]> bbs.cooldavid.org Git - net-next-2.6.git/log
net-next-2.6.git
16 years ago[RTNETLINK]: Fix warning for !CONFIG_KMOD
Thomas Graf [Tue, 31 Jul 2007 21:13:50 +0000 (14:13 -0700)]
[RTNETLINK]: Fix warning for !CONFIG_KMOD

replay label is unused otherwise.

Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV4] ip_options.c: kmalloc + memset conversion to kzalloc
Mariusz Kozlowski [Tue, 31 Jul 2007 21:06:45 +0000 (14:06 -0700)]
[IPV4] ip_options.c: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[DECNET]: kmalloc + memset conversion to kzalloc
Mariusz Kozlowski [Tue, 31 Jul 2007 21:05:56 +0000 (14:05 -0700)]
[DECNET]: kmalloc + memset conversion to kzalloc

Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: ethtool_perm_addr only has one implementation
Matthew Wilcox [Tue, 31 Jul 2007 21:00:29 +0000 (14:00 -0700)]
[NET]: ethtool_perm_addr only has one implementation

All drivers implement ethtool get_perm_addr the same way -- by calling
the generic function.  So we can inline the generic function into the
caller and avoid going through the drivers.

Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: ethtool ops are the only way
Matthew Wilcox [Tue, 31 Jul 2007 21:00:02 +0000 (14:00 -0700)]
[NET]: ethtool ops are the only way

During the transition to the ethtool_ops way of doing things, we supported
calling the device's ->do_ioctl method to allow unconverted drivers to
continue working.  Those days are long behind us, all in-tree drivers
use the ethtool_ops way, and so we no longer need to support this.

The bonding driver is the biggest beneficiary of this; it no longer
needs to call ioctl() as a fallback if ethtool_ops aren't supported.

Also put a proper copyright statement on ethtool.c.

Signed-off-by: Matthew Wilcox <matthew@wil.cx>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PPPOE]: Improve hashing function in hash_item().
Florian Zumbiehl [Tue, 31 Jul 2007 20:47:57 +0000 (13:47 -0700)]
[PPPOE]: Improve hashing function in hash_item().

The new code produces the same results as the old version and is
~ 3 to 6 times faster for 4-bit hashes on the CPUs I tested.

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[XFRM]: State selection update to use inner addresses.
Joakim Koskela [Thu, 26 Jul 2007 07:08:42 +0000 (00:08 -0700)]
[XFRM]: State selection update to use inner addresses.

This patch modifies the xfrm state selection logic to use the inner
addresses where the outer have been (incorrectly) used. This is
required for beet mode in general and interfamily setups in both
tunnel and beet mode.

Signed-off-by: Joakim Koskela <jookos@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Diego Beltrami <diego.beltrami@gmail.com>
Signed-off-by: Miika Komu <miika@iki.fi>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPSEC]: Ensure that state inner family is set
Herbert Xu [Tue, 31 Jul 2007 09:04:32 +0000 (02:04 -0700)]
[IPSEC]: Ensure that state inner family is set

Similar to the issue we had with template families which
specified the inner families of policies, we need to set
the inner families of states as the main xfrm user Openswan
leaves it as zero.

af_key is unaffected because the inner family is set by it
and not the KM.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TCP]: Bidir flow must not disregard SACK blocks for lost marking
Ilpo Järvinen [Tue, 31 Jul 2007 02:51:12 +0000 (19:51 -0700)]
[TCP]: Bidir flow must not disregard SACK blocks for lost marking

It's possible that new SACK blocks that should trigger new LOST
markings arrive with new data (which previously made is_dupack
false). In addition, I think this fixes a case where we get
a cumulative ACK with enough SACK blocks to trigger the fast
recovery (is_dupack would be false there too).

I'm not completely pleased with this solution because readability
of the code is somewhat questionable as 'is_dupack' in SACK case
is no longer about dupacks only but would mean something like
'lost_marker_work_todo' too... But because of Eifel stuff done
in CA_Recovery, the FLAG_DATA_SACKED check cannot be placed to
the if statement which seems attractive solution. Nevertheless,
I didn't like adding another variable just for that either... :-)

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TCP]: Fix ratehalving with bidirectional flows
Ilpo Järvinen [Tue, 31 Jul 2007 02:48:37 +0000 (19:48 -0700)]
[TCP]: Fix ratehalving with bidirectional flows

Actually, the ratehalving seems to work too well, as cwnd is
reduced on every second ACK even though the packets in flight
remains unchanged. Recoveries in a bidirectional flows suffer
quite badly because of this, both NewReno and SACK are affected.

After this patch, rate halving is performed for ACK only if
packets in flight was supposedly changed too.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PPPOL2TP]: Add CONFIG_INET Kconfig dependency.
James Chapman [Tue, 31 Jul 2007 01:51:48 +0000 (18:51 -0700)]
[PPPOL2TP]: Add CONFIG_INET Kconfig dependency.

PPPOL2TP uses UDP so it obviously depends on CONFIG_INET.

Signed-off-by: James Chapman <jchapman@katalix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Page offsets and lengths need to be __u32.
David S. Miller [Tue, 31 Jul 2007 01:47:03 +0000 (18:47 -0700)]
[NET]: Page offsets and lengths need to be __u32.

Based upon a report from Stephen Rothwell.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[AF_UNIX]: Make code static.
Adrian Bunk [Tue, 31 Jul 2007 01:05:45 +0000 (18:05 -0700)]
[AF_UNIX]: Make code static.

The following code can now become static:
- struct unix_socket_table
- unix_table_lock

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NETFILTER]: Make nf_ct_ipv6_skip_exthdr() static.
Adrian Bunk [Tue, 31 Jul 2007 01:04:57 +0000 (18:04 -0700)]
[NETFILTER]: Make nf_ct_ipv6_skip_exthdr() static.

nf_ct_ipv6_skip_exthdr() can now become static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PKTGEN]: make get_ipsec_sa() static and non-inline
Adrian Bunk [Tue, 31 Jul 2007 01:04:09 +0000 (18:04 -0700)]
[PKTGEN]: make get_ipsec_sa() static and non-inline

Non-static inline code usually doesn't makes sense.

In this case making is static and non-inline is the correct solution.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PPPoE]: move lock_sock() in pppoe_sendmsg() to the right location
Florian Zumbiehl [Tue, 31 Jul 2007 00:49:13 +0000 (17:49 -0700)]
[PPPoE]: move lock_sock() in pppoe_sendmsg() to the right location

and the last one for now: Acquire the sock lock in pppoe_sendmsg()
before accessing the sock - and in particular avoid releasing the lock
even though it hasn't been acquired.

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PPPoX/E]: return ENOTTY on unknown ioctl requests
Florian Zumbiehl [Tue, 31 Jul 2007 00:48:23 +0000 (17:48 -0700)]
[PPPoX/E]: return ENOTTY on unknown ioctl requests

here another patch for the PPPoX/E code that makes sure that ENOTTY is
returned for unknown ioctl requests rather than 0 (and removes another
unneeded initializer which I didn't bother creating a separate patch for).

Signed-off-by: Florian Zumbiehl <florz@florz.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV6]: ipv6_addr_type() doesn't know about RFC4193 addresses.
Dave Johnson [Tue, 31 Jul 2007 00:19:31 +0000 (17:19 -0700)]
[IPV6]: ipv6_addr_type() doesn't know about RFC4193 addresses.

ipv6_addr_type() doesn't check for 'Unique Local IPv6 Unicast
Addresses' (RFC4193) and returns IPV6_ADDR_RESERVED for that range.

SCTP uses this function and will fail bind() and connect() calls that
use RFC4193 addresses, SCTP will also ignore inbound connections from
RFC4193 addresses if listening on IPV6_ADDR_ANY.

There may be other users of ipv6_addr_type() that could also have
problems.

Signed-off-by: Dave Johnson <djohnson@sw.starentnetworks.com>
Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Fix prio_tune() handling of root qdisc.
Peter P Waskiewicz Jr [Tue, 31 Jul 2007 00:13:45 +0000 (17:13 -0700)]
[NET]: Fix prio_tune() handling of root qdisc.

Fix the check in prio_tune() to see if sch->parent is TC_H_ROOT instead of
sch->handle to load or reject the qdisc for multiqueue devices.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Fix sch_api to properly set sch->parent on the root.
Patrick McHardy [Tue, 31 Jul 2007 00:11:50 +0000 (17:11 -0700)]
[NET]: Fix sch_api to properly set sch->parent on the root.

Fix sch_api to correctly set sch->parent for both ingress and egress
qdiscs in qdisc_create().

Signed-off-by: Patrick McHardy <trash@kaber.net>
Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET_SCHED]: Fix prio/ingress classification logic error
Patrick McHardy [Tue, 31 Jul 2007 00:07:14 +0000 (17:07 -0700)]
[NET_SCHED]: Fix prio/ingress classification logic error

Fix handling of empty or completely non-matching filter chains. In
that case -1 is returned and tcf_result is uninitialized, the
qdisc should fall back to default classification in that case.

Noticed by PJ Waskiewicz <peter.p.waskiewicz.jr@intel.com>.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV6]: Remove circular dependency on if_inet6.h
Herbert Xu [Tue, 31 Jul 2007 00:05:49 +0000 (17:05 -0700)]
[IPV6]: Remove circular dependency on if_inet6.h

net/if_inet6.h includes linux/ipv6.h which also tries to include
net/if_inet6.h.  Since the latter only needs it for forward
declarations, we can fix this by adding the declarations.

A number of files are implicitly including net/if_inet6.h through
linux/ipv6.h.  They also use net/ipv6.h so this patch includes
net/if_inet6.h there.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV4/IPV6]: Fail registration if inet device construction fails
Herbert Xu [Tue, 31 Jul 2007 00:04:52 +0000 (17:04 -0700)]
[IPV4/IPV6]: Fail registration if inet device construction fails

Now that netdev notifications can fail, we can use this to signal
errors during registration for IPv4/IPv6.  In particular, if we
fail to allocate memory for the inet device, we can fail the netdev
registration.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Allow netdev REGISTER/CHANGENAME events to fail
Herbert Xu [Tue, 31 Jul 2007 00:03:38 +0000 (17:03 -0700)]
[NET]: Allow netdev REGISTER/CHANGENAME events to fail

This patch adds code to allow errors to be passed up from event
handlers of NETDEV_REGISTER and NETDEV_CHANGENAME.  It also adds
the notifier_from_errno/notifier_to_errnor helpers to pass the
errno value up to the notifier caller.

If an error is detected when a device is registered, it causes
that operation to fail.  A NETDEV_UNREGISTER will be sent to
all event handlers.

Similarly if NETDEV_CHANGENAME fails the original name is restored
and a new NETDEV_CHANGENAME event is sent.

As such all event handlers must be idempotent with respect to
these events.

When an event handler is registered NETDEV_REGISTER events are
sent for all devices currently registered.  Should any of them
fail, we will send NETDEV_GOING_DOWN/NETDEV_DOWN/NETDEV_UNREGISTER
events to that handler for the devices which have already been
registered with it.  The handler registration itself will fail.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET] loopback: Panic if registration fails
Herbert Xu [Mon, 30 Jul 2007 23:37:19 +0000 (16:37 -0700)]
[NET] loopback: Panic if registration fails

Because IPv4 and IPv6 both depend on the presence of the loopback
device to function, failure in registration the loopback device should
be fatal.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Take dev_base_lock when moving device name hash list entry
Herbert Xu [Mon, 30 Jul 2007 23:35:46 +0000 (16:35 -0700)]
[NET]: Take dev_base_lock when moving device name hash list entry

When we added name-based hashing the dev_base_lock was designated as the
lock to take when changing the name hash list.  Unfortunately, because
it was a preexisting lock that just happened to be taken in the right
spots we neglected to take it in dev_change_name.

The race can affect calles of __dev_get_by_name that do so without taking
the RTNL.  They may end up walking down the wrong hash chain and end up
missing the device that they're looking for.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: Call uninit if necessary in register_netdevice
Herbert Xu [Mon, 30 Jul 2007 23:29:40 +0000 (16:29 -0700)]
[NET]: Call uninit if necessary in register_netdevice

This patch makes register_netdevice call dev->uninit if the regsitration
fails after dev->init has completed successfully.  Very few drivers use
the init/uninit calls but at least one (drivers/net/wan/sealevel.c) may
leak without this change.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPVS]: Use skb_forward_csum
Herbert Xu [Mon, 30 Jul 2007 23:20:12 +0000 (16:20 -0700)]
[IPVS]: Use skb_forward_csum

As a path that forwards packets, IPVS should be using
skb_forward_csum instead of directly setting ip_summed
to CHECKSUM_NONE.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[PKTGEN]: Add missing KERN_* tags to printk()s.
David S. Miller [Mon, 30 Jul 2007 23:11:48 +0000 (16:11 -0700)]
[PKTGEN]: Add missing KERN_* tags to printk()s.

Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[BLUETOOTH] l2cap: don't mangle cmd.len
Al Viro [Sun, 29 Jul 2007 07:18:23 +0000 (00:18 -0700)]
[BLUETOOTH] l2cap: don't mangle cmd.len

Since nobody uses it after we convert it to host-endian,
no need to do that at all.  At that point l2cap is endian-clean.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[BLUETOOTH]: pass (host-endian) cmd length as explicit argument to l2cap_conf_req()
Al Viro [Sun, 29 Jul 2007 07:17:25 +0000 (00:17 -0700)]
[BLUETOOTH]: pass (host-endian) cmd length as explicit argument to l2cap_conf_req()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[BLUETOOTH] l2cap: endianness annotations
Al Viro [Sun, 29 Jul 2007 07:16:36 +0000 (00:16 -0700)]
[BLUETOOTH] l2cap: endianness annotations

no code changes, just documenting existing types

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[BLUETOOTH]: Fix endianness bug in l2cap_sock_listen()
Al Viro [Sun, 29 Jul 2007 07:15:18 +0000 (00:15 -0700)]
[BLUETOOTH]: Fix endianness bug in l2cap_sock_listen()

We loop through psm values, calling __l2cap_get_sock_by_addr(psm, ...)
until we get NULL; then we set ->psm of our socket to htobs(psm).
IOW, we find unused psm value and put it into our socket.  So far, so
good, but...  __l2cap_get_sock_by_addr() compares its argument with
->psm of sockets.  IOW, the entire thing works correctly only on
little-endian.  On big-endian we'll get "no socket with such psm"
on the first iteration, since we won't find a socket with ->psm == 0x1001.
We will happily conclude that 0x1001 is unused and slap htobs(0x1001)
(i.e. 0x110) into ->psm of our socket.  Of course, the next time around
the same thing will repeat and we'll just get a fsckload of sockets
with the same ->psm assigned.

Fix: pass htobs(psm) to __l2cap_get_sock_by_addr() there.  All other
callers are already passing little-endian values and all places that
store something in ->psm are storing little-endian.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IRDA]: Fix rfcomm use-after-free
Marcel Holtmann [Thu, 26 Jul 2007 07:12:25 +0000 (00:12 -0700)]
[IRDA]: Fix rfcomm use-after-free

Adrian Bunk wrote:
> Commit 8de0a15483b357d0f0b821330ec84d1660cadc4e added the following
> use-after-free in net/bluetooth/rfcomm/tty.c:
>
> <--  snip  -->
>
> ...
> static int rfcomm_dev_add(struct rfcomm_dev_req *req, struct rfcomm_dlc *dlc)
> {
> ...
>         if (IS_ERR(dev->tty_dev)) {
>                 list_del(&dev->list);
>                 kfree(dev);
>                 return PTR_ERR(dev->tty_dev);
>         }
> ...
>
> <--  snip  -->
>
> Spotted by the Coverity checker.

really good catch. I fully overlooked that one. The attached patch
should fix it.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[IPV6]: Don't update ADVMSS on routes where the MTU is not also updated
Simon Arlott [Thu, 26 Jul 2007 07:09:55 +0000 (00:09 -0700)]
[IPV6]: Don't update ADVMSS on routes where the MTU is not also updated

The ADVMSS value was incorrectly updated for ALL routes when the MTU
is updated because it's outside the effect of the if statement's
condition.

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET_DMA]: remove unused dma_memcpy_to_kernel_iovec
Shannon Nelson [Thu, 26 Jul 2007 07:05:53 +0000 (00:05 -0700)]
[NET_DMA]: remove unused dma_memcpy_to_kernel_iovec

Al Viro pointed out that dma_memcpy_to_kernel_iovec() really was
unreachable and thus unused.  The code originally was there to support
in-kernel dma needs, but since it remains unused, we'll pull it out.

Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TIPC]: fix tipc_link_create error handling
Florian Westphal [Thu, 26 Jul 2007 07:05:07 +0000 (00:05 -0700)]
[TIPC]: fix tipc_link_create error handling

if printbuf allocation or tipc_node_attach_link() fails, invalid
references to the link are left in the associated node and bearer
structures.
Fix by allocating printbuf early and moving timer initialization
and the addition of the new link to the b_ptr->links list after
tipc_node_attach_link() succeeded.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[NET]: kernel-doc fixes
Randy Dunlap [Thu, 26 Jul 2007 07:03:29 +0000 (00:03 -0700)]
[NET]: kernel-doc fixes

Fix kernel-doc omissions in net/:

Warning(linux-2.6.23-rc1//net/core/dev.c:2728): No description found for parameter 'addr'
Warning(linux-2.6.23-rc1//net/core/dev.c:2752): No description found for parameter 'addr'
Warning(linux-2.6.23-rc1//net/core/dev.c:3839): No description found for parameter 'net_dma'
Warning(linux-2.6.23-rc1//net/core/dev.c:3877): No description found for parameter 'state'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TCP]: htcp - use measured rtt
Stephen Hemminger [Thu, 26 Jul 2007 06:50:28 +0000 (23:50 -0700)]
[TCP]: htcp - use measured rtt

Change HTCP to use measured RTT rather than smooth RTT.
Srtt is computed using the TCP receive timestamp
options, so it is vulnerable to hostile receivers. To avoid any problems
this might cause use the measured RTT instead.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TCP]: cubic - eliminate use of receive time stamp
Stephen Hemminger [Thu, 26 Jul 2007 06:50:06 +0000 (23:50 -0700)]
[TCP]: cubic - eliminate use of receive time stamp

Remove use of received timestamp option value from RTT calculation in Cubic.
A hostile receiver may be returning a larger timestamp option than the original
value. This would cause the sender to believe the malevolent receiver had
a larger RTT and because Cubic tries to provide some RTT friendliness, the
sender would then favor the liar.

Instead, use the jiffie resolutionRTT value already computed and
passed back after ack.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years ago[TCP]: congestion control API pass RTT in microseconds
Stephen Hemminger [Thu, 26 Jul 2007 06:49:34 +0000 (23:49 -0700)]
[TCP]: congestion control API pass RTT in microseconds

This patch changes the API for the callback that is done after an ACK is
received. It solves a couple of issues:

  * Some congestion controls want higher resolution value of RTT
    (controlled by TCP_CONG_RTT_SAMPLE flag). These don't really want a ktime, but
    all compute a RTT in microseconds.

  * Other congestion control could use RTT at jiffies resolution.

To keep API consistent the units should be the same for both cases, just the
resolution should change.

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
16 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6.23
Linus Torvalds [Tue, 31 Jul 2007 04:54:37 +0000 (21:54 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6.23

* master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6.23:
  sh: Fix fs.h removal from mm.h regressions.
  sh: fix get_wchan() for SH kernels without framepointers
  sh: arch/sh/boot - fix shell usage
  rtc: rtc-sh: Correct sh_rtc_set_time() for some SH-3 parts.
  sh: remove support for sh7300 and solution engine 7300
  sh: Add sh to the CC_OPTIMIZE_FOR_SIZE dependencies.
  sh: Kill off virt_to_bus()/bus_to_virt().
  sh: sh-sci - fix SH7708 support
  sh: Restrict DSP support to specific CPUs.
  sh: Silence sq compile warning on sh4 nommu.
  sh: Kill the rest of the SE73180 cruft.
  sh: remove support for sh73180 and solution engine 73180
  sh: remove old broken pint code
  sh: Reclaim beginning of P3 space for vmalloc area.
  sh: Fix Dreamcast DMA issues.
  sh: Add kmap_coherent()/kunmap_coherent() interface for SH-4.

16 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh64-2.6
Linus Torvalds [Tue, 31 Jul 2007 04:54:16 +0000 (21:54 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh64-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh64-2.6:
  sh64: Kill off virt_to_bus()/bus_to_virt().
  sh64: Fix irq_intc build failure.
  sh64: Fix fs.h removal from mm.h regressions.

16 years agosh64: Kill off virt_to_bus()/bus_to_virt().
Paul Mundt [Tue, 31 Jul 2007 04:11:25 +0000 (13:11 +0900)]
sh64: Kill off virt_to_bus()/bus_to_virt().

Follows the SH change.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
16 years agosh64: Fix irq_intc build failure.
Paul Mundt [Tue, 31 Jul 2007 04:07:37 +0000 (13:07 +0900)]
sh64: Fix irq_intc build failure.

Needs interrupt.h:

  CC      arch/sh64/kernel/irq_intc.o
arch/sh64/kernel/irq_intc.c: In function 'make_intc_irq':
arch/sh64/kernel/irq_intc.c:179: error: implicit declaration of function 'disable_irq_nosync'
make[1]: *** [arch/sh64/kernel/irq_intc.o] Error 1

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
16 years agosh64: Fix fs.h removal from mm.h regressions.
Paul Mundt [Tue, 31 Jul 2007 04:03:02 +0000 (13:03 +0900)]
sh64: Fix fs.h removal from mm.h regressions.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
16 years agosh: Fix fs.h removal from mm.h regressions.
Paul Mundt [Tue, 31 Jul 2007 04:01:43 +0000 (13:01 +0900)]
sh: Fix fs.h removal from mm.h regressions.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
16 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Mon, 30 Jul 2007 23:38:05 +0000 (16:38 -0700)]
Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  Fix a potential NULL pointer dereference in mace_interrupt() in drivers/net/pcmcia/nmclan_cs.c
  PATCH kernel 2.6.22] PCMCIA-NETDEV : modify smc91c92_cs.c to become SMP safe
  S2io: Increment received packet count correctly
  S2io: Fix crash when resetting adapter
  S2io: Mask spurious interrupts
  S2IO: Implementing review comments from old patches
  S2IO: Checking for the return value of pci map function
  S2IO: Removing MSI support from driver
  S2IO: Removing 3 buffer mode support from the driver
  netxen: drop redudant spinlock
  netxen: Fix interrupt handling for multiport adapters
  netxen: re-init station address after h/w init
  tulip: Remove tulip maintainer
  forcedeth: mac address correct
  gfar: Fix modpost warning
  lib8390: comment on locking by Alan Cox
  Fix a potential NULL pointer dereference in write_bulk_callback() in drivers/net/usb/pegasus.c

16 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Mon, 30 Jul 2007 23:36:33 +0000 (16:36 -0700)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/ipath: Workaround problem of errormask register being overwritten
  IB/ipath: Fix some issues with buffer cancel and sendctrl register update
  IB/ipath: Use faster put_tid_2 routine after initialization
  IB/ipath: Remove unsafe fastrcvint code from interrupt handler
  IB/ehca: Move extern declarations from .c files to .h files
  IB/mlx4: Whitespace fix
  IB/ehca: Fix include order to better match kernel style
  mlx4_core: Remove kfree() in mlx4_mr_alloc() error flow
  RDMA/amso1100: Initialize the wait_queue_head_t in the c2_qp structure

16 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6
Linus Torvalds [Mon, 30 Jul 2007 23:35:35 +0000 (16:35 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6:
  modules: better error messages when modules fail to load due to a sysfs problem.
  kobject: update documentation
  kset: kernel-doc cleanups
  driver core: revert "device" link creation check
  stable_api_nonsense.txt: Disambiguate the use of "this" by using "that" to refer to the syscall interface
  Fix Doc/sysfs-rules typos
  kernel-doc fixes for PCI and drivers/base/
  kobject: put kobject_actions in kobject.h
  kobject: fix link error when CONFIG_HOTPLUG is disabled
  HOWTO: sync Japanese HOWTO
  HOWTO: adjust translation header of Japanese stable_api_nonsense.txt

16 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Mon, 30 Jul 2007 23:34:00 +0000 (16:34 -0700)]
Merge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6:
  USB: "sparse" cleanups for usb gadgets
  usb-serial: Fix edgeport regression on non-EPiC devices
  USB: more pxa2xx_udc dead code removal
  USB: NIKON D50 is an unusual device
  USB: drivers/usb/serial/sierra.c: make 3 functions static
  USB: fix BUG: sleeping function called from invalid context at /home/jeremy/hg/xen/paravirt/linux/drivers/usb/core/urb.c:524, in_atomic():1, irqs_disabled():0
  USB: mct_u232: Convert to proper speed handling API
  digi_acceleport: Drag the driver kicking and screaming into coding style
  cp2101: Remove broken termios optimisation, use proper speed API
  USB: Fix a bug in usb_start_wait_urb
  USB: fix scatterlist PIO case (IOMMU)
  USB: fix usb_serial_suspend(): buggy code
  USB: yet another quirky device
  USB: Add CanonScan LiDE30 to the quirk list
  USB: even more quirks
  USB: usb.h kernel-doc additions
  USB: more quirky devices
  USB: Don't let usb-storage steal Blackberry Pearl
  USB: devices misc: Trivial patch to build the IOWARRIOR when it is selected in Kconfig

16 years agomodules: better error messages when modules fail to load due to a sysfs problem.
Greg Kroah-Hartman [Mon, 30 Jul 2007 18:26:38 +0000 (11:26 -0700)]
modules: better error messages when modules fail to load due to a sysfs problem.

This helps people when debugging problems like the ones that were in the
recent -mm releases.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agokobject: update documentation
Cornelia Huck [Fri, 27 Jul 2007 11:41:10 +0000 (13:41 +0200)]
kobject: update documentation

Update kobject documentation:

- Update structure definitions.
- Remove documentation of removed struct subsystem.

(First shot, uevent_ops probably need some documentation as well.)

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agokset: kernel-doc cleanups
Randy Dunlap [Fri, 27 Jul 2007 16:36:26 +0000 (09:36 -0700)]
kset: kernel-doc cleanups

Removed kernel-doc marker (/**) from struct kset -- each struct member
still needs annotation.

Corrected one parameter name so that kernel-doc matches the macro.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agodriver core: revert "device" link creation check
Cornelia Huck [Wed, 25 Jul 2007 07:58:08 +0000 (09:58 +0200)]
driver core: revert "device" link creation check

driver core: revert "device" link creation check

Commit 2ee97caf0a6602f749ddbfdb1449e383e1212707 introduced an extra
check on when to create the "device" symlink. Unfortunately, this
breaks input, so let's revert to the old behaviour.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Acked-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agostable_api_nonsense.txt: Disambiguate the use of "this" by using "that" to refer...
Juan Lang [Tue, 24 Jul 2007 20:24:19 +0000 (13:24 -0700)]
stable_api_nonsense.txt: Disambiguate the use of "this" by using "that" to refer to the syscall interface

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoFix Doc/sysfs-rules typos
Randy Dunlap [Tue, 24 Jul 2007 04:05:02 +0000 (21:05 -0700)]
Fix Doc/sysfs-rules typos

Fix typos only (spelling, grammar, duplicate words, etc.).

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Kay Sievers <kay.sievers@vrfy.org
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agokernel-doc fixes for PCI and drivers/base/
Randy Dunlap [Tue, 24 Jul 2007 04:42:11 +0000 (21:42 -0700)]
kernel-doc fixes for PCI and drivers/base/

Fix undocumented function parameters in PCI and drivers/base.

Warning(linux-2.6.23-rc1//drivers/pci/pci.c:1526): No description found for parameter 'rq'
Warning(linux-2.6.23-rc1//drivers/base/firmware_class.c:245): No description found for parameter 'bin_attr'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agokobject: put kobject_actions in kobject.h
Greg Kroah-Hartman [Tue, 9 Apr 2002 19:14:34 +0000 (12:14 -0700)]
kobject: put kobject_actions in kobject.h

This prevents the extern declaration in the driver core.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agokobject: fix link error when CONFIG_HOTPLUG is disabled
Cornelia Huck [Fri, 20 Jul 2007 11:58:13 +0000 (13:58 +0200)]
kobject: fix link error when CONFIG_HOTPLUG is disabled

Leaving kobject_actions[] in kobject_uevent.c, but putting it outside
the #ifdef looks indeed like the best solution to me. This way, we
avoid adding #ifdef CONFIG_HOTPLUG into core.c, when all other
functions called do not need such a thing.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoHOWTO: sync Japanese HOWTO
Tsugikazu Shibata [Thu, 19 Jul 2007 02:24:54 +0000 (11:24 +0900)]
HOWTO: sync Japanese HOWTO

Signed-off-by: Tsugikazu Shibata <tshibata@ab.jp.nec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoHOWTO: adjust translation header of Japanese stable_api_nonsense.txt
IKEDA, Munehiro [Thu, 19 Jul 2007 02:36:56 +0000 (11:36 +0900)]
HOWTO: adjust translation header of Japanese stable_api_nonsense.txt

Signed-off-by: IKEDA, Munehiro <m-ikeda@ds.jp.nec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: "sparse" cleanups for usb gadgets
David Brownell [Mon, 30 Jul 2007 19:31:07 +0000 (12:31 -0700)]
USB: "sparse" cleanups for usb gadgets

This removes complaints about the gadget stack which are generated by
the currrent "sparse":  it doesn't like the fact that zero is the null
pointer.  (Last I checked, C guarantees that's correct ...)

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agousb-serial: Fix edgeport regression on non-EPiC devices
Adam Kropelin [Sun, 29 Jul 2007 15:03:29 +0000 (11:03 -0400)]
usb-serial: Fix edgeport regression on non-EPiC devices

Fix serious regression on non-EPiC edgeport usb-serial devices. Baud
rate and MCR/LCR registers are not being written on these models due
to apparent copy-n-paste errors introduced with EPiC support.

Failure reported by Nick Pasich <Nick@NickAndBarb.net>.

Signed-off-by: Adam Kropelin <akropel1@rochester.rr.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: more pxa2xx_udc dead code removal
David Brownell [Sat, 21 Jul 2007 18:43:35 +0000 (11:43 -0700)]
USB: more pxa2xx_udc dead code removal

Remove some more dead code from the pxa2xx_udc driver:  support
for a no-longer-undocumented hardware "test mode".  Newer chips
made this the default, evidently as the best workaround for deep
silicon bugs.  The interest was that this seemed to be the only
way to kick in the (documented!) double buffering capability.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: NIKON D50 is an unusual device
Milinevsky Dmitry [Fri, 20 Jul 2007 23:58:53 +0000 (16:58 -0700)]
USB: NIKON D50 is an unusual device

This short patch allows NIKON D50 to be mounted as UMS[unusual device]
on Linux niam 2.6.22-rc7-cfs-v18 #2 PREEMPT Tue Jul 3 22:35:53 EEST
2007 i686 Intel(R) Celeron(R) M processor 1.50GHz GenuineIntel
GNU/Linux,
some previous kernels...

lsusb -v
Bus 001 Device 006: ID 04b0:0409 Nikon Corp.
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               2.00
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0
  bDeviceProtocol         0
  bMaxPacketSize0        64
  idVendor           0x04b0 Nikon Corp.
  idProduct          0x0409
  bcdDevice            1.00
  iManufacturer           1 NIKON
  iProduct                2 NIKON DSC D50
  iSerial                 0
  bNumConfigurations      1
  Configuration Descriptor:
    bLength                 9
    bDescriptorType         2
    wTotalLength           32
    bNumInterfaces          1
    bConfigurationValue     1
    iConfiguration          0
    bmAttributes         0xc0
      Self Powered
    MaxPower                2mA
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        0
      bAlternateSetting       0
      bNumEndpoints           2
      bInterfaceClass         8 Mass Storage
      bInterfaceSubClass      6 SCSI
      bInterfaceProtocol     80 Bulk (Zip)
      iInterface              0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x01  EP 1 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x82  EP 2 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0200  1x 512 bytes
        bInterval               0
Device Qualifier (for other device speed):
  bLength                10
  bDescriptorType         6
  bcdUSB               2.00
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0
  bDeviceProtocol         0
  bMaxPacketSize0        64
  bNumConfigurations      1
Device Status:     0x0001
  Self Powered

Signed-off-by: Milinevsky Dmitry <niam.niam@gmail.com>
Cc: Oliver Neukum <oliver@neukum.org>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: <caglar@pardus.org.tr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: drivers/usb/serial/sierra.c: make 3 functions static
Adrian Bunk [Sun, 29 Jul 2007 14:59:02 +0000 (16:59 +0200)]
USB: drivers/usb/serial/sierra.c: make 3 functions static

This patch makes three needlessly global functions static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Kevin Lloyd <linux@sierrawireless.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix BUG: sleeping function called from invalid context at /home/jeremy/hg/xen...
Oliver Neukum [Tue, 24 Jul 2007 13:13:42 +0000 (15:13 +0200)]
USB: fix BUG: sleeping function called from invalid context at /home/jeremy/hg/xen/paravirt/linux/drivers/usb/core/urb.c:524, in_atomic():1, irqs_disabled():0

Clearly there's a bug in
drivers/usb/serial/usb-serial.c:usb_serial_put().  It shouldn't call
kref_put() while holding a spinlock.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: mct_u232: Convert to proper speed handling API
Alan Cox [Thu, 26 Jul 2007 18:01:10 +0000 (19:01 +0100)]
USB: mct_u232: Convert to proper speed handling API

Signed-off-by: Alan Cox <alan@redhat.com>
Acked-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agodigi_acceleport: Drag the driver kicking and screaming into coding style
Alan Cox [Thu, 26 Jul 2007 17:57:52 +0000 (18:57 +0100)]
digi_acceleport: Drag the driver kicking and screaming into coding style

- The outbreak of acute bracketitus has been cured
- The belief that brackets should have spaces everywhere likewise
- Various other coding style tweaks
- Use baud rates not Bfoo in the speed setup switch

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agocp2101: Remove broken termios optimisation, use proper speed API
Alan Cox [Thu, 26 Jul 2007 17:54:12 +0000 (18:54 +0100)]
cp2101: Remove broken termios optimisation, use proper speed API

I've also enabled the commented out support for 7200, 14400, 55854,
127117 and 3686400 baud as you can now set such rates in the kernel.

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Fix a bug in usb_start_wait_urb
Alan Stern [Tue, 24 Jul 2007 22:23:23 +0000 (18:23 -0400)]
USB: Fix a bug in usb_start_wait_urb

This patch (as941) fixes a bug recently added to the USB synchronous
API.  The status of a completed URB must be preserved separately
across a completion callback.  Also, the actual_length value isn't
available until after the URB has fully completed.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix scatterlist PIO case (IOMMU)
David Brownell [Sun, 22 Jul 2007 22:13:13 +0000 (15:13 -0700)]
USB: fix scatterlist PIO case (IOMMU)

Update the scatterlist logic so that PIO options are also disabled
when an IOMMU may have coalesced pages during dma_map_sg() ... it's
not just HIGHMEM that can make trouble supporting both PIO and DMA
based host controller drivers.

There also seems to be a cross-arch issue here, with 64bit powerpc
not using an IOMMU define ... and its IOMMU_VMERGE config can always
be overridden on the kernel command line.  So this is better, but
still imperfect.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: fix usb_serial_suspend(): buggy code
Oliver Neukum [Mon, 23 Jul 2007 06:58:39 +0000 (08:58 +0200)]
USB: fix usb_serial_suspend(): buggy code

Am Montag 23 Juli 2007 schrieb Adrian Bunk:
> Commit ec22559e0b7a05283a3413bda5d177e42c950e23 added the following
> function to drivers/usb/serial/usb-serial.c:
>
[..]
>
> The Coverity checker spotted the inconsequent NULL checking for "serial".
>
> Looking at the code it also doesn't seem to have been intended to always
> return 0.

Coverity is right. The check for NULL is wrongly done and the error
return is lost.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: yet another quirky device
Oliver Neukum [Mon, 30 Jul 2007 10:09:59 +0000 (12:09 +0200)]
USB: yet another quirky device

another quirky scanner.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Add CanonScan LiDE30 to the quirk list
Johann Felix Soden [Sat, 28 Jul 2007 15:10:33 +0000 (17:10 +0200)]
USB: Add CanonScan LiDE30 to the quirk list

This patch adds CanoScan N1240U/LiDE30 (Scanner) to the list of quirky USB
devices.

Signed-off-by: Johann Felix Soden <johfel@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: even more quirks
Oliver Neukum [Fri, 20 Jul 2007 09:34:53 +0000 (11:34 +0200)]
USB: even more quirks

The number of quirky devices seems to be large.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: usb.h kernel-doc additions
Randy Dunlap [Thu, 19 Jul 2007 17:21:37 +0000 (10:21 -0700)]
USB: usb.h kernel-doc additions

Add kernel-doc entries in <linux/usb.h> for:

Warning(linux-2.6.22-git12//include/linux/usb.h:162): No description found for parameter 'intf_assoc'
Warning(linux-2.6.22-git12//include/linux/usb.h:268): No description found for parameter 'intf_assoc[USB_MAXIADS]'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: more quirky devices
Oliver Neukum [Thu, 19 Jul 2007 11:00:15 +0000 (13:00 +0200)]
USB: more quirky devices

our list of devices which cannot be suspended keeps growing.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: Don't let usb-storage steal Blackberry Pearl
Jeremy Katz [Thu, 19 Jul 2007 13:37:42 +0000 (09:37 -0400)]
USB: Don't let usb-storage steal Blackberry Pearl

The Blackberry Pearl can run in two modes; a usb-storage only mode
and a mode that allows access via mass storage and to its database.
The berry_charge module will set the device to dual mode and thus we
should ignore its native mode if that module is built

Signed-off-by: Jeremy Katz <katzj@redhat.com>
Signed-off-by: Phil Dibowitz <phil@ipom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoUSB: devices misc: Trivial patch to build the IOWARRIOR when it is selected in Kconfig
Juergen Beisert [Thu, 19 Jul 2007 15:02:59 +0000 (17:02 +0200)]
USB: devices misc: Trivial patch to build the IOWARRIOR when it is selected in Kconfig

Trivial patch to build the IOWARRIOR when it is selected in Kconfig.

Signed-off-by: Juergen Beisert <jbe@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
16 years agoIB/ipath: Workaround problem of errormask register being overwritten
Dave Olson [Fri, 20 Jul 2007 21:41:26 +0000 (14:41 -0700)]
IB/ipath: Workaround problem of errormask register being overwritten

On some system hardware, we are seeing moderately common cases of the
chip errormask register being overwritten due to a chip bug in iba6120
that is triggered by a vendor-specific PCIe broadcast message.  This
patch merely checks periodically, and corrects it if needed (the
overwrite can cause us to not get error and hardware error
interrupts).  Also, make dd->ipath_errormask the one, true canonical
source for kr_errormask, and remove references to ipath_ignorederrs as
it is currently unused.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: John Gregor <john.gregor@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
16 years agoIB/ipath: Fix some issues with buffer cancel and sendctrl register update
Dave Olson [Fri, 20 Jul 2007 21:23:37 +0000 (14:23 -0700)]
IB/ipath: Fix some issues with buffer cancel and sendctrl register update

There was confused use of INFINIPATH_S_PIOBUFAVAILUPD (value) and
IPATH_S_PIOBUFAVAILUPD (bit position).  Also, some callers of
ipath_cancel_sends() need kr_sendctrl restored, and some want to do it
later.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
16 years agoIB/ipath: Use faster put_tid_2 routine after initialization
Dave Olson [Fri, 20 Jul 2007 20:34:02 +0000 (13:34 -0700)]
IB/ipath: Use faster put_tid_2 routine after initialization

At one time the ipath_minrev field was initialized prior to the
ipath_init_iba6120_funcs call, but that is no longer the case, so the
slower put_tid routine was always being used.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
16 years agoIB/ipath: Remove unsafe fastrcvint code from interrupt handler
Dave Olson [Fri, 20 Jul 2007 19:50:55 +0000 (12:50 -0700)]
IB/ipath: Remove unsafe fastrcvint code from interrupt handler

The fastrcvint code's purpose was to avoid reading the interrupt
status if kernel packets were in the receive queue (to reduce
overhead).  Because intstatus was not read, we could miss the error
interrupt bit indicating freeze mode, since it only delivers a single
interrupt, even if still pending after intclear is written.

This patch removes that unsafe optimization.

Signed-off-by: Dave Olson <dave.olson@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
16 years agoFix a potential NULL pointer dereference in mace_interrupt() in drivers/net/pcmcia...
Micah Gruber [Tue, 24 Jul 2007 02:44:56 +0000 (10:44 +0800)]
Fix a potential NULL pointer dereference in mace_interrupt() in drivers/net/pcmcia/nmclan_cs.c

This patch fixes a potential null dereference bug where we dereference
DEV before a null check. This patch simply moves the dereferencing after
the null check.

Signed-off-by: Micah Gruber <micah.gruber@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoPATCH kernel 2.6.22] PCMCIA-NETDEV : modify smc91c92_cs.c to become SMP safe
Komuro [Mon, 23 Jul 2007 12:36:06 +0000 (21:36 +0900)]
PATCH kernel 2.6.22] PCMCIA-NETDEV : modify smc91c92_cs.c to become SMP safe

protect smc_start_xmit, smc_interrupt and media_check by spin_lock.

Signed-off-by: Komuro <komurojun-mbn@nifty.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2io: Increment received packet count correctly
Ramkrishna Vepa [Thu, 26 Jul 2007 02:43:12 +0000 (19:43 -0700)]
S2io: Increment received packet count correctly

 - Fix to increment the received packet count correctly.

(Resending; Removed HTML sections in the patch)

Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2io: Fix crash when resetting adapter
Ramkrishna Vepa [Thu, 26 Jul 2007 02:40:33 +0000 (19:40 -0700)]
S2io: Fix crash when resetting adapter

- Removed the call to pci_set_power_state to reset the adapter as it was resulting
  in system crash on some platforms.

(Resending; Removed HTML sections in the patch)

Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2io: Mask spurious interrupts
Ramkrishna Vepa [Thu, 26 Jul 2007 02:35:09 +0000 (19:35 -0700)]
S2io: Mask spurious interrupts

- Mask single and double bit ETQ ecc errors to inhibit spurious interrupts.

(Resending; Removed HTML sections in the patch)

Signed-off-by: Santosh Rastapur <santosh.rastapur@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2IO: Implementing review comments from old patches
Veena Parat [Mon, 23 Jul 2007 06:39:43 +0000 (02:39 -0400)]
S2IO: Implementing review comments from old patches

 - Incorporated Jeff Garzik's comments on coding standards

Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2IO: Checking for the return value of pci map function
Veena Parat [Mon, 23 Jul 2007 06:37:14 +0000 (02:37 -0400)]
S2IO: Checking for the return value of pci map function

 - Checking for the return value of pci map function
 - Implemented Francois Romieu's comments on eliminating code duplication
   using goto
 - Implemented Francois Romieu's comments on using a temporary variable for
   accessing statistics structure

Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2IO: Removing MSI support from driver
Veena Parat [Mon, 23 Jul 2007 06:23:54 +0000 (02:23 -0400)]
S2IO: Removing MSI support from driver

 - Removed MSI support from driver - unused feature
 - Replaced request_mem_region with pci_request_regions

Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoS2IO: Removing 3 buffer mode support from the driver
Veena Parat [Mon, 23 Jul 2007 06:20:51 +0000 (02:20 -0400)]
S2IO: Removing 3 buffer mode support from the driver

 - Removed 3 buffer mode support from driver - unused feature
 - Incorporated Jeff Garzik's comments on elimination of inline typecasting
 - Code cleanup : Removed a few extra spaces

Signed-off-by: Veena Parat <veena.parat@neterion.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agonetxen: drop redudant spinlock
Dhananjay Phadke [Fri, 27 Jul 2007 17:42:11 +0000 (23:12 +0530)]
netxen: drop redudant spinlock

Some leftover code that makes use of adapter->lock in tx_timeout function,
which resets the interface under this lock. In close() when the workqueue
is flushed, prints the warning about sleeping with interrupts disabled
(when spinlock debug is enabled). The lock was required with private netxen
IOCTLs, which were removed a while ago.

Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agonetxen: Fix interrupt handling for multiport adapters
dhananjay@netxen.com [Wed, 25 Jul 2007 14:43:12 +0000 (20:13 +0530)]
netxen: Fix interrupt handling for multiport adapters

This patch fixes masking of interrupts on multiport adapters. Also disables
interrupts upon ifdown interface. The wrong mask could result in interrupt
flood after interface is down.

Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agonetxen: re-init station address after h/w init
dhananjay@netxen.com [Wed, 25 Jul 2007 14:43:11 +0000 (20:13 +0530)]
netxen: re-init station address after h/w init

This is a workaround for firmware bug with 2nd port of multiport adapter,
where MAC address is reset. Driver just needs to overwrite it with the
value read from PROM.

Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agotulip: Remove tulip maintainer
Valerie Henson [Mon, 30 Jul 2007 19:10:52 +0000 (13:10 -0600)]
tulip: Remove tulip maintainer

Remove Val Henson as tulip maintainer and let her roam free, FREE!

Signed-off-by: Val Henson <val@nmt.edu>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agoforcedeth: mac address correct
Ayaz Abdulla [Fri, 27 Jul 2007 03:46:00 +0000 (23:46 -0400)]
forcedeth: mac address correct

In older chipsets, the mac address was stored in reversed order.
However, in newer chipsets, the mac address is in correct order. This
patch takes those newer chipsets into account and does not rely on a
special bit setup by BIOS'.

Signed-Off-By: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
16 years agogfar: Fix modpost warning
Kumar Gala [Thu, 26 Jul 2007 05:52:34 +0000 (00:52 -0500)]
gfar: Fix modpost warning

Fix the following modpost warning:

WARNING: vmlinux.o(.init.text+0x1aa6c): Section mismatch: reference to .exit.text:gfar_mdio_exit (between 'gfar_init' and 'gfar_mdio_init')

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>