- reimplements accounting with respect to the extension infrastructure,
- makes one global version of seq_print_acct() instead of two seq_print_counters(),
- makes it possible to enable it at boot time (for CONFIG_SYSCTL/CONFIG_SYSFS=n),
- makes it possible to enable/disable it at runtime by sysctl or sysfs,
- extends counters from 32bit to 64bit,
- renames ip_conntrack_counter -> nf_conn_counter,
- enables accounting code unconditionally (no longer depends on CONFIG_NF_CT_ACCT),
- set initial accounting enable state based on CONFIG_NF_CT_ACCT
- removes buggy IPCT_COUNTER_FILLING event handling.
If accounting is enabled newly created connections get additional acct extend.
Old connections are not changed as it is not possible to add a ct_extend area
to confirmed conntrack. Accounting is performed for all connections with
acct extend regardless of a current state of "net.netfilter.nf_conntrack_acct".
Signed-off-by: Krzysztof Piotr Oledzki <ole@ans.pl> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Add NLA_PUT_BE64 macro required for 64bit counters in netfilter
Signed-off-by: Krzysztof Piotr Oledzki <ole@ans.pl> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
This is a trivial patch against the hdlcdrv module that fixes its CRC
calculation. The finished CRC was overwriting the first two bytes of
each packet rather than being appended to the end.
I've tested this with 2.6.8 and 2.6.10-rc1, but hdlcdrv hasn't changed
much recently so it should work with many other kernel versions.
Signed-off-by: Micah Dowty <micah@navi.cx> Acked-by: Thomas Sailer <t.sailer@alumni.ethz.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
After discussions with Jamal and Herbert on netdev, we should
provide at least minimal prioritization at the qdisc level
even in multiqueue situations.
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 21 Jul 2008 15:34:49 +0000 (08:34 -0700)]
sunhme: Remove stop/wake TX queue calls in set-multicast-list handler.
Based upon a bug report by Alexander Beregalov and commentary
from Ben Hutchings.
These are totally unnecessary, in particular because this
driver's ->hard_start_xmit() handler takes the same driver
spinlock that the set-multicast-list handler uses.
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 21 Jul 2008 15:28:37 +0000 (08:28 -0700)]
atl1: Do not wake queue before queue has been started.
Based upon a bug report by Alexey Dobriyan, the patch is
also tested by him and confirmed to fix the problem.
Packet flow during link state events should not be done by
waking and stopping the TX queue anyways, that is handled
transparently by netif_carrier_{on,off}().
So, remove the netif_{wake,stop}_queue() calls in the link
check code, and add the necessary netif_start_queue() call
to atl1_up().
Signed-off-by: David S. Miller <davem@davemloft.net>
Milan Broz [Mon, 21 Jul 2008 11:00:37 +0000 (12:00 +0100)]
dm: introduce merge_bvec_fn
Introduce a bvec merge function for device mapper devices
for dynamic size restrictions.
This code ensures the requested biovec lies within a single
target and then calls a target-specific function to check
against any constraints imposed by underlying devices.
Signed-off-by: Milan Broz <mbroz@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Fix a race condition that returns incorrect data when a write causes an
exception to be allocated whilst a read is still in flight.
The race condition happens as follows:
* A read to non-reallocated sector in the snapshot is submitted so that the
read is routed to the original device.
* A write to the original device is submitted. The write causes an exception
that reallocates the block. The write proceeds.
* The original read is dequeued and reads the wrong data.
This race can be triggered with CFQ scheduler and one thread writing and
multiple threads reading simultaneously.
(This patch relies upon the earlier dm-kcopyd-per-device.patch to avoid a
deadlock.)
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Whenever a snapshot read gets mapped through to the origin, track it in
a per-snapshot hash table indexed by chunk number, using memory allocated
from a new per-snapshot mempool.
We need to track these reads to avoid race conditions which will be fixed
by patches that follow.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Adrian Bunk [Mon, 21 Jul 2008 11:00:27 +0000 (12:00 +0100)]
dm log: make dm_dirty_log init and exit static
dm_dirty_log_{init,exit}() can now become static.
Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Alasdair G Kergon <agk@redhat.com>
arch/x86/kvm/built-in.o:(.text.fixup+0x1): relocation truncated to fit: R_X86_64_32 against `.text'
arch/x86/kvm/built-in.o:(.text.fixup+0xb): relocation truncated to fit: R_X86_64_32 against `.text'
arch/x86/kvm/built-in.o:(.text.fixup+0x15): relocation truncated to fit: R_X86_64_32 against `.text'
arch/x86/kvm/built-in.o:(.text.fixup+0x1f): relocation truncated to fit: R_X86_64_32 against `.text'
arch/x86/kvm/built-in.o:(.text.fixup+0x29): relocation truncated to fit: R_X86_64_32 against `.text'
Introduced by commit 4ecac3fd. The problem is that 'push' will default
to 32-bit, which is not wide enough as a fixup address. (and which would
crash on any real fixup event even if it was wide enough)
Introduce KVM_EX_PUSH to get the proper address push width on 64-bit too.
All modifications and most access to the mddev->disks list are made
under the reconfig_mutex lock. However there are three places where
the list is walked without any locking. If a reconfig happens at this
time, havoc (and oops) can ensue.
So use RCU to protect these accesses:
- wrap them in rcu_read_{,un}lock()
- use list_for_each_entry_rcu
- add to the list with list_add_rcu
- delete from the list with list_del_rcu
- delay the 'free' with call_rcu rather than schedule_work
Note that export_rdev did a list_del_init on this list. In almost all
cases the entry was not in the list anymore so it was a no-op and so
safe. It is no longer safe as after list_del_rcu we may not touch
the list_head.
An audit shows that export_rdev is called:
- after unbind_rdev_from_array, in which case the delete has
already been done,
- after bind_rdev_to_array fails, in which case the delete isn't needed.
- before the device has been put on a list at all (e.g. in
add_new_disk where reading the superblock fails).
- and in autorun devices after a failure when the device is on a
different list.
So remove the list_del_init call from export_rdev, and add it back
immediately before the called to export_rdev for that last case.
Note also that ->same_set is sometimes used for lists other than
mddev->list (e.g. candidates). In these cases rcu is not needed.
md: only count actual openers as access which prevent a 'stop'
Open isn't the only thing that increments ->active. e.g. reading
/proc/mdstat will increment it briefly. So to avoid false positives
in testing for concurrent access, introduce a new counter that counts
just the number of times the md device it open.
Andre Noll [Mon, 21 Jul 2008 07:05:22 +0000 (17:05 +1000)]
md: Make mddev->array_size sector-based.
This patch renames the array_size field of struct mddev_s to array_sectors
and converts all instances to use units of 512 byte sectors instead of 1k
blocks.
Signed-off-by: Andre Noll <maan@systemlinux.org> Signed-off-by: NeilBrown <neilb@suse.de>
Andre Noll [Mon, 21 Jul 2008 04:42:07 +0000 (14:42 +1000)]
md: Fix check for overlapping devices.
The checks in overlaps() expect all parameters either in block-based
or sector-based quantities. However, its single caller passes two
rdev->data_offset arguments as well as two rdev->size arguments, the
former being sector counts while the latter are measured in 1K blocks.
This could cause rdev_size_store() to accept an invalid size from user
space. Fix it by passing only sector-based quantities to overlaps().
Signed-off-by: Andre Noll <maan@systemlinux.org> Signed-off-by: NeilBrown <neilb@suse.de>
Neil Brown [Sat, 12 Jul 2008 00:37:50 +0000 (10:37 +1000)]
md: Tidy up rdev_size_store a bit:
- used strict_strtoull in place of simple_strtoull
- use my_mddev in place of rdev->mddev (they have the same value)
and more significantly,
- don't adjust mddev->size to fit, rather reject changes which make
rdev->size smaller than mddev->size
Adjusting mddev->size is a hangover from bind_rdev_to_array which
does a similar thing. But it really is a better design to insist that
mddev->size is set as required, then the rdev->sizes are set to allow
for that. The previous way invites confusion.
Merge branch 'for-2.6.27' of git://linux-nfs.org/~bfields/linux
* 'for-2.6.27' of git://linux-nfs.org/~bfields/linux: (51 commits)
nfsd: nfs4xdr.c do-while is not a compound statement
nfsd: Use C99 initializers in fs/nfsd/nfs4xdr.c
lockd: Pass "struct sockaddr *" to new failover-by-IP function
lockd: get host reference in nlmsvc_create_block() instead of callers
lockd: minor svclock.c style fixes
lockd: eliminate duplicate nlmsvc_lookup_host call from nlmsvc_lock
lockd: eliminate duplicate nlmsvc_lookup_host call from nlmsvc_testlock
lockd: nlm_release_host() checks for NULL, caller needn't
file lock: reorder struct file_lock to save space on 64 bit builds
nfsd: take file and mnt write in nfs4_upgrade_open
nfsd: document open share bit tracking
nfsd: tabulate nfs4 xdr encoding functions
nfsd: dprint operation names
svcrdma: Change WR context get/put to use the kmem cache
svcrdma: Create a kmem cache for the WR contexts
svcrdma: Add flush_scheduled_work to module exit function
svcrdma: Limit ORD based on client's advertised IRD
svcrdma: Remove unused wait q from svcrdma_xprt structure
svcrdma: Remove unneeded spin locks from __svc_rdma_free
svcrdma: Add dma map count and WARN_ON
...
projects / linux-2.6 / log