When rtl8192ce is in operation, kmemleak indicates a number of leaks, yet
when the driver is removed all are gone. These false positives happen
in two locations:

unreferenced object 0xffff880041124000 (size 9536):
  comm "work_for_cpu", pid 9295, jiffies 4295037203 (age 20596.320s)
  hex dump (first 32 bytes):
    33 00 00 01 01 6d 00 00 00 00 b1 0e 21 0b 00 00  3....m......!...
    01 01 6d 00 00 00 00 8b 20 c0 e9 00 00 01 01 6d  ..m..... ......m
  backtrace:
    [<ffffffff8137a1b1>] kmemleak_alloc+0x21/0x50
    [<ffffffff8138d01f>] kmalloc_large_node+0x9a/0xa6
    [<ffffffff811335a5>] __kmalloc_node_track_caller+0x175/0x3b0
    [<ffffffff812ddfe3>] __alloc_skb+0x73/0x230
    [<ffffffff812de1b8>] dev_alloc_skb+0x18/0x30
    [<ffffffffa091a3b7>] rtl_pci_probe+0x10e0/0x17d2 [rtlwifi]
-- snip --
unreferenced object 0xffff8800b4d3f600 (size 256):
  comm "kworker/u:2", pid 13221, jiffies 4297830173 (age 9424.568s)
  hex dump (first 32 bytes):
    1c d6 45 b1 00 88 ff ff 1c d6 45 b1 00 88 ff ff  ..E.......E.....
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<ffffffff8137a1b1>] kmemleak_alloc+0x21/0x50
    [<ffffffff81130f43>] kmem_cache_alloc_node+0x153/0x270
    [<ffffffff812ddfb6>] __alloc_skb+0x46/0x230
    [<ffffffff812de1b8>] dev_alloc_skb+0x18/0x30
    [<ffffffffa093264a>] rtl92c_set_fw_rsvdpagepkt+0x22a/0x5c0 [rtl8192c_common]
-- snip --
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In https://bugzilla.kernel.org/show_bug.cgi?id=42976, a system with driver
rtl8192se used as an AP suffers from "Out of SW-IOMMU space" errors. These
are caused by the DMA buffers used for beacons never being unmapped.

This bug was also reported at
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/961618Reported-and-Tested-by: NDa Xue <da@lessconfused.com>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In https://bugzilla.kernel.org/show_bug.cgi?id=42976, a system with driver
rtl8192se used as an AP suffers from "Out of SW-IOMMU space" errors. These
are caused by the DMA buffers used for beacons never being unmapped.

This bug was also reported at
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/961618Reported-and-Tested-by: NDa Xue <da@lessconfused.com>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Fixes problem where caller would think routine succeeded when it failed
leading to divide by zero panic.
Signed-off-by: NSimon Graham <simon.graham@virtualcomputer.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Fixes problem where caller would think routine succeeded when it failed
leading to divide by zero panic.

(This also reverts an earlier attempt, commit 42bc0c97 "rtlwifi: Return
correct failure code on error". -- JWL)
Signed-off-by: NSimon Graham <simon.graham@virtualcomputer.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Callers of rtl_pci_init expect zero to be returned on error. Returning
the error code leads to, amongst other things, divide by zero panics
attempting to use the ring size that is set to zero.
Signed-off-by: NSimon Graham <simon.graham@virtualcomputer.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Handle previous allocation failures when freeing device memory
Signed-off-by: NSimon Graham <simon.graham@virtualcomputer.com>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Memory allocated by ieee80211_alloc_hw() will get orphaned
if any subsequent initializations fail.

Also don't pci_set_drvdata(pdev, NULL) until just before disabling
the PCI device. Functions called by rtl_deinit_core(hw) may eventually need
the context (when its actually implemented).

Cc: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Cc: John W. Linville <linville@tuxdriver.com>
Cc: linux-wireless@vger.kernel.org
Cc: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NTim Gardner <tim.gardner@canonical.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

the return value from _rtl_pci_switch_clk_req is not used by any of its callers.
Signed-off-by: NDevendra.Naga <devendra.aaru@gmail.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch addresses a kernel bugzilla report and two recent mail threads.

The kernel bugzilla report is https://bugzilla.kernel.org/show_bug.cgi?id=42632,
which reports a udev timeout on boot.

The first mail thread, which was on LKML (http://lkml.indiana.edu/hypermail/
linux/kernel/1112.3/00965.html) was for a WARNING that occurs after a
suspend/resume cycle for rtl8192cu.

The scond mail thread (http://marc.info/?l=linux-wireless&m=132655490826766&w=2)
concerned changes in udev that break drivers that delay while firmware is loaded
on modprobe.

This patch converts all rtlwifi-based drivers to use the asynchronous firmware
loading mechanism. Drivers rtl8192ce, rtl8192cu and rtl8192de share a common
callback routine. Driver rtl8192se needs different handling of the firmware,
thus it has its own code.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Although the rtlwifi family of devices contains 11 copies of the pr_fmt
macro, the macro is not defined for all routines that need it. By moving
the macro to wifi.h, a single copy is available for all routines.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In https://bugzilla.redhat.com/show_bug.cgi?id=771656, a kernel bug was
triggered due to a failed skb allocation that was not checked. This event
lead to an audit of all memory allocations in the complete rtlwifi family
of drivers. This patch fixes the rest.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Consolidate printks to avoid possible message interleaving
and reduce the object size.

Remove unnecessary RT_ASSERT parentheses.
Align arguments.
Coalesce formats.
Remove unnecessary __func__ use as the macro uses it.

$ size drivers/net/wireless/rtlwifi/built-in.o*
   text	   data	    bss	    dec	    hex	filename
 588901	  55333	 127216	 771450	  bc57a	drivers/net/wireless/rtlwifi/built-in.o.new
 590002	  55333	 127560	 772895	  bcb1f	drivers/net/wireless/rtlwifi/built-in.o.old
Signed-off-by: NJoe Perches <joe@perches.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Consolidate printks to avoid possible message interleaving
and reduce the object size.

Remove unnecessary RT_TRACE parentheses.

Miscellaneous typo and grammar fixes.
Add missing newlines to formats.
Remove duplicate KERN_DEBUG prefixes.
Coalesce formats.
Align arguments.

$ size drivers/net/wireless/rtlwifi/built-in.o*
   text    data     bss     dec     hex filename
 594841   55333  129680  779854   be64e drivers/net/wireless/rtlwifi/built-in.o.new
 607022   55333  138720  801075   c3933 drivers/net/wireless/rtlwifi/built-in.o.old
Signed-off-by: NJoe Perches <joe@perches.com>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

DaveM said:
   Please, this kind of stuff rots forever and not using bool properly
   drives me crazy.

Joe Perches <joe@perches.com> gave me the spatch script:

	@@
	bool b;
	@@
	-b = 0
	+b = false
	@@
	bool b;
	@@
	-b = 1
	+b = true

I merely installed coccinelle, read the documentation and took credit.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Leaving leisure power save mode can take some time, so it's better to
perform that action in process context with interrupts enabled. This
patch changes lps_leave tasklet to work.
Reported-by: NPhilipp Dreimann <philipp@dreimann.net>
Tested-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Mike McCormack <mikem@ring3k.org>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: NStanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The recent discussion regarding an incorrect return of IRQ_HANDLED
from rt2800pci caused me to look at this PCI interrupt routine. I
discovered that changes were needed.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The interrupt routine for PCI devices has a special exit that
executes the same instructions as does the normal exit.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

These were getting the macros from an implicit module.h
include via device.h, but we are planning to clean that up.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

drivers/net: Add export.h to wireless/brcm80211/brcmfmac/bcmsdh.c

This relatively recently added file uses EXPORT_SYMBOL and hence
needs export.h included so that it is compatible with the module.h
split up work.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NJustin P. Mattock <justinmattock@gmail.com>
Signed-off-by: NVitaliy Ivanov <vitalivanov@gmail.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The PCIe driver used raw reads and writes on the PCIe hardware.
As all of these are only affecting the configuration space, all of
then can be converted to pci_{read,write}_config_XX calls.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This driver uses information from the self member of the pci_bus struct to
get information regarding the bridge to which the PCIe device is attached.
Unfortunately, this member is not established on all architectures, which
leads to a kernel oops.

Skipping the entire block that uses the self member to determine the bridge
vendor will only affect RTL8192DE devices as that driver sets the ASPM support
flag differently when the bridge vendor is Intel. If the self member is
available, there is no functional change.

This patch fixes Bugzilla No. 40212.
Reported-by: NHubert Liao <liao.hubertt@gmail.com>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@kernel.org> [back to 2.6.38]
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This should be unnecessary if synchronize_irq is used.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Use PCI_VENDOR_ID_* from pci_ids.h instead of creating #define locally.
Signed-off-by: NJon Mason <jdmason@kudzu.us>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The PCIE capability offset is saved during PCI bus walking.  It will
remove an unnecessary search in the PCI configuration space if this
value is referenced instead of reacquiring it.

Also, remove unnecessary and unused #defines for PCI.
Signed-off-by: NJon Mason <jdmason@kudzu.us>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

offchan_deley should be offchan_delay
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

There are two devices with PCI ID 0x10ec:0x8192, namely RTL8192E and
RTL8192SE. The method of distinguishing them is by the revision ID
at offset 0x8 of the PCI configuration space. If the value is 0x10,
then the device uses rtl8192se for a driver.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Once we realize a bad packet was received, don't
waste time unmapping it, freeing it, then allocation
a new skb and mapping it, just resubmit the existing
skb.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In the case we fail to allocate a new skb, the old skb should
be resubmitted unmodified.

Fixes bug introduced in a9e12869.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Acked-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Should pass along packet if there's no CRC and no hardware error.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Should pass along packet if there's no CRC and no hardware error.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Ownership of an rx buffer should only be given to the hardware
after all other changes are written, otherwise there's
a potential race.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This removes the need to use IRQ safe spinlocks in many places.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

_rtl_pci_rx_interrupt uses rtlpci->rx_ring[rx_queue_idx].idx a
few times, so store it in a separate variable.
Signed-off-by: NMike McCormack <mikem@ring3k.org>
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Although a previous fix handles the kernel panics that result from
failure to allocate a new RX buffer, memory fragmentation can be
reduced if the amsdu_8k capability is disabled as new buffers need only
be of O(0), not O(2).
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>