linux-stable.git/drivers/net/ethernet, branch v3.4.76

net: smc91: fix crash regression on the versatile

2013-12-12T06:34:11+00:00

commit a0c20fb02592d372e744d1d739cda3e1b3defaae upstream.

After commit e9e4ea74f06635f2ffc1dffe5ef40c854faa0a90
"net: smc91x: dont't use SMC_outw for fixing up halfword-aligned data"
The Versatile SMSC LAN91C111 is crashing like this:

------------[ cut here ]------------
kernel BUG at /home/linus/linux/drivers/net/ethernet/smsc/smc91x.c:599!
Internal error: Oops - BUG: 0 [#1] ARM
Modules linked in:
CPU: 0 PID: 43 Comm: udhcpc Not tainted 3.13.0-rc1+ #24
task: c6ccfaa0 ti: c6cd0000 task.ti: c6cd0000
PC is at smc_hardware_send_pkt+0x198/0x22c
LR is at smc_hardware_send_pkt+0x24/0x22c
pc : []    lr : []    psr: 20000013
sp : c6cd1d08  ip : 00000001  fp : 00000000
r10: c02adb08  r9 : 00000000  r8 : c6ced802
r7 : c786fba0  r6 : 00000146  r5 : c8800000  r4 : c78d6000
r3 : 0000000f  r2 : 00000146  r1 : 00000000  r0 : 00000031
Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 0005317f  Table: 06cf4000  DAC: 00000015
Process udhcpc (pid: 43, stack limit = 0xc6cd01c0)
Stack: (0xc6cd1d08 to 0xc6cd2000)
1d00:                   00000010 c8800000 c78d6000 c786fba0 c78d6000 c01be868
1d20: c01be7a4 00004000 00000000 c786fba0 c6c12b80 c0208554 000004d0 c780fc60
1d40: 00000220 c01fb734 00000000 00000000 00000000 c6c9a440 c6c12b80 c78d6000
1d60: c786fba0 c6c9a440 00000000 c021d1d8 00000000 00000000 c6c12b80 c78d6000
1d80: c786fba0 00000001 c6c9a440 c02087f8 c6c9a4a0 00080008 00000000 00000000
1da0: c78d6000 c786fba0 c78d6000 00000138 00000000 00000000 00000000 00000000
1dc0: 00000000 c027ba74 00000138 00000138 00000001 00000010 c6cedc00 00000000
1de0: 00000008 c7404400 c6cd1eec c6cd1f14 c067a73c c065c0b8 00000000 c067a740
1e00: 01ffffff 002040d0 00000000 00000000 00000000 00000000 00000000 ffffffff
1e20: 43004400 00110022 c6cdef20 c027ae8c c6ccfaa0 be82d65c 00000014 be82d3cc
1e40: 00000000 00000000 00000000 c01f2870 00000000 00000000 00000000 c6cd1e88
1e60: c6ccfaa0 00000000 00000000 00000000 00000000 00000000 00000000 00000000
1e80: 00000000 00000000 00000031 c7802310 c7802300 00000138 c7404400 c0771da0
1ea0: 00000000 c6cd1eec c7800340 00000138 be82d65c 00000014 be82d3cc c6cd1f08
1ec0: 00000014 00000000 c7404400 c7404400 00000138 c01f4628 c78d6000 00000000
1ee0: 00000000 be82d3cc 00000138 c6cd1f08 00000014 c6cd1ee4 00000001 00000000
1f00: 00000000 00000000 00080011 00000002 06000000 ffffffff 0000ffff 00000002
1f20: 06000000 ffffffff 0000ffff c00928c8 c065c520 c6cd1f58 00000003 c009299c
1f40: 00000003 c065c520 c7404400 00000000 c7404400 c01f2218 c78106b0 c7441cb0
1f60: 00000000 00000006 c06799fc 00000000 00000000 00000006 00000000 c01f3ee0
1f80: 00000000 00000000 be82d678 be82d65c 00000014 00000001 00000122 c00139c8
1fa0: c6cd0000 c0013840 be82d65c 00000014 00000006 be82d3cc 00000138 00000000
1fc0: be82d65c 00000014 00000001 00000122 00000000 00000000 00018cb1 00000000
1fe0: 00003801 be82d3a8 0003a0c7 b6e9af08 60000010 00000006 00000000 00000000
[] (smc_hardware_send_pkt+0x198/0x22c) from [] (smc_hard_start_xmit+0xc4/0x1e8)
[] (smc_hard_start_xmit+0xc4/0x1e8) from [] (dev_hard_start_xmit+0x460/0x4cc)
[] (dev_hard_start_xmit+0x460/0x4cc) from [] (sch_direct_xmit+0x94/0x18c)
[] (sch_direct_xmit+0x94/0x18c) from [] (dev_queue_xmit+0x238/0x42c)
[] (dev_queue_xmit+0x238/0x42c) from [] (packet_sendmsg+0xbe8/0xd28)
[] (packet_sendmsg+0xbe8/0xd28) from [] (sock_sendmsg+0x84/0xa8)
[] (sock_sendmsg+0x84/0xa8) from [] (SyS_sendto+0xb8/0xdc)
[] (SyS_sendto+0xb8/0xdc) from [] (ret_fast_syscall+0x0/0x2c)
Code: e3130002 1a000001 e3130001 0affffcd (e7f001f2)
---[ end trace 81104fe70e8da7fe ]---
Kernel panic - not syncing: Fatal exception in interrupt

This is because the macro operations in smc91x.h defined
for Versatile are missing SMC_outsw() as used in this
commit.

The Versatile needs and uses the same accessors as the other
platforms in the first if(...) clause, just switch it to using
that and we have one problem less to worry about.

This includes a hunk of a patch from Will Deacon fixin
the other 32bit platforms as well: Innokom, Ramses, PXA,
PCM027.

Checkpatch complains about spacing, but I have opted to
follow the style of this .h-file.

Cc: Russell King 
Cc: Nicolas Pitre 
Cc: Eric Miao 
Cc: Jonathan Cameron 
Signed-off-by: Will Deacon 
Signed-off-by: Linus Walleij 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

8139cp: re-enable interrupts after tx timeout

2013-11-29T18:50:32+00:00

commit 01ffc0a7f1c1801a2354719dedbc32aff45b987d upstream.

Recovery doesn't work too well if we leave interrupts disabled...

Signed-off-by: David Woodhouse 
Acked-by: Francois Romieu 
Signed-off-by: David S. Miller 
Cc: Luis Henriques 
Signed-off-by: Greg Kroah-Hartman

davinci_emac.c: Fix IFF_ALLMULTI setup

2013-11-04T12:23:41+00:00

[ Upstream commit d69e0f7ea95fef8059251325a79c004bac01f018 ]

When IFF_ALLMULTI flag is set on interface and IFF_PROMISC isn't,
emac_dev_mcast_set should only enable RX of multicasts and reset
MACHASH registers.

It does this, but afterwards it either sets up multicast MACs
filtering or disables RX of multicasts and resets MACHASH registers
again, rendering IFF_ALLMULTI flag useless.

This patch fixes emac_dev_mcast_set, so that multicast MACs filtering and
disabling of RX of multicasts are skipped when IFF_ALLMULTI flag is set.

Tested with kernel 2.6.37.

Signed-off-by: Mariusz Ceier 
Acked-by: Mugunthan V N 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

bnx2x: record rx queue for LRO packets

2013-11-04T12:23:41+00:00

[ Upstream commit 60e66fee56b2256dcb1dc2ea1b2ddcb6e273857d ]

RPS support is kind of broken on bnx2x, because only non LRO packets
get proper rx queue information. This triggers reorders, as it seems
bnx2x like to generate a non LRO packet for segment including TCP PUSH
flag : (this might be pure coincidence, but all the reorders I've
seen involve segments with a PUSH)

11:13:34.335847 IP A > B: . 415808:447136(31328) ack 1 win 457 
11:13:34.335992 IP A > B: . 447136:448560(1424) ack 1 win 457 
11:13:34.336391 IP A > B: . 448560:479888(31328) ack 1 win 457 
11:13:34.336425 IP A > B: P 511216:512640(1424) ack 1 win 457 
11:13:34.336423 IP A > B: . 479888:511216(31328) ack 1 win 457 
11:13:34.336924 IP A > B: . 512640:543968(31328) ack 1 win 457 
11:13:34.336963 IP A > B: . 543968:575296(31328) ack 1 win 457 

We must call skb_record_rx_queue() to properly give to RPS (and more
generally for TX queue selection on forward path) the receive queue
information.

Similar fix is needed for skb_mark_napi_id(), but will be handled
in a separate patch to ease stable backports.

Signed-off-by: Eric Dumazet 
Cc: Willem de Bruijn 
Cc: Eilon Greenstein 
Acked-by: Dmitry Kravkov 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

net: mv643xx_eth: fix orphaned statistics timer crash

2013-11-04T12:23:40+00:00

[ Upstream commit f564412c935111c583b787bcc18157377b208e2e ]

The periodic statistics timer gets started at port _probe() time, but
is stopped on _stop() only. In a modular environment, this can cause
the timer to access already deallocated memory, if the module is unloaded
without starting the eth device. To fix this, we add the timer right
before the port is started, instead of at _probe() time.

Signed-off-by: Sebastian Hesselbarth 
Acked-by: Jason Cooper 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

net: mv643xx_eth: update statistics timer from timer context only

2013-11-04T12:23:40+00:00

[ Upstream commit 041b4ddb84989f06ff1df0ca869b950f1ee3cb1c ]

Each port driver installs a periodic timer to update port statistics
by calling mib_counters_update. As mib_counters_update is also called
from non-timer context, we should not reschedule the timer there but
rather move it to timer-only context.

Signed-off-by: Sebastian Hesselbarth 
Acked-by: Jason Cooper 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

ll_temac: Reset dma descriptors indexes on ndo_open

2013-10-13T22:42:48+00:00

[ Upstream commit 7167cf0e8cd10287b7912b9ffcccd9616f382922 ]

The dma descriptors indexes are only initialized on the probe function.

If a packet is on the buffer when temac_stop is called, the dma
descriptors indexes can be left on a incorrect state where no other
package can be sent.

So an interface could be left in an usable state after ifdow/ifup.

This patch makes sure that the descriptors indexes are in a proper
status when the device is open.

Signed-off-by: Ricardo Ribalda Delgado 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

via-rhine: fix VLAN priority field (PCP, IEEE 802.1p)

2013-10-13T22:42:48+00:00

[ Upstream commit 207070f5221e2a901d56a49df9cde47d9b716cd7 ]

Outgoing packets sent by via-rhine have their VLAN PCP field off by one
(when hardware acceleration is enabled). The TX descriptor expects only VID
and PCP (without a CFI/DEI bit).

Peter Boström noticed and reported the bug.

Signed-off-by: Roger Luethi 
Cc: Peter Boström 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

gianfar: Change default HW Tx queue scheduling mode

2013-10-13T22:42:47+00:00

commit b98b8babd6e3370fadb7c6eaacb00eb2f6344a6c upstream.

This is primarily to address transmission timeout occurrences, when
multiple H/W Tx queues are being used concurrently. Because in
the priority scheduling mode the controller does not service the
Tx queues equally (but in ascending index order), Tx timeouts are
being triggered rightaway for a basic test with multiple simultaneous
connections like:
iperf -c  -n 100M -P 8

resulting in kernel trace:
NETDEV WATCHDOG: eth1 (fsl-gianfar): transmit queue  timed out
------------[ cut here ]------------
WARNING: at net/sched/sch_generic.c:255
...
and controller reset during intense traffic, and possibly further
complications.

This patch changes the default H/W Tx scheduling setting (TXSCHED)
for multi-queue devices, from priority scheduling mode to a weighted
round robin mode with equal weights for all H/W Tx queues, and
addresses the issue above.

Signed-off-by: Claudiu Manoil 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

sfc: Fix efx_rx_buf_offset() for recycled pages

2013-10-01T16:10:52+00:00

This bug fix is only for stable branches older than 3.10.  The bug was
fixed upstream by commit 2768935a4660 ('sfc: reuse pages to avoid DMA
mapping/unmapping costs'), but that change is totally unsuitable for
stable.

Commit b590ace09d51 ('sfc: Fix efx_rx_buf_offset() in the presence of
swiotlb') added an explicit page_offset member to struct
efx_rx_buffer, which must be set consistently with the u.page and
dma_addr fields.  However, it failed to add the necessary assignment
in efx_resurrect_rx_buffer().  It also did not correct the calculation
of efx_rx_buffer::dma_addr in efx_resurrect_rx_buffer(), which assumes
that DMA-mapping a page will result in a page-aligned DMA address
(exactly what swiotlb violates).

Add the assignment of efx_rx_buffer::page_offset and change the
calculation of dma_addr to make use of it.

Signed-off-by: Ben Hutchings 
Acked-by: David Miller 
Signed-off-by: Greg Kroah-Hartman