Pavel Emelyanov [Fri, 28 Mar 2008 23:38:43 +0000 (16:38 -0700)]
[SOCK]: Introduce a percpu inuse counters array (v2).
And redirect sock_prot_inuse_add and _get to use one.
As far as the dereferences are concerned. Before the patch we made
1 dereference to proto->inuse.add call, the call itself and then
called the __get_cpu_var() on a static variable. After the patch we
make a direct call, then one dereference to proto->inuse_idx and
then the same __get_cpu_var() on a still static variable. So this
patch doesn't seem to produce performance penalty on SMP.
This is not per-net yet, but I will deliberately make NET_NS=y case
separated from NET_NS=n one, since it'll cost us one-or-two more
dereferences to get the struct net and the inuse counter.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Acked-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Fri, 28 Mar 2008 23:38:17 +0000 (16:38 -0700)]
[SOCK]: Enumerate struct proto-s to facilitate percpu inuse accounting (v2).
The inuse counters are going to become a per-cpu array. Introduce an
index for this array on the struct proto.
To handle the case of proto register-unregister-register loop the
bitmap is used. All its bits manipulations are protected with
proto_list_lock and a sanity check for the bitmap being exhausted is
also added.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Acked-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[NET] NEIGHBOUR: Extract hash/lookup functions for pneigh entries.
Extract hash function for pneigh entries from pneigh_lookup(),
__pneigh_lookup() and pneigh_delete() as pneigh_hash().
Extract core of pneigh_lookup() and __pneigh_lookup() as
__pneigh_lookup_1().
Ilpo Järvinen [Fri, 28 Mar 2008 00:54:29 +0000 (17:54 -0700)]
[SCTP]: Remove sctp_add_cmd_sf wrapper bloat
With a was number of callsites sctp_add_cmd_sf wrapper bloats
kernel by some amount. Due to unlikely tracking allyesconfig,
with the initial result were around ~7kB (thus caught my
attention) while a non-debug config produced only ~2.3kB effect.
I (ij) proposed first a patch to uninline it but Vlad responded
with a patch that removed the only sctp_add_cmd call which is
wrapped by sctp_add_cmd_sf (I wasn't sure if I could do that).
I did minor cleanup to Vlad's patch.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Thomas Graf [Thu, 27 Mar 2008 23:08:03 +0000 (16:08 -0700)]
[ESP]: Ensure IV is in linear part of the skb to avoid BUG() due to OOB access
ESP does not account for the IV size when calling pskb_may_pull() to
ensure everything it accesses directly is within the linear part of a
potential fragment. This results in a BUG() being triggered when the
both the IPv4 and IPv6 ESP stack is fed with an skb where the first
fragment ends between the end of the esp header and the end of the IV.
This bug was found by Dirk Nehring <dnehring@gmx.net> .
Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Denis V. Lunev [Thu, 27 Mar 2008 21:26:30 +0000 (14:26 -0700)]
[NETNS]: Do no include NET related headers if CONFIG_NET is not set.
This fix broken compilation for 'allnoconfig'. This was introduced by
Introduced by commit 1218854afa6f659be90b748cf1bc7badee954a35 ("[NET]
NETNS: Omit seq_net_private->net without CONFIG_NET_NS.")
Signed-off-by: Denis V. Lunev <den@openvz.org> Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Denis V. Lunev [Thu, 27 Mar 2008 21:25:53 +0000 (14:25 -0700)]
[NETNS]: Compile NET /proc support only if CONFIG_NET is set.
This fix broken compilation for 'allnoconfig'. This was introduced by
Introduced by commit 1218854afa6f659be90b748cf1bc7badee954a35 ("[NET]
NETNS: Omit seq_net_private->net without CONFIG_NET_NS.")
Signed-off-by: Denis V. Lunev <den@openvz.org> Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Holger Schurig [Wed, 26 Mar 2008 09:04:44 +0000 (10:04 +0100)]
libertas: reduce debug output
This patch tries to make dmesg logs between different runs easier
to compare by
* removing the jiffies (use CONFIG_PRINTK_TIME if you need
timing)
* remove the line numbers, they change with each applied patch
It also changes the deprecated __FUNCTION__ to __func__ to make
checkpatch.pl happy.
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Wed, 26 Mar 2008 22:21:47 +0000 (23:21 +0100)]
mac80211: reorder fields to make some structures smaller
This patch reorders some fields in various structures to have
less padding within the structures, making them smaller. It
doesn't yet make any type adjustments, but often size_t is used
for example for IE lengths which is total overkill since size_t
will be 8 bytes long on 64-bit yet the length can at most fill
a u8.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ron Rindjunsky [Wed, 26 Mar 2008 18:36:03 +0000 (20:36 +0200)]
mac80211: A-MPDU MLME use dynamic allocation
This patch alters the A-MPDU MLME in sta_info to use dynamic allocation,
thus drastically improving memory usage - from a constant ~2 Kbyte in
the previous (static) allocation to a lower limit of ~200 Byte and an upper
limit of ~2 Kbyte.
Signed-off-by: Ron Rindjunsky <ron.rindjunsky@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Johannes Berg [Wed, 26 Mar 2008 13:14:55 +0000 (14:14 +0100)]
cfg80211: don't export ieee80211_get_channel
This patch makes ieee80211_get_channel a static inline defined in
cfg80211's header file which simply calls __ieee80211_get_channel
to avoid symbol clashes with the ieee80211 code.
The problem was pointed out by David Miller, thanks!
Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Cc: David Miller <davem@davemloft.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Holger Schurig [Wed, 26 Mar 2008 12:26:55 +0000 (13:26 +0100)]
libertas: the compact flash driver is no longer experimental
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Holger Schurig [Wed, 26 Mar 2008 08:57:45 +0000 (09:57 +0100)]
libertas: remove CMD_802_11_PWR_CFG
This has nowhere been used. Note: in the firmware manual this was
documented as CMD_802_11_PA_CFG. If we ever need it, we can/should
re-implement it as a direct command.
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Holger Schurig [Wed, 26 Mar 2008 08:57:14 +0000 (09:57 +0100)]
libertas: kill useless #define LBS_MONITOR_OFF 0
Signed-off-by: Holger Schurig <hs4233@mail.mn-solutions.de> Acked-by: Dan Williams <dcbw@redhat.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tomas Winkler [Tue, 25 Mar 2008 23:33:41 +0000 (16:33 -0700)]
iwlwifi: iwl_priv - clean up in types of members
This patch fix types of is_open and iw_mode members
of iwl_priv sturct
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
iwlwifi: allow a default callback for ASYNC host commands
This patch provides a default callback for ASYNC host commands instead
of calling to BUG_ON. Most of the callbacks are now just empty functions
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Ron Rindjunsky [Tue, 25 Mar 2008 23:33:39 +0000 (16:33 -0700)]
mac80211: fix wrong Rx A-MPDU control via debugfs
This patch eliminate the use of buf_size as a trigger in favor of a new
flag to control Rx A-MPDU sessions through debugfs
Signed-off-by: Ron Rindjunsky <ron.rindjunsky@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tomas Winkler [Tue, 25 Mar 2008 23:33:38 +0000 (16:33 -0700)]
iwlwifi: improve NIC i/o debug prints information
This patch gives the function's caller name in case NIC
access reference count was not used by it.
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Ron Rindjunsky <ron.rindjunsky@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Tomas Winkler [Tue, 25 Mar 2008 23:33:37 +0000 (16:33 -0700)]
iwlwifi: rename iwl-4965-io.h to iwl-io.h
This patch renames iwl-4965-io.h back to iw-io.h
it also remove 4965 from all functions it supplies
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: Reinette Chatre <reinette.chatre@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Mohamed Abbas [Tue, 25 Mar 2008 23:33:36 +0000 (16:33 -0700)]
iwlwifi: Add led support
This patch add LEDS support to 3965 and 4965 drivers. It is based on
led trigger and class. For our drivers we needed to avoid two things.
1- We receive led trigger on/off on each Rx\Tx frame. In our driver
we can not call led command like that. In this driver once driver
receive a start of traffic it call the led command to start blinking
then we count all bytes of Tx and Rx frame, after two second we count the
blink rate of last two second then id blink rate changed we call the led
commands
2- Since we can call led command very often, we make sure we call the
led command after we receive the statistics notification so
we don't need to wake up the ucode id it is in sleep state.
This patch was tested with 4965 and 3945.
Signed-off-by: Mohamed Abbas <mabbas@linux.intel.com> Signed-off-by: Ian Schram<ischram@telenet.be> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Benjamin Thery [Wed, 26 Mar 2008 23:53:30 +0000 (16:53 -0700)]
[NETNS][IPV6] flowlabels - make proc per namespace
Make /proc/net/ip6_flowlabel show only flow labels belonging to the
current network namespace.
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Benjamin Thery [Wed, 26 Mar 2008 23:53:08 +0000 (16:53 -0700)]
[NETNS][IPV6] flowlabels - make flowlabels per namespace
This patch introduces a new member, fl_net, in struct ip6_flowlabel.
This allows to create labels with the same value in different namespaces.
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Lezcano [Wed, 26 Mar 2008 23:52:32 +0000 (16:52 -0700)]
[NETNS][IPV6] anycast - handle several network namespace
Make use of the network namespace information to have this protocol to
handle several network namespace.
Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Wed, 26 Mar 2008 23:51:09 +0000 (16:51 -0700)]
[IPSEC]: Fix BEET output
The IPv6 BEET output function is incorrectly including the inner
header in the payload to be protected. This causes a crash as
the packet doesn't actually have that many bytes for a second
header.
The IPv4 BEET output on the other hand is broken when it comes
to handling an inner IPv6 header since it always assumes an
inner IPv4 header.
This patch fixes both by making sure that neither BEET output
function touches the inner header at all. All access is now
done through the protocol-independent cb structure. Two new
attributes are added to make this work, the IP header length
and the IPv4 option length. They're filled in by the inner
mode's output function.
Thanks to Joakim Koskela for finding this problem.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 26 Mar 2008 23:49:40 +0000 (16:49 -0700)]
[IPV6]: Fix potential net leak and oops in ipv6 routing code.
The commits f3db4851 ([NETNS][IPV6] ip6_fib - fib6_clean_all handle several
network namespaces) and 69ddb805 ([NETNS][IPV6] route6 - Make proc entry
/proc/net/rt6_stats per namespace) made some proc files per net.
Both of them introduced potential OOPS - get_proc_net can return NULL, but
this check is lost - and a struct net leak - in case single_open() fails the
previously got net is not put.
Kill all these bugs with one patch.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Acked-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Allan Stephens [Wed, 26 Mar 2008 23:48:21 +0000 (16:48 -0700)]
[TIPC]: Cosmetic cleanup of TIPC polling logic
This patch eliminates an unnecessary poll-related routine
by merging it into TIPC's main polling routine, and updates
the comments associated with this code.
Signed-off-by: Allan Stephens <allan.stephens@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 26 Mar 2008 23:27:22 +0000 (16:27 -0700)]
[VLAN]: Reduce memory consumed by vlan_groups
Currently each vlan_groupd contains 8 pointers on arrays with 512
pointers on struct net_device each :) Such a construction "in many
cases ... wastes memory".
My proposal is to allow for some of these arrays pointers be NULL,
meaning that there are no devices in it. When a new device is added
to the vlan_group, the appropriate array is allocated.
The check in vlan_group_get_device's is safe, since the pointer
vg->vlan_devices_arrays[x] can only switch from NULL to not-NULL.
The vlan_group_prealloc_vid() is guarded with rtnl lock and is
also safe.
I've checked (I hope that) all the places, that use these arrays
and found, that the register_vlan_dev is the only place, that can
put a vlan device on an empty vlan_group.
Rough calculations shows, that after the patch a setup with a
single vlan dev (or up to 512 vlans with sequential vids) will
occupy approximately 8 times less memory.
The question I have is - does this patch makes sense, or a totally
new structures are required to store the vlan_devs?
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Patrick McHardy <kaber@trash.net>
Pavel Emelyanov [Wed, 26 Mar 2008 09:27:09 +0000 (02:27 -0700)]
[ICMP]: Dst entry leak in icmp_send host re-lookup code (v2).
Commit 8b7817f3a959ed99d7443afc12f78a7e1fcc2063 ([IPSEC]: Add ICMP host
relookup support) introduced some dst leaks on error paths: the rt
pointer can be forgotten to be put. Fix it bu going to a proper label.
Found after net namespace's lo refused to unregister :) Many thanks to
Den for valuable help during debugging.
Herbert pointed out, that xfrm_lookup() will put the rtable in case
of error itself, so the first goto fix is redundant.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Wed, 26 Mar 2008 09:12:11 +0000 (02:12 -0700)]
[NET]: Fix multicast device ioctl checks
SIOCADDMULTI/SIOCDELMULTI check whether the driver has a set_multicast_list
method to determine whether it supports multicast. Drivers implementing
secondary unicast support use set_rx_mode however.
Check for both dev->set_multicast_mode and dev->set_rx_mode to determine
multicast capabilities.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Pavel Emelyanov [Wed, 26 Mar 2008 08:56:24 +0000 (01:56 -0700)]
[NETNS][ICMP]: Make ctl tables for ICMP sysctls per-net.
Add some flesh to ipv4_sysctl_init_net and ipv4_sysctl_exit_net,
i.e. copy the table, alter .data pointers and register it per-net.
Other ipv4_table's sysctls are now global, but this is going to
change once sysctl permissions patches migrate from -mm tree to
mainline in 2.6.26 merge window :)
Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Denis V. Lunev [Wed, 26 Mar 2008 07:48:17 +0000 (00:48 -0700)]
[NETNS]: Compilation warnings under CONFIG_NET_NS.
Recent commits from YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
have been introduced a several compilation warnings
'assignment discards qualifiers from pointer target type'
due to extra const modifier in the inline call parameters of
{dev|sock|twsk}_net_set.
Drop it.
Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Wed, 26 Mar 2008 07:16:29 +0000 (00:16 -0700)]
[UML]: uml-net: don't set IFF_ALLMULTI in set_multicast_list
IFF_ALLMULTI is an indication from the network stack to the driver
to disable multicast filters, drivers should never set it directly.
Since the UML networking device doesn't have any filtering capabilites,
it doesn't the set_multicast_list function at all, it is kept so userspace
can still issue SIOCADDMULTI/SIOCDELMULTI ioctls however.
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Patrick McHardy [Wed, 26 Mar 2008 07:15:17 +0000 (00:15 -0700)]
[VLAN]: Don't copy ALLMULTI/PROMISC flags from underlying device
Changing these flags requires to use dev_set_allmulti/dev_set_promiscuity
or dev_change_flags. Setting it directly causes two unwanted effects:
- the next dev_change_flags call will notice a difference between
dev->gflags and the actual flags, enable promisc/allmulti
mode and incorrectly update dev->gflags
- this keeps the underlying device in promisc/allmulti mode until
the VLAN device is deleted
Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
- Fix Rx/Tx HW interrupt coalescing counter reset logic. Disabling
is required before resetting the counter.
- Update the Default both Rx and Tx coalescing timer
threshold. Formerly 4 is set which is equal to 1.5 frame at the line
rate of 1GbE interface, and it doesn't match to the coalescing frame
count which is set to 16. Threashold 21 is matched to frame count 16.
Signed-off-by: Dai Haruki <dai.haruki@freescale.com> Signed-off-by: Andy Fleming <afleming@freescale.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Andy Fleming [Mon, 24 Mar 2008 15:53:27 +0000 (10:53 -0500)]
gianfar: Only process completed frames
If the LAST bit is not set in the RxBD, it's possible we're processing
an incomplete frame, which is bad. While we're at it, add a constant
for the error bitmask, so the whole if-clause fits on one line,
and is more legible.
Signed-off-by: Dai Haruki <dai.haruki@freescale.com> Signed-off-by: Andy Fleming <afleming@freescale.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Dai Haruki [Mon, 24 Mar 2008 15:53:26 +0000 (10:53 -0500)]
gianfar: Fix frame size calculation when hardware VLAN acceleration is on
In gfar_change_mtu(), the frame size needs to be increased to account
for the extra 4 bytes VLAN adds to the ethernet header. However,
it was being increased by the length of the whole header (18 bytes),
which is wrong.
Signed-off-by: Dai Haruki <dai.haruki@freescale.com> Signed-off-by: Andy Fleming <afleming@freescale.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Anton Vorontsov [Tue, 4 Mar 2008 16:41:32 +0000 (19:41 +0300)]
phy/broadcom: add support for BCM5481 PHY
This patch adds support for BCM5481 PHY. Unfortunately it's hard to
get specifications for this PHY, so its special register 0x18 isn't
annotated properly (but we know it's used to set up the delays).
I've kept the magic numbers, so we'll not forget to fix it at the
first opportunity, and will name that register and its bits correctly.
Marcin Slusarz [Tue, 4 Mar 2008 23:19:20 +0000 (15:19 -0800)]
qla3xxx: convert byte order of constant instead of variable
Convert byte order of constant instead of variable which can be done at
compile time (vs run time)
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com> Acked-by: Ron Mercer <ron.mercer@qlogic.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Florian Fainelli [Wed, 19 Mar 2008 16:14:51 +0000 (17:14 +0100)]
Add support the Korina (IDT RC32434) Ethernet MAC
This patch adds support for the IDT rc32434 Ethernet MAC
we can find in the IDT boards and the Mikrotik RB500.
Driver references some code from the linux-mips RB500
support.
Signed-off-by: Florian Fainelli <florian.fainelli@telecomint.eu> Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: Philip Rischel <rischelp@idt.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Al Viro [Wed, 19 Mar 2008 09:44:29 +0000 (09:44 +0000)]
ni52: switch to ioremap()
isa_bus_to_virt() is the wrong thing to do here; it happens
to work on i386, but only by accident. What we want is
normal ioremap/readb/etc. set - it's all in iomem.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Al Viro [Wed, 19 Mar 2008 09:44:09 +0000 (09:44 +0000)]
ni52: clean up initialization of priv
* initialize spinlock once
* check586() used to be done before we'd allocated ->priv; these days
it's there from the very beginning, so we don't have to play with
private copy. Consequently, we don't need to mess with reinitializing
->base, etc. afterwards.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Al Viro [Wed, 19 Mar 2008 09:43:49 +0000 (09:43 +0000)]
ni52: kill the second argument of check586(), get rid of bogus pointer
Passing ISA bus address explicitly cast to char * only to cast it back to
unsigned long is dumb; so's passing it at all when it's always dev->mem_start...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Al Viro [Wed, 19 Mar 2008 09:43:29 +0000 (09:43 +0000)]
ni52 trivial iomem annotations
Driver is still broken, though; partially from Alan's checkpatch-induced
fun, partially from layers of ancient mess ;-)
By the end of the series... hell, might be even worth trying to stick
such card into old alpha or ppc with an ISA slot and see if it work -
would be for the first time ever in case of alpha and for the first
time since at least 2.5.3 in case of ppc...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Jesse Brandeburg [Fri, 21 Mar 2008 18:07:03 +0000 (11:07 -0700)]
ixgb: remove irq_sem
ixgb can remove irq_sem by auditing all the call sites to make sure
that each of them makes sure the adapter is in the correct state
before re-enabling interrupts. after doing this to all of our other
drivers it is becoming easier.
Jesse Brandeburg [Fri, 21 Mar 2008 18:06:58 +0000 (11:06 -0700)]
e1000: remove irq_sem
irq_sem was just a hack to prevent interrupts from being enabled
unexpectedly in deep call paths. Simply finding those call paths and
fixing them by hand results in a driver that behaves as we expect and
doesn't need the atomic at all.
Jesse Brandeburg [Fri, 21 Mar 2008 18:06:32 +0000 (11:06 -0700)]
ixgb: add explicit state checking
In order to remove the irq_sem code we need to implement strict
adapter state checking to prevent accidental double up or downs
or resets. This code is largely copied from e1000/e1000e.
Michael Buesch [Tue, 25 Mar 2008 17:04:46 +0000 (18:04 +0100)]
b44: Truncate PHY address
Some ROMs on embedded devices store incorrect values for
the PHY address of the ethernet device.
It looks like the number is sign-extended.
Truncate the value by applying the PHY-address mask to it.
The patch was tested on a bcm47xx embedded system (where the bug
triggers) and a bcm4400 PCI card.
Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: Jeff Garzik <jeff@garzik.org>
Jussi Kivilinna [Sun, 23 Mar 2008 10:45:44 +0000 (12:45 +0200)]
rndis_host: fix oops when query for OID_GEN_PHYSICAL_MEDIUM fails
When query for OID_GEN_PHYSICAL_MEDIUM fails, uninitialized pointer
'phym' is being accessed in generic_rndis_bind(), resulting OOPS.
Patch fixes phym to be initialized and setup correctly when
rndis_query() for physical medium fails.
Bug was introduced by following commit:
commit 039ee17d1baabaa21783a0d5ab3e8c6d8c794bdf
Author: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Date: Sun Jan 27 23:34:33 2008 +0200
Reported-by: Dmitri Monakhov <dmonakhov@openvz.org> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David Brownell <dbrownell@users.sourceforge.net> Signed-off-by: Jeff Garzik <jeff@garzik.org>
The problem is that reg_lock is used with plain spin_lock() in
drivers/net/cxgb3/sge.c but is used with spin_lock_irqsave() in
drivers/net/cxgb3/cxgb3_offload.c. This is technically a false
positive, since the uses in sge.c are only in the initialization and
cleanup paths and cannot overlap with any use in interrupt context.
The best fix is probably just to use spin_lock_irq() with reg_lock in
sge.c. Even though it's not strictly required for correctness, it
avoids triggering lockdep and the extra overhead of disabling
interrupts is not important at all in the initialization and cleanup
slow paths.
Signed-off-by: Roland Dreier <rolandd@cisco.com> Signed-off-by: Jeff Garzik <jeff@garzik.org>
projects / linux-2.6 / log