The addition of RX event merging support means we don't reliably
detect dropped RX events now. Currently we will only detect them if
the previous event for the RX queue had the CONT bit set.
Only accept RX completion events as merged if the
GET_CAPABILITIES_OUT_RX_BATCHING bit is set in datapath_caps (which it
won't be for the low-latency datapath) and the CONT bit is not set on
the event.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
To run BISTs the MC goes down in to a special mode where it will only
respond to MCDI from the testing PF, and TX, RX and event queues are
torn down. Other PFs get a message as it goes down to tell them it's
going down.
When the other PFs get this message, they check the soft status
register to tell when the MC has rebooted after BIST mode and they can
start recovery.
[bwh: Convert the test result to 1 or -1 as for earlier NICs]
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
The MC firmware is cooperatively multitasking and its scheduler will
send an event when a task yields after running for more than the
expected maximum time. This can be useful for firmware development
but does not usually indicate a serious error and does not help to
detect a lockup (there is a hardware watchdog that does that).
Change the message and reduce log level accordingly.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
There is an as-yet unexplained bug that sometimes prevents (or delays)
the driver seeing the completion event for a completed MCDI request on
the SFC9120. The requested configuration change will have happened
but the driver assumes it to have failed, and this can result in
further failures. We can mitigate this by polling for completion
after unsuccessfully waiting for an event.
Fixes: 8127d661e7 ('sfc: Add support for Solarflare SFC9100 family')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
rx_prefix_size is 4-bytes aligned on Falcon/Siena (16 bytes), but it is equal
to 14 on EF10. So, it should be taken into account if arch requires IP header
to be 4-bytes aligned (via NET_IP_ALIGN).
Fixes: 8127d661e7 ('sfc: Add support for Solarflare SFC9100 family')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
There is a single MCDI PTP operation for setting the frequency
adjustment and applying a time offset to the hardware clock. When
applying a time offset we should not change the frequency adjustment.
These two operations can now be requested separately but this requires
a flash firmware update. Keep using the single operation, but
remember and repeat the previous frequency adjustment.
Fixes: 7c236c43b8 ('sfc: Add support for IEEE-1588 PTP')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
This disables PTP when we bring the interface down to avoid getting
unmatched RX timestamp events, and tries to re-enable it when bringing
the interface up.
[bwh: Make efx_ptp_stop() safe on Falcon. Introduce
efx_ptp_{start,stop}_datapath() functions; we'll expand them later.]
Fixes: 7c236c43b8 ('sfc: Add support for IEEE-1588 PTP')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
In case of a flood of PTP packets, the timestamp peripheral and MC
firmware on the SFN[56]322F boards may not be able to provide
timestamp events for all packets. Don't complain too much about this.
Fixes: 7c236c43b8 ('sfc: Add support for IEEE-1588 PTP')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Limit syslog flood if a PTP packet storm occurs.
Fixes: 7c236c43b8 ('sfc: Add support for IEEE-1588 PTP')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
efx_ptp_is_ptp_tx() must be robust against skbs from raw sockets that
have invalid IPv4 and UDP headers.
Add checks that:
- the transport header has been found
- there is enough space between network and transport header offset
for an IPv4 header
- there is enough space after the transport header offset for a
UDP header
Fixes: 7c236c43b8 ('sfc: Add support for IEEE-1588 PTP')
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Use PCI standard macro dev_is_pci() instead of hardcoding.
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace local macro DFX_BUS_PCI() with PCI standard macro
dev_is_pci().
Acked-by: Maciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix spelling errors in tun driver.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several files refer to an old address for the Free Software Foundation
in the file header comment. Resolve by replacing the address with
the URL <http://www.gnu.org/licenses/> so that we do not have to keep
updating the header comments anytime the address changes.
CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Veaceslav Falico <vfalico@redhat.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: Haiyang Zhang <haiyangz@microsoft.com>
CC: "K. Y. Srinivasan" <kys@microsoft.com>
CC: Paul Mackerras <paulus@samba.org>
CC: Ian Campbell <ian.campbell@citrix.com>
CC: Wei Liu <wei.liu2@citrix.com>
CC: Rusty Russell <rusty@rustcorp.com.au>
CC: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several files refer to an old address for the Free Software Foundation
in the file header comment. Resolve by replacing the address with
the URL <http://www.gnu.org/licenses/> so that we do not have to keep
updating the header comments anytime the address changes.
CC: Oliver Neukum <oliver@neukum.org>
CC: Steve Glendinning <steve.glendinning@shawell.net>
CC: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several files refer to an old address for the Free Software Foundation
in the file header comment. Resolve by replacing the address with
the URL <http://www.gnu.org/licenses/> so that we do not have to keep
updating the header comments anytime the address changes.
CC: Samuel Ortiz <samuel@sortiz.org>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several files refer to an old address for the Free Software Foundation
in the file header comment. Resolve by replacing the address with
the URL <http://www.gnu.org/licenses/> so that we do not have to keep
updating the header comments anytime the address changes.
CC: Santosh Raspatur <santosh@chelsio.com>
CC: Dimitris Michailidis <dm@chelsio.com>
CC: Michael Chan <mchan@broadcom.com>
CC: Santiago Leon <santil@linux.vnet.ibm.com>
CC: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
CC: Olof Johansson <olof@lixom.net>
CC: Manish Chopra <manish.chopra@qlogic.com>
CC: Sony Chacko <sony.chacko@qlogic.com>
CC: Rajesh Borundia <rajesh.borundia@qlogic.com>
CC: Nicolas Pitre <nico@fluxnic.net>
CC: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When running in a network namespace whose only link to the outside
world is a macvlan device, not being able to create a macvtap off of
it is a real pain.
So modify macvtap creation to automatically forward a creation of a
macvtap on a macvlan to become a creation of a macvtap on the
underlying network device, just like is currently done with
macvlan-on-macvlan devices.
v2: Use netif_is_macvlan and macvlan_dev_real_dev helpers to make it
more clear what we're doing.
Signed-off-by: Kevin Wallace <kevin@pentabarf.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings says:
====================
SIOCGHWTSTAMP ioctl
1. Add the SIOCGHWTSTAMP ioctl and update the timestamping
documentation.
2. Implement SIOCGHWTSTAMP in most drivers that support SIOCSHWTSTAMP.
3. Add a test program to exercise SIOC{G,S}HWTSTAMP.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When driver registration fails, we need to clean up the resources allocated
before. mlx4_core missed destroying the workqueue allocated.
This patch destroys the workqueue when registration fails.
Signed-off-by: Wei Yang <weiyang@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
xenvif_start_xmit() relies on checking vif->task for NULL to determine
whether the vif is ready to accept packets. The task thread is stopped in
xenvif_disconnect() but task is not set to NULL. Thus, on a re-connect the
check will give a false positive.
Also since commit ea732dff5c (Handle backend
state transitions in a more robust way) it should not be possible for
xenvif_connect() to be called if the vif is already connected so change the
check of vif->tx_irq to a BUG_ON() and also add a BUG_ON(vif->task).
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Cc: Wei Liu <wei.liu2@citrix.com>
Cc: Ian Campbell <ian.campbell@citrix.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Few network drivers really supports frag_list : virtual drivers.
Some drivers wrongly advertise NETIF_F_FRAGLIST feature.
If skb with a frag_list is given to them, packet on the wire will be
corrupt.
Remove this flag, as core networking stack will make sure to
provide packets that can be sent without corruption.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Cc: Anirudha Sarangi <anirudh@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On tx submit the driver always dma_map_single() FEC_ENET_TX_FRSIZE (=2048)
bytes. This works because we don't overwrite any memory after the data buffer,
we remove it from cache if it was there. So we hurt performace in case the
mapping of a smaller area makes a difference.
There is also a bug: If the data area starts shortly before the end of
RAM say 0xc7fffa10 and the RAM ends at 0xc8000000 then we have enough
space to fit the data area (according to skb->len) but we would map beyond
end of ram if we are using 2048. In v2.6.31 (against which kernel this patch
made) there is the following check in dma_cache_maint():
|BUG_ON(!virt_addr_valid(start) || !virt_addr_valid(start + size - 1));
Since the area starting at 0xc8000000 is no longer virt_addr_valid() we
BUG() during dma_map_single(). The BUG() statement was removed in v3.5-rc1 as
per 2dc6a016 ("ARM: dma-mapping: use asm-generic/dma-mapping-common.h").
This patch was tested on v2.6.31 and then forward-ported and compile
tested only against the net tree. I think it is still worth fixing
mainline even after the BUG() statement is gone.
Tested-by: Fugang Duan <B38611@freescale.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Fugang Duan <B38611@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When only one port of the two port is pinned out, then dt probe is failing
because second port phy is not found. fixing this by checking the number of
slaves and breaking the loop.
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Modify tg3_chip_reset() and tg3_close() to check if the PCI network
adapter device is accessible at all in order to skip poking it or
trying to handle a carrier loss in vain when that's not the case.
Introduce a special PCI helper function pci_device_is_present()
for this purpose.
Of course, this uncovers the lack of the appropriate RTNL locking
in tg3_suspend() and tg3_resume(), so add that locking in there
too.
These changes prevent tg3 from burning a CPU at 100% load level for
solid several seconds after the Thunderbolt link is disconnected from
a Matrox DS1 docking station.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Building brcmfmac for sparc64 gave the following warning:
CC [M] drivers/net/wireless/brcm80211/brcmfmac/bcmsdh_sdmmc.o
bcmsdh_sdmmc.c: In function 'brcmf_sdioh_request_byte':
bcmsdh_sdmmc.c:89:6: warning: 'err_ret' may be used uninitialized
in this function [-Wuninitialized]
Inspecting the code it indeed had a path of execution in
which the return value was used uninitialized. This patch
fixes that code path.
Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Reviewed-by: Franky Lin <frankyl@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
If kmalloc fails wcn36xx_smd_rsp_process will attempt to dereference
a NULL pointer. There might be a better error recovery then just
printing an error, but printing an error message is better then the
current behaviour.
Signed-off-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Code in ath9k_hw_set_clockrate function indicates that ah->curchan
(and thus chan local variable) may be NULL. If that is indeed the
case, IS_CHAN_HT40(chan) check has to be performed only in branch
where chan is not NULL. Moving the code under already existing
if condition fixes this issue.
Signed-off-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The EEPROM parameter to determine whether the bias
strength values for XLNA have to be applied is part
of the miscConfiguration field and not featureEnable.
Cc: stable@vger.kernel.org
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Bit 5 in the miscConfiguration field of the base EEPROM
header denotes whether QuickDrop is enabled or not. Fix
the incorrect usage of BIT(1) and also make sure that
this is done only for the required chips.
Cc: stable@vger.kernel.org
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Pull networking updates from David Miller:
"Here is a pile of bug fixes that accumulated while I was in Europe"
1) In fixing kernel leaks to userspace during copying of socket
addresses, we broke a case that used to work, namely the user
providing a buffer larger than the in-kernel generic socket address
structure. This broke Ruby amongst other things. Fix from Dan
Carpenter.
2) Fix regression added by byte queue limit support in 8139cp driver,
from Yang Yingliang.
3) The addition of MSG_SENDPAGE_NOTLAST buggered up a few sendpage
implementations, they should just treat it the same as MSG_MORE.
Fix from Richard Weinberger and Shawn Landden.
4) Handle icmpv4 errors received on ipv6 SIT tunnels correctly, from
Oussama Ghorbel. In particular we should send an ICMPv6 unreachable
in such situations.
5) Fix some regressions in the recent genetlink fixes, in particular
get the pmcraid driver to use the new safer interfaces correctly.
From Johannes Berg.
6) macvtap was converted to use a per-cpu set of statistics, but some
code was still bumping tx_dropped elsewhere. From Jason Wang.
7) Fix build failure of xen-netback due to missing include on some
architectures, from Andy Whitecroft.
8) macvtap double counts received packets in statistics, fix from Vlad
Yasevich.
9) Fix various cases of using *_STATS_BH() when *_STATS() is more
appropriate. From Eric Dumazet and Hannes Frederic Sowa.
10) Pktgen ipsec mode doesn't update the ipv4 header length and checksum
properly after encapsulation. Fix from Fan Du.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (61 commits)
net/mlx4_en: Remove selftest TX queues empty condition
{pktgen, xfrm} Update IPv4 header total len and checksum after tranformation
virtio_net: make all RX paths handle erors consistently
virtio_net: fix error handling for mergeable buffers
virtio_net: Fixed a trivial typo (fitler --> filter)
netem: fix gemodel loss generator
netem: fix loss 4 state model
netem: missing break in ge loss generator
net/hsr: Support iproute print_opt ('ip -details ...')
net/hsr: Very small fix of comment style.
MAINTAINERS: Added net/hsr/ maintainer
ipv6: fix possible seqlock deadlock in ip6_finish_output2
ixgbe: Make ixgbe_identify_qsfp_module_generic static
ixgbe: turn NETIF_F_HW_L2FW_DOFFLOAD off by default
ixgbe: ixgbe_fwd_ring_down needs to be static
e1000: fix possible reset_task running after adapter down
e1000: fix lockdep warning in e1000_reset_task
e1000: prevent oops when adapter is being closed and reset simultaneously
igb: Fixed Wake On LAN support
inet: fix possible seqlock deadlocks
...
Remove waiting for TX queues to become empty during selftest.
This check is not necessary for any purpose, and might put
the driver into an infinite loop.
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
receive mergeable now handles errors internally.
Do same for big and small packet paths, otherwise
the logic is too hard to follow.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet noticed that if we encounter an error
when processing a mergeable buffer, we don't
dequeue all of the buffers from this packet,
the result is almost sure to be loss of networking.
Jason Wang noticed that we also leak a page and that we don't decrement
the rq buf count, so we won't repost buffers (a resource leak).
Fix both issues.
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Michael Dalton <mwdalton@google.com>
Reported-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
"MAC filter" sounds more reasonable than "MAC fitler".
Signed-off-by: Thomas Huth <thuth@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Correct a namespace complaint by making the function static
and moving the prototype into the .c file.
Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
NETIF_F_HW_L2FW_DOFFLOAD allows upper layer net devices such
as macvlan to use queues in the hardware to directly submit and
receive skbs.
This creates a subtle change in the datapath though. One change
being the skb may no longer use the root devices qdisc.
Because users may not expect this we can't enable the feature
by default unless the hardware can offload all the software
functionality above it. So for now disable it by default and
let users opt in.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
When compiling with -Wstrict-prototypes gcc catches a static
I missed.
./ixgbe_main.c:4254: warning: no previous prototype for 'ixgbe_fwd_ring_down'
Reported-by: Phillip Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
On e1000_down(), we should ensure every asynchronous work is canceled
before proceeding. Since the watchdog_task can schedule other works
apart from itself, it should be stopped first, but currently it is
stopped after the reset_task. This can result in the following race
leading to the reset_task running after the module unload:
e1000_down_and_stop(): e1000_watchdog():
---------------------- -----------------
cancel_work_sync(reset_task)
schedule_work(reset_task)
cancel_delayed_work_sync(watchdog_task)
The patch moves cancel_delayed_work_sync(watchdog_task) at the beginning
of e1000_down_and_stop() thus ensuring the race is impossible.
Cc: Tushar Dave <tushar.n.dave@intel.com>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The patch fixes the following lockdep warning, which is 100%
reproducible on network restart:
======================================================
[ INFO: possible circular locking dependency detected ]
3.12.0+ #47 Tainted: GF
-------------------------------------------------------
kworker/1:1/27 is trying to acquire lock:
((&(&adapter->watchdog_task)->work)){+.+...}, at: [<ffffffff8108a5b0>] flush_work+0x0/0x70
but task is already holding lock:
(&adapter->mutex){+.+...}, at: [<ffffffffa0177c0a>] e1000_reset_task+0x4a/0xa0 [e1000]
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (&adapter->mutex){+.+...}:
[<ffffffff810bdb5d>] lock_acquire+0x9d/0x120
[<ffffffff816b8cbc>] mutex_lock_nested+0x4c/0x390
[<ffffffffa017233d>] e1000_watchdog+0x7d/0x5b0 [e1000]
[<ffffffff8108b972>] process_one_work+0x1d2/0x510
[<ffffffff8108ca80>] worker_thread+0x120/0x3a0
[<ffffffff81092c1e>] kthread+0xee/0x110
[<ffffffff816c3d7c>] ret_from_fork+0x7c/0xb0
-> #0 ((&(&adapter->watchdog_task)->work)){+.+...}:
[<ffffffff810bd9c0>] __lock_acquire+0x1710/0x1810
[<ffffffff810bdb5d>] lock_acquire+0x9d/0x120
[<ffffffff8108a5eb>] flush_work+0x3b/0x70
[<ffffffff8108b5d8>] __cancel_work_timer+0x98/0x140
[<ffffffff8108b693>] cancel_delayed_work_sync+0x13/0x20
[<ffffffffa0170cec>] e1000_down_and_stop+0x3c/0x60 [e1000]
[<ffffffffa01775b1>] e1000_down+0x131/0x220 [e1000]
[<ffffffffa0177c12>] e1000_reset_task+0x52/0xa0 [e1000]
[<ffffffff8108b972>] process_one_work+0x1d2/0x510
[<ffffffff8108ca80>] worker_thread+0x120/0x3a0
[<ffffffff81092c1e>] kthread+0xee/0x110
[<ffffffff816c3d7c>] ret_from_fork+0x7c/0xb0
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&adapter->mutex);
lock((&(&adapter->watchdog_task)->work));
lock(&adapter->mutex);
lock((&(&adapter->watchdog_task)->work));
*** DEADLOCK ***
3 locks held by kworker/1:1/27:
#0: (events){.+.+.+}, at: [<ffffffff8108b906>] process_one_work+0x166/0x510
#1: ((&adapter->reset_task)){+.+...}, at: [<ffffffff8108b906>] process_one_work+0x166/0x510
#2: (&adapter->mutex){+.+...}, at: [<ffffffffa0177c0a>] e1000_reset_task+0x4a/0xa0 [e1000]
stack backtrace:
CPU: 1 PID: 27 Comm: kworker/1:1 Tainted: GF 3.12.0+ #47
Hardware name: System manufacturer System Product Name/P5B-VM SE, BIOS 0501 05/31/2007
Workqueue: events e1000_reset_task [e1000]
ffffffff820f6000 ffff88007b9dba98 ffffffff816b54a2 0000000000000002
ffffffff820f5e50 ffff88007b9dbae8 ffffffff810ba936 ffff88007b9dbac8
ffff88007b9dbb48 ffff88007b9d8f00 ffff88007b9d8780 ffff88007b9d8f00
Call Trace:
[<ffffffff816b54a2>] dump_stack+0x49/0x5f
[<ffffffff810ba936>] print_circular_bug+0x216/0x310
[<ffffffff810bd9c0>] __lock_acquire+0x1710/0x1810
[<ffffffff8108a5b0>] ? __flush_work+0x250/0x250
[<ffffffff810bdb5d>] lock_acquire+0x9d/0x120
[<ffffffff8108a5b0>] ? __flush_work+0x250/0x250
[<ffffffff8108a5eb>] flush_work+0x3b/0x70
[<ffffffff8108a5b0>] ? __flush_work+0x250/0x250
[<ffffffff8108b5d8>] __cancel_work_timer+0x98/0x140
[<ffffffff8108b693>] cancel_delayed_work_sync+0x13/0x20
[<ffffffffa0170cec>] e1000_down_and_stop+0x3c/0x60 [e1000]
[<ffffffffa01775b1>] e1000_down+0x131/0x220 [e1000]
[<ffffffffa0177c12>] e1000_reset_task+0x52/0xa0 [e1000]
[<ffffffff8108b972>] process_one_work+0x1d2/0x510
[<ffffffff8108b906>] ? process_one_work+0x166/0x510
[<ffffffff8108ca80>] worker_thread+0x120/0x3a0
[<ffffffff8108c960>] ? manage_workers+0x2c0/0x2c0
[<ffffffff81092c1e>] kthread+0xee/0x110
[<ffffffff81092b30>] ? __init_kthread_worker+0x70/0x70
[<ffffffff816c3d7c>] ret_from_fork+0x7c/0xb0
[<ffffffff81092b30>] ? __init_kthread_worker+0x70/0x70
== The issue background ==
The problem occurs, because e1000_down(), which is called under
adapter->mutex by e1000_reset_task(), tries to synchronously cancel
e1000 auxiliary works (reset_task, watchdog_task, phy_info_task,
fifo_stall_task), which take adapter->mutex in their handlers. So the
question is what does adapter->mutex protect there?
The adapter->mutex was introduced by commit 0ef4ee ("e1000: convert to
private mutex from rtnl") as a replacement for rtnl_lock() taken in the
asynchronous handlers. It targeted on fixing a similar lockdep warning
issued when e1000_down() was called under rtnl_lock(), and it fixed it,
but unfortunately it introduced the lockdep warning described above.
Anyway, that said the source of this bug is that the asynchronous works
were made to take rtnl_lock() some time ago, so let's look deeper and
find why it was added there.
The rtnl_lock() was added to asynchronous handlers by commit 338c15
("e1000: fix occasional panic on unload") in order to prevent
asynchronous handlers from execution after the module is unloaded
(e1000_down() is called) as it follows from the comment to the commit:
> Net drivers in general have an issue where timers fired
> by mod_timer or work threads with schedule_work are running
> outside of the rtnl_lock.
>
> With no other lock protection these routines are vulnerable
> to races with driver unload or reset paths.
>
> The longer term solution to this might be a redesign with
> safer locks being taken in the driver to guarantee no
> reentrance, but for now a safe and effective fix is
> to take the rtnl_lock in these routines.
I'm not sure if this locking scheme fixed the problem or just made it
unlikely, although I incline to the latter. Anyway, this was long time
ago when e1000 auxiliary works were implemented as timers scheduling
real work handlers in their routines. The e1000_down() function only
canceled the timers, but left the real handlers running if they were
running, which could result in work execution after module unload.
Today, the e1000 driver uses sane delayed works instead of the pair
timer+work to implement its delayed asynchronous handlers, and the
e1000_down() synchronously cancels all the works so that the problem
that commit 338c15 tried to cope with disappeared, and we don't need any
locks in the handlers any more. Moreover, any locking there can
potentially result in a deadlock.
So, this patch reverts commits 0ef4ee and 338c15.
Fixes: 0ef4eedc2e ("e1000: convert to private mutex from rtnl")
Fixes: 338c15e470 ("e1000: fix occasional panic on unload")
Cc: Tushar Dave <tushar.n.dave@intel.com>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This change is based on a similar change made to e1000e support in
commit bb9e44d0d0 ("e1000e: prevent oops when adapter is being closed
and reset simultaneously"). The same issue has also been observed
on the older e1000 cards.
Here, we have increased the RESET_COUNT value to 50 because there are too
many accesses to e1000 nic on stress tests to e1000 nic, it is not enough
to set RESET_COUT 25. Experimentation has shown that it is enough to set
RESET_COUNT 50.
Signed-off-by: yzhu1 <yanjun.zhu@windriver.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
