When iputting the inode, We may leave the delayed nodes if they have some
delayed items that have not been dealt with. So when the inode is read again,
we must look up the relative delayed node, and use the information in it to
initialize the inode. Or we will get inconsonant inode information, it may
cause that the same directory index number is allocated again, and hit the
following oops:
[ 5447.554187] err add delayed dir index item(name: pglog_0.965_0) into the
insertion tree of the delayed node(root id: 262, inode id: 258, errno: -17)
[ 5447.569766] ------------[ cut here ]------------
[ 5447.575361] kernel BUG at fs/btrfs/delayed-inode.c:1301!
[SNIP]
[ 5447.790721] Call Trace:
[ 5447.793191] [<ffffffffa0641c4e>] btrfs_insert_dir_item+0x189/0x1bb [btrfs]
[ 5447.800156] [<ffffffffa0651a45>] btrfs_add_link+0x12b/0x191 [btrfs]
[ 5447.806517] [<ffffffffa0651adc>] btrfs_add_nondir+0x31/0x58 [btrfs]
[ 5447.812876] [<ffffffffa0651d6a>] btrfs_create+0xf9/0x197 [btrfs]
[ 5447.818961] [<ffffffff8111f840>] vfs_create+0x72/0x92
[ 5447.824090] [<ffffffff8111fa8c>] do_last+0x22c/0x40b
[ 5447.829133] [<ffffffff8112076a>] path_openat+0xc0/0x2ef
[ 5447.834438] [<ffffffff810c58e2>] ? __perf_event_task_sched_out+0x24/0x44
[ 5447.841216] [<ffffffff8103ecdd>] ? perf_event_task_sched_out+0x59/0x67
[ 5447.847846] [<ffffffff81121a79>] do_filp_open+0x3d/0x87
[ 5447.853156] [<ffffffff811e126c>] ? strncpy_from_user+0x43/0x4d
[ 5447.859072] [<ffffffff8111f1f5>] ? getname_flags+0x2e/0x80
[ 5447.864636] [<ffffffff8111f179>] ? do_getname+0x14b/0x173
[ 5447.870112] [<ffffffff8111f1b7>] ? audit_getname+0x16/0x26
[ 5447.875682] [<ffffffff8112b1ab>] ? spin_lock+0xe/0x10
[ 5447.880882] [<ffffffff81112d39>] do_sys_open+0x69/0xae
[ 5447.886153] [<ffffffff81112db1>] sys_open+0x20/0x22
[ 5447.891114] [<ffffffff813b9aab>] system_call_fastpath+0x16/0x1b
Fix it by reusing the old delayed node.
Reported-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Tested-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Evidently, the device sometimes wants to write back
to command buffers, even if I see no reason why it
should. Allow it to do that.
Tested-by: Andy Lutomirski <luto@mit.edu>
Tested-by: Kyle McMartin <kyle@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
When we stop the device while a command is in
flight that uses multiple TBs, we can leak the
DMA buffers for the second and higher TBs. Fix
this by using iwlagn_unmap_tfd() as we do when
we normally recover the entry.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
A recently added feature to the firmware enables the driver to retrieve
firmware logs via the host bus (SDIO or SPI).
There are two modes of operation:
1. On-demand: The FW collects its log in an internal ring buffer. This
buffer can later be read, for example, upon recovery.
2. Continuous: The FW pushes the FW logs as special packets in the RX
path.
Reading the internal ring buffer does not involve the FW. Thus, as long
as the HW is not in ELP, it should be possible to read the logs, even if
the FW crashes.
A sysfs binary file named "fwlog" was added to support this feature,
letting a monitor process read the FW messages. The log is transferred
from the FW only when available, so the reading process might block.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
During recovery work commands sent to the FW could fail and schedule
additional recovery work. Since the chip is going to be powered off,
avoid recursive recoveries.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The FW initialization might depend on the FW revision, so check for any
FW quirks right after booting it.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When resuming (after wowlan), we want the rx packets (which is
usually the wake-up packet itself) to be passed to mac80211 only
after the resume notifier was completed, and mac80211 is up and
running (otherwise, the packets will be dropped).
By enqueueing the netstack_work to a freezable workqueue, we can
guarantee the rx processing to occur only after mac80211 was resumed.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
wl1271_flush_deferred_work(), which calls ieee80211_rx() and
ieee80211_tx_status(), is called from a process context.
hence, use ieee80211_tx_status_ni() instead of ieee80211_tx_status().
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When resuming while connected (without wowlan), the interface
is already IF_OPER_UP, so we won't get the notifier callback,
and hence never complete the association (from wl12xx perspective)
This situation, among other potential problems, prevents the
station from entering psm.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Some platforms don't support the wake_irq, so disable wowlan
in this case, and avoid the "Unbalanced IRQ wake disable"
warning on disable_irq_wake().
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Beacon filtering needs to be enabled so AP won't wake up by
by every received beacon.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
We set wl->wow_enabled on every suspend(), so we need to clear it
on every resume().
(we can't rely on setting wl->wow_enabled=false in suspend(),
as it being called only when wowlan triggers are configured)
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Since mac80211 calls suspend/resume only when wowlan triggers
exist, there is no need to check for triggers existance in the
callbacks as well.
Add a WARN_ON() to verify it.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The Low Power DRPw (LPD) mode contains several optimizations
that designed to reduce power consumption. The purpose
is to save current consumption in RX and Listen mode.
LPD setting apply only for wl127x AP mode (not wl128x)
Signed-off-by: Shahar Levi <shahar_levi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
If our SDIO function has its runtime PM disabled, don't try to
manipulate its runtime PM status at all. This way we can still
power on cards plugged to mmc hosts that are not MMC_CAP_POWER_OFF_CARD.
Reported-and-tested-by: Tim Yamin <tim@kangatronix.co.uk>
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The firmware, in practice, treats the channels in three separate
blocks, one for each band (bg, a and j). Instead of using a single
array and doing some magic with indices, split the array in 3 to make
it more readable.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Due to rebase error, the patch for commit cb5ae0 ("wl12xx: configure
rates when working in ibss mode") was wrong - a blob was added
into the wrong function. fix it.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
ioc->ioc_data is rcu protectd, so uses correct API to access it.
This doesn't change any behavior, but just make code consistent.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: stable@kernel.org # after ab4bd22d
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
I got a rcu warnning at boot. the ioc->ioc_data is rcu_deferenced, but
doesn't hold rcu_read_lock.
Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Cc: stable@kernel.org # after ab4bd22d
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
crc7 is used only in wl12xx_spi.
Remove redundant crc7.h includes, and update Kconfig to select CRC7
only if WL12XX_SPI is being selected.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Beacon early termination doesn't help much in the 5GHz band and masks
channel switch IE Beacons. Thus, change the code to use BET only in
2.4GHz.
[Reworded the commit log slightly -- Luca.]
Signed-off-by: Shahar Levi <shahar_levi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Allow control over rx_streaming interval and operation mode
(always/only on coex) via debugfs.
e.g.
echo 100 > /debug/ieee80211/phy0/wl12xx/rx_streaming/interval
echo 1 > /debug/ieee80211/phy0/wl12xx/rx_streaming/always
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When rx_streaming.interval is non-zero, use automatic rx streaming.
Enable rx streaming on the each rx/tx packet, and disable it
rx_streaming.duration msecs later.
When rx_streaming.always=0 (default), rx streaming is enabled only
when there is a coex operation.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
wl12xx supports the "rx streaming" feature:
When in ps mode, and @timeout msecs have been passed since
the last rx/tx, it issues trigger packets (QoS-null/PS-Poll packets,
according to the ac type) in const intervals (in order to reduce
the rx time).
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Those have little value. Remove those to
make the driver less noisy.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Nops aren't needed. When we actually need
those calls, then we add them with meat
and barbecue sauce.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
->init and ->reset are optional - at least
sdio.c doesn't implement them - so allow those
pointers to be NULL.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Those have little value. Remove those to make
the driver less noisy.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
That's only needed during init anyway, let's free
some space after we're done probing.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Results on dummy device can be seen in my netconf 2011
slides. These results are for a 10Gige IXGBE intel
nic - on another i5 machine, very similar specs to
the one used in the netconf2011 results.
It turns out - this is a hell lot worse than dummy
and so this patch is even more beneficial for 10G.
Test setup:
----------
System under test sending packets out.
Additional box connected directly dropping packets.
Installed prio qdisc on the eth device and default
netdev default length of 1000 used as is.
The 3 prio bands each were set to 100 (didnt factor in
the results).
5 packet runs were made and the middle 3 picked.
results
-------
The "cpu" column indicates the which cpu the sample
was taken on,
The "Pkt runx" carries the number of packets a cpu
dequeued when forced to be in the "dequeuer" role.
The "avg" for each run is the number of times each
cpu should be a "dequeuer" if the system was fair.
3.0-rc4 (plain)
cpu Pkt run1 Pkt run2 Pkt run3
================================================
cpu0 21853354 21598183 22199900
cpu1 431058 473476 393159
cpu2 481975 477529 458466
cpu3 23261406 23412299 22894315
avg 11506948 11490372 11486460
3.0-rc4 with patch and default weight 64
cpu Pkt run1 Pkt run2 Pkt run3
================================================
cpu0 13205312 13109359 13132333
cpu1 10189914 10159127 10122270
cpu2 10213871 10124367 10168722
cpu3 13165760 13164767 13096705
avg 11693714 11639405 11630008
As you can see the system is still not perfect but
is a lot better than what it was before...
At the moment we use the old backlog weight, weight_p
which is 64 packets. It seems to be reasonably fine
with that value.
The system could be made more fair if we reduce the
weight_p (as per my presentation), but we are going
to affect the shared backlog weight. Unless deemed
necessary, I think the default value is fine. If not
we could add yet another knob.
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use pr_fmt, pr_<level> and netdev_<level> as appropriate.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Initialization of this field to "all priorities" must be done before MCC queue
creation. As soon as the MCC queue is created, an event modifying this value
may be received.
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Problem initially reproted and fixed by Eric Dumazet <eric.dumazet@gmail.com>
netdev_stats_update() resets netdev->stats and then accumulates stats from
various rings. This is wrong as stats readers can sometimes catch zero values.
Use temporary variables instead for accumulating per-ring values.
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
cifs: propagate errors from cifs_get_root() to mount(2)
cifs: tidy cifs_do_mount() up a bit
cifs: more breakage on mount failures
cifs: close sget() races
cifs: pull freeing mountdata/dropping nls/freeing cifs_sb into cifs_umount()
cifs: move cifs_umount() call into ->kill_sb()
cifs: pull cifs_mount() call up
sanitize cifs_umount() prototype
cifs: initialize ->tlink_tree in cifs_setup_cifs_sb()
cifs: allocate mountdata earlier
cifs: leak on mount if we share superblock
cifs: don't pass superblock to cifs_mount()
cifs: don't leak nls on mount failure
cifs: double free on mount failure
take bdi setup/destruction into cifs_mount/cifs_umount
Acked-by: Steve French <smfrench@gmail.com>
This makes things a little clearer and prevents us from running old code
on a new chipset that may not be supported.
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewied-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Keith Packard <keithp@keithp.com>
SCSI defines discard alignment as the offset to the first
optimal discard. In the case of SD/MMC, that is always zero
which is the default.
SCSI defines discard granularity as a hint of a optimal
discard size. That is much better expressed by the MMC
"preferred erase size" (pref_erase) field.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
For example, an eMMC with 2 boot partitions will have 3 threads.
The names change from:
40 ? 00:00:00 mmcqd/0
41 ? 00:00:00 mmcqd/0
42 ? 00:00:00 mmcqd/0
to:
40 ? 00:00:00 mmcqd/0
41 ? 00:00:00 mmcqd/0boot0
42 ? 00:00:00 mmcqd/0boot1
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andrei Warkentin <andreiw@motorola.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
The erase timeout calculation may depend on clock rate
which is zero if the clock is gated, so use
mmc_host_clk_rate() which allows for that case.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
The MMC block driver and other drivers (e.g. mmc-test) will expect
the card to be switched to the User Data Area eMMC partition when
they start. Hence the MMC block driver should ensure it is that
way when it is removed.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andrei Warkentin <andreiw@motorola.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
