Count the number of TX doorbells we issue to the qdio layer.
Also count the number of actual frames in a TX buffer, and then
use this data along with the byte count during TX completion.
We'll make additional use of the frame count in a subsequent patch.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We're down to a single bit flag for MAC-address related status, reflect
that in the info struct.
Also set up the flag during initialization instead of clearing it during
shutdown - one more little step towards unifying the shutdown code.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The logic that deals with errors from qeth_l3_get_unique_id() is quite
complex: it sets card->unique_id to 0xfffe, additionally flags it as
UNIQUE_ID_NOT_BY_CARD and later takes this flag as cue to not propagate
card->unique_id to dev->dev_id. With dev->dev_id thus holding 0,
addrconf_ifid_eui48() applies its default behaviour.
Get rid of all the special bit masks, and just return the old uid in
case of an error. For the vast majority of cases this will be 0 (and so
we still get the desired default behaviour) - with the rare exception
where qeth_l3_get_unique_id() might have been called earlier but the
initialization then failed at a later point.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When the support for polling drivers was initially added, it only
considered Input Queue 0. But as QDIO interrupts are actually for the
full device and not a single queue, this doesn't really fit for
configurations where multiple Input Queues are used.
Rework the qdio code so that interrupts for a polling driver are not
split up into actions for each queue. Instead deliver the interrupt as
a single event, and let the driver decide which queue needs what action.
When re-enabling the QDIO interrupt via qdio_start_irq(), this means
that the qdio code needs to
(1) put _all_ eligible queues back into a state where they raise IRQs,
(2) and afterwards check _all_ eligible queues for new work to bridge
the race window.
On the qeth side of things (as the only qdio polling driver), we can now
add CQ polling support to the main NAPI poll routine. It doesn't consume
NAPI budget, and to avoid hogging the CPU we yield control after
completing one full queue worth of buffers.
The subsequent qdio_start_irq() will check for any additional work, and
have us re-schedule the NAPI instance accordingly.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Whenever all completed RX buffers have been processed
(ie. rx->b_count == 0), we call down to the HW layer to scan for
additional buffers. If no further buffers are available, the code
breaks out of the while-loop.
So we never reach the 'process an RX buffer' step with rx->b_count == 0,
eliminate that check and one level of indentation.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The main NAPI poll routine should eventually handle more types of work,
beyond just the RX ring.
Split off the RX poll logic into a separate function, and simplify the
nested while-loop.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since RX buffers may contain multiple packets, qeth's NAPI poll code can
exhaust its budget in the middle of an RX buffer. Thus we keep track of
our current position within the active RX buffer, so we can resume
processing here in the next NAPI poll period.
Clean up that code by tracking the index of the active buffer element,
instead of a pointer to it.
Also simplify the code that advances to the next RX buffer when the
current buffer has been fully processed.
v2: - remove QDIO_ELEMENT_NO() macro (davem)
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently ENA only provides the PCI remove() handler, used during rmmod
for example. This is not called on shutdown/kexec path; we are potentially
creating a failure scenario on kexec:
(a) Kexec is triggered, no shutdown() / remove() handler is called for ENA;
instead pci_device_shutdown() clears the master bit of the PCI device,
stopping all DMA transactions;
(b) Kexec reboot happens and the device gets enabled again, likely having
its FW with that DMA transaction buffered; then it may trigger the (now
invalid) memory operation in the new kernel, corrupting kernel memory area.
This patch aims to prevent this, by implementing a shutdown() handler
quite similar to the remove() one - the difference being the handling
of the netdev, which is unregistered on remove(), but following the
convention observed in other drivers, it's only detached on shutdown().
This prevents an odd issue in AWS Nitro instances, in which after the 2nd
kexec the next one will fail with an initrd corruption, caused by a wild
DMA write to invalid kernel memory. The lspci output for the adapter
present in my instance is:
00:05.0 Ethernet controller [0200]: Amazon.com, Inc. Elastic Network
Adapter (ENA) [1d0f:ec20]
Suggested-by: Gavin Shan <gshan@redhat.com>
Signed-off-by: Guilherme G. Piccoli <gpiccoli@canonical.com>
Acked-by: Sameeh Jubran <sameehj@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The lib files should not be defined as TEST_PROGS, or we will run them
in run_kselftest.sh.
Also remove ethtool_lib.sh exec permission.
Fixes: 81573b18f2 ("selftests/net/forwarding: add Makefile to install tests")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Find some tests are missed in Makefile by running:
for file in $(ls *.sh); do grep -q $file Makefile || echo $file; done
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP recvmsg() calls skb_copy_datagram_iter(), which
calls an indirect function (cb pointing to simple_copy_to_iter())
for every MSS (fragment) present in the skb.
CONFIG_RETPOLINE=y forces a very expensive operation
that we can avoid thanks to indirect call wrappers.
This patch gives a 13% increase of performance on
a single flow, if the bottleneck is the thread reading
the TCP socket.
Fixes: 950fcaecd5 ("datagram: consolidate datagram copy to iter helpers")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.c: In function cxgb4_get_free_ftid:
drivers/net/ethernet/chelsio/cxgb4/cxgb4_filter.c:547:23:
warning: variable tab set but not used [-Wunused-but-set-variable]
commit 8d174351f2 ("cxgb4: rework TC filter rule insertion across regions")
involved this, remove it.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A single fix in this pull request to correctly handle the size of
read-only zone files (from me).
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQSRPv8tYSvhwAzJdzjdoc3SxdoYdgUCXnrJJAAKCRDdoc3SxdoY
dgZXAQDK88T4sdtFq1Fl1PuP+oyzHml+xgNo0djZQOdicnD6qQD8CgMGDFQQG4dv
Ral+67qEyvUABGt0Vkmy29wuN8El6wQ=
=+1D9
-----END PGP SIGNATURE-----
Merge tag 'zonefs-5.6-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs
Pull zonefs fix from Damien Le Moal:
"A single fix from me to correctly handle the size of read-only zone
files"
* tag 'zonefs-5.6-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs:
zonfs: Fix handling of read-only zones
The write pointer of zones in the read-only consition is defined as
invalid by the SCSI ZBC and ATA ZAC specifications. It is thus not
possible to determine the correct size of a read-only zone file on
mount. Fix this by handling read-only zones in the same manner as
offline zones by disabling all accesses to the zone (read and write)
and initializing the inode size of the read-only zone to 0).
For zones found to be in the read-only condition at runtime, only
disable write access to the zone and keep the size of the zone file to
its last updated value to allow the user to recover previously written
data.
Also fix zonefs documentation file to reflect this change.
Signed-off-by: Damien Le Moal <damien.lemoal@wdc.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contains Netfilter fixes for net:
1) A new selftest for nf_queue, from Florian Westphal. This test
covers two recent fixes: 07f8e4d0fd ("tcp: also NULL skb->dev
when copy was needed") and b738a185be ("tcp: ensure skb->dev is
NULL before leaving TCP stack").
2) The fwd action breaks with ifb. For safety in next extensions,
make sure the fwd action only runs from ingress until it is extended
to be used from a different hook.
3) The pipapo set type now reports EEXIST in case of subrange overlaps.
Update the rbtree set to validate range overlaps, so far this
validation is only done only from userspace. From Stefano Brivio.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAl56fu0ACgkQSD+KveBX
+j4g1Qf/edWPTCMGK4eb0jBPUvnxkoGPYj4cq0tfeUaY7Q4r95RzNjuW3gTotzGV
JZymoC2OoWQxUR2Ye0FkM1C/RQFIAHinEX/KFOMJ6PL+k4+micXeIGNfVo3aflO0
kaTcBdgZKqFS5hpRtWZc/DVRWckqJYtaAJEFliQbYGwmfiZNoNr0/ZeU+/DX2dHn
bQkRHQZ3Zq43P4FhVBSyrfmsxUI71k7GtCdJ5G4i80e8qCCARKZDx7q1FRC0k6fh
a84+7NxpFRkl+kT+se/bxcQaFht49YSJVauGMKK8Ae+pz0XEaNrYsFz9zQY/s7W6
4a62hzlHuVmUwteZfH+secZzCnOkUw==
=rO1d
-----END PGP SIGNATURE-----
Merge tag 'mlx5-fixes-2020-03-24' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
Mellanox, mlx5 fixes 2020-03-24
This series introduces some fixes to mlx5 driver.
From Aya, Fixes to the RX error recovery flows
From Leon, Fix IB capability mask
Please pull and let me know if there is any problem.
For -stable v5.5
('net/mlx5_core: Set IB capability mask1 to fix ib_srpt connection failure')
For -stable v5.4
('net/mlx5e: Fix ICOSQ recovery flow with Striding RQ')
('net/mlx5e: Do not recover from a non-fatal syndrome')
('net/mlx5e: Fix missing reset of SW metadata in Striding RQ reset')
('net/mlx5e: Enhance ICOSQ WQE info fields')
The above patch ('net/mlx5e: Enhance ICOSQ WQE info fields')
will fail to apply cleanly on v5.4 due to a trivial contextual conflict,
but it is an important fix, do I need to do something about it or just
assume Greg will know how to handle this ?
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The original change fixed an issue on RTL8168b by mimicking the vendor
driver behavior to disable MSI on chip versions before RTL8168d.
This however now caused an issue on a system with RTL8168c, see [0].
Therefore leave MSI disabled on RTL8168b, but re-enable it on RTL8168c.
[0] https://bugzilla.redhat.com/show_bug.cgi?id=1792839
Fixes: 003bd5b4a7 ("r8169: don't use MSI before RTL8168d")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We are having multiple review cycles with all vendors trying
to implement devlink-info. Let's expand the documentation with
more information about what's implemented and motivation behind
this interface in an attempt to make the implementations easier.
Describe what each info section is supposed to contain, and make
some references to other HW interfaces (PCI caps).
Document how firmware management is expected to look, to make
it clear how devlink-info and devlink-flash work in concert.
Name some future work.
v2: - improve wording
v3: - improve wording
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The DT binding for this PHY describes an *optional* clock property.
Due to a bug in the error handling logic, we are actually ignoring this
clock *all* of the time so far.
Fix this by using devm_clk_get_optional() to handle this clock properly.
Fixes: b78ac6ecd1 ("net: phy: mdio-bcm-unimac: Allow configuring MDIO clock divider")
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It looks like the VSC8584 PHY driver is rolling its own RGMII delay
configuration code, despite the fact that the logic is mostly the same.
In fact only the register layout and position for the RGMII controls has
changed. So we need to adapt and parameterize the PHY-dependent bit
fields when calling the new generic function.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Antoine Tenart <antoine.tenart@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andre Przywara says:
====================
net: axienet: Update error handling and add 64-bit DMA support
a minor update, fixing the 32-bit build breakage, and brightening up
Dave's christmas tree. Rebased against latest net-next/master.
This series is based on net-next as of today (9970de8b01), which
includes Russell's fixes [1], solving the SGMII issues I have had.
[1] https://lore.kernel.org/netdev/E1j6trA-0003GY-N1@rmk-PC.armlinux.org.uk/
Changelog v2 .. v3:
- Use two "left-shifts by 16" to fix builds with 32-bit phys_addr_t
- reorder variable declarations
Changelog v1 .. v2:
- Add Reviewed-by: tags from Radhey
- Extend kerndoc documentation
- Convert DMA error handler tasklet to work queue
- log DMA mapping errors
- mark DMA mapping error checks as unlikely (in "hot" paths)
- return NETDEV_TX_OK on TX DMA mapping error (increasing TX drop counter)
- Request eth IRQ as an optional IRQ
- Remove no longer needed MDIO IRQ register names
- Drop DT propery check for address width, assume full 64 bit
This series updates the Xilinx Axienet driver to work on our board
here. One big issue was broken SGMII support, which Russell fixed already
(in net-next).
While debugging and understanding the driver, I found several problems
in the error handling and cleanup paths, which patches 2-7 address.
Patch 8 removes a annoying error message, patch 9 paves the way for newer
revisions of the IP. The next patch adds mii-tool support, just for good
measure.
The next four patches add support for 64-bit DMA. This is an integration
option on newer IP revisions (>= v7.1), and expects MSB bits in formerly
reserved registers. Without writing to those MSB registers, the state
machine won't trigger, so it's mandatory to access them, even if they
are zero. Patches 11 and 12 prepare the code by adding accessors, to
wrap this properly and keep it working on older IP revisions.
Patch 13 enables access to the MSB registers, by trying to write a
non-zero value to them and checking if that sticks. Older IP revisions
always read those registers as zero.
Patch 14 then adjusts the DMA mask, based on the autodetected MSB
feature. It uses the full 64 bits in this case, the rest of the system
(actual physical addresses in use) should provide a natural limit if the
chip has connected fewer address lines. If not, the parent DT node can
use a dma-range property.
The Xilinx PG138 and PG021 documents (in versions 7.1 in both cases)
were used for this series.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
With all DMA address accesses wrapped, we can actually support 64-bit
DMA if this option was chosen at IP integration time.
If the IP has been configured for an address width greater than 32 bits,
we assume the full 64 bit DMA width is working. In practise this will be
limited by the actual system address bus width, which will ideally be the
same as the DMA IP address width.
If this is not the case, the actual width can still be configured using a
dma-ranges property in the parent of the MAC node.
This increases the DMA mask on those systems to let the kernel choose
buffers from memory at higher addresses.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When newer revisions of the Axienet IP are configured for a 64-bit bus,
we *need* to write to the MSB part of the an address registers,
otherwise the IP won't recognise this as a DMA start condition.
This is even true when the actual DMA address comes from the lower 4 GB.
To autodetect this configuration, at probe time we write all 1's to such
an MSB register, and see if any bits stick. If this is configured for a
32-bit bus, those MSB registers are RES0, so reading back 0 indicates
that no MSB writes are necessary.
On the other hands reading anything other than 0 indicated the need to
write the MSB registers, so we set the respective flag.
The actual DMA mask stays at 32-bit for now. To help bisecting, a
separate patch will enable allocations from higher addresses.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer revisions of the AXI DMA IP (>= v7.1) support 64-bit addresses,
both for the descriptors itself, as well as for the buffers they are
pointing to.
This is realised by adding "MSB" words for the next and phys pointer
right behind the existing address word, now named "LSB". These MSB words
live in formerly reserved areas of the descriptor.
If the hardware supports it, write both words when setting an address.
The buffer address is handled by two wrapper functions, the two
occasions where we set the next pointers are open coded.
For now this is guarded by a flag which we don't set yet.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer versions of the Xilink DMA IP support busses with more than 32
address bits, by introducing an MSB word for the registers holding DMA
pointers (tail/current, RX/TX descriptor addresses).
On IP configured for more than 32 bits, it is also *required* to write
both words, to let the IP recognise this as a start condition for an
MM2S request, for instance.
Wrap the DMA pointer writes with a separate function, to add this
functionality later. For now we stick to the lower 32 bits.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
mii-tool is useful for debugging, and all it requires to work is to wire
up the ioctl ops function pointer.
Add this to the axienet driver to enable mii-tool.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer revisions of the IP don't have these registers. Since we don't
really use them, just drop them from the ethtools dump.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
According to the DT binding, the Ethernet core interrupt is optional.
Use platform_get_irq_optional() to avoid the error message when the
IRQ is not specified.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Especially with the default 32-bit DMA mask, DMA buffers are a limited
resource, so their allocation can fail.
So as the DMA API documentation requires, add error checking code after
dma_map_single() calls to catch the case where we run out of "low" memory.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Factor out the code that cleans up a number of connected TX descriptors,
as we will need it to properly roll back a failed _xmit() call.
There are subtle differences between cleaning up a successfully sent
chain (unknown number of involved descriptors, total data size needed)
and a chain that was about to set up (number of descriptors known), so
cater for those variations with some extra parameters.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since 0 is a valid DMA address, we cannot use the physical address to
check whether a TX descriptor is valid and is holding a DMA mapping.
Use the "cntrl" member of the descriptor to make this decision, as it
contains at least the length of the buffer, so 0 points to an
uninitialised buffer.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When axienet_dma_bd_init() bails out during the initialisation process,
it might do so with parts of the structure already allocated and
initialised, while other parts have not been touched yet. Before
returning in this case, we call axienet_dma_bd_release(), which does not
take care of this corner case.
This is most obvious by the first loop happily dereferencing
lp->rx_bd_v, which we actually check to be non NULL *afterwards*.
Make sure we only unmap or free already allocated structures, by:
- directly returning with -ENOMEM if nothing has been allocated at all
- checking for lp->rx_bd_v to be non-NULL *before* using it
- only unmapping allocated DMA RX regions
This avoids NULL pointer dereferences when initialisation fails.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we fail allocating the DMA buffers in axienet_dma_bd_init(), we
report this error, but carry on with initialisation nevertheless.
This leads to a kernel panic when the driver later wants to send a
packet, as it uses uninitialised data structures.
Make the axienet_device_reset() routine return an error value, as it
contains the DMA buffer initialisation. Make sure we propagate the error
up the chain and eventually fail the driver initialisation, to avoid
relying on non-initialised buffers.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The DMA error handler routine is currently a tasklet, scheduled to run
after the DMA error IRQ was handled.
However it needs to take the MDIO mutex, which is not allowed to do in a
tasklet. A kernel (with debug options) complains consequently:
[ 614.050361] net eth0: DMA Tx error 0x174019
[ 614.064002] net eth0: Current BD is at: 0x8f84aa0ce
[ 614.080195] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:935
[ 614.109484] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 40, name: kworker/u4:4
[ 614.135428] 3 locks held by kworker/u4:4/40:
[ 614.149075] #0: ffff000879863328 ((wq_completion)rpciod){....}, at: process_one_work+0x1f0/0x6a8
[ 614.177528] #1: ffff80001251bdf8 ((work_completion)(&task->u.tk_work)){....}, at: process_one_work+0x1f0/0x6a8
[ 614.209033] #2: ffff0008784e0110 (sk_lock-AF_INET-RPC){....}, at: tcp_sendmsg+0x24/0x58
[ 614.235429] CPU: 0 PID: 40 Comm: kworker/u4:4 Not tainted 5.6.0-rc3-00926-g4a165a9d5921 #26
[ 614.260854] Hardware name: ARM Test FPGA (DT)
[ 614.274734] Workqueue: rpciod rpc_async_schedule
[ 614.289022] Call trace:
[ 614.296871] dump_backtrace+0x0/0x1a0
[ 614.308311] show_stack+0x14/0x20
[ 614.318751] dump_stack+0xbc/0x100
[ 614.329403] ___might_sleep+0xf0/0x140
[ 614.341018] __might_sleep+0x4c/0x80
[ 614.352201] __mutex_lock+0x5c/0x8a8
[ 614.363348] mutex_lock_nested+0x1c/0x28
[ 614.375654] axienet_dma_err_handler+0x38/0x388
[ 614.389999] tasklet_action_common.isra.15+0x160/0x1a8
[ 614.405894] tasklet_action+0x24/0x30
[ 614.417297] efi_header_end+0xe0/0x494
[ 614.429020] irq_exit+0xd0/0xd8
[ 614.439047] __handle_domain_irq+0x60/0xb0
[ 614.451877] gic_handle_irq+0xdc/0x2d0
[ 614.463486] el1_irq+0xcc/0x180
[ 614.473451] __tcp_transmit_skb+0x41c/0xb58
[ 614.486513] tcp_write_xmit+0x224/0x10a0
[ 614.498792] __tcp_push_pending_frames+0x38/0xc8
[ 614.513126] tcp_rcv_established+0x41c/0x820
[ 614.526301] tcp_v4_do_rcv+0x8c/0x218
[ 614.537784] __release_sock+0x5c/0x108
[ 614.549466] release_sock+0x34/0xa0
[ 614.560318] tcp_sendmsg+0x40/0x58
[ 614.571053] inet_sendmsg+0x40/0x68
[ 614.582061] sock_sendmsg+0x18/0x30
[ 614.593074] xs_sendpages+0x218/0x328
[ 614.604506] xs_tcp_send_request+0xa0/0x1b8
[ 614.617461] xprt_transmit+0xc8/0x4f0
[ 614.628943] call_transmit+0x8c/0xa0
[ 614.640028] __rpc_execute+0xbc/0x6f8
[ 614.651380] rpc_async_schedule+0x28/0x48
[ 614.663846] process_one_work+0x298/0x6a8
[ 614.676299] worker_thread+0x40/0x490
[ 614.687687] kthread+0x134/0x138
[ 614.697804] ret_from_fork+0x10/0x18
[ 614.717319] xilinx_axienet 7fe00000.ethernet eth0: Link is Down
[ 615.748343] xilinx_axienet 7fe00000.ethernet eth0: Link is Up - 1Gbps/Full - flow control off
Since tasklets are not really popular anymore anyway, lets convert this
over to a work queue, which can sleep and thus can take the MDIO mutex.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to axienet, the temac driver is now architecture agnostic, and
can be at least compiled for several architectures.
Especially the fact that this is a soft IP for implementing in FPGAs
makes the current restriction rather pointless, as it could literally
appear on any architecture, as long as an FPGA is connected to the bus.
The driver hasn't been actually tried on any hardware, it is just a
drive-by patch when doing the same for axienet (a similar patch for
axienet is already merged).
This (temac and axienet) have been compile-tested for:
alpha hppa64 microblaze mips64 powerpc powerpc64 riscv64 s390 sparc64
(using kernel.org cross compilers).
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Radhey Shyam Pandey <radhey.shyam.pandey@xilinx.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, ethtool feature mask for checksum command is ORed with
NETIF_F_FCOE_CRC_BIT, which is bit's position number, instead of the
actual feature bit - NETIF_F_FCOE_CRC.
The invalid bitmask here might affect unrelated features when toggling
TX checksumming. For example, TX checksumming is always mistakenly
reported as enabled on the netdevs tested (mlx5, virtio_net).
Fixes: f70bb06563 ("ethtool: update mapping of features to legacy ioctl requests")
Signed-off-by: Vladyslav Tarasiuk <vladyslavt@mellanox.com>
Reviewed-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
cxgb4_ptp_fineadjtime() doesn't pass the signedness of offset delta
in FW_PTP_CMD. Fix it by passing correct sign.
Signed-off-by: Raju Rangoju <rajur@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
use readl_poll_timeout() to replace the poll codes for simplify
iproc_mdio_wait_for_idle() function
Signed-off-by: Dejin Zheng <zhengdejin5@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Not only did this wheel did not need reinventing, but there is also
an issue with it: It doesn't remove the VLAN header in a way that
preserves the L2 payload checksum when that is being provided by the DSA
master hw. It should recalculate checksum both for the push, before
removing the header, and for the pull afterwards. But the current
implementation is quite dizzying, with pulls followed immediately
afterwards by pushes, the memmove is done before the push, etc. This
makes a DSA master with RX checksumming offload to print stack traces
with the infamous 'hw csum failure' message.
So remove the dsa_8021q_remove_header function and replace it with
something that actually works with inet checksumming.
Fixes: d461933638 ("net: dsa: tag_8021q: Create helper function for removing VLAN header")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Second set of patches for v5.7. Lots of cleanup patches this time, but
of course various new features as well fixes.
When merging with wireless-drivers this pull request has a conflict in:
drivers/net/wireless/intel/iwlwifi/pcie/drv.c
To solve that just drop the changes from commit cf52c8a776 in
wireless-drivers and take the hunk from wireless-drivers-next as is.
The list of specific subsystem device IDs are not necessary after
commit d6f2134a38 (in wireless-drivers-next) anymore, the detection
is based on other characteristics of the devices.
Major changes:
qtnfmac
* support WPA3 SAE and OWE in AP mode
ath10k
* support for getting btcoex settings from Device Tree
* support QCA9377 SDIO device
ath11k
* add HE rate accounting
* add thermal sensor and cooling devices
mt76
* MT7663 support for the MT7615 driver
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJeechaAAoJEG4XJFUm622b/IcH/jzdxSwI/dRp2ZrEG+UFDNyZ
ChTdgW1k+WWLrTKaw5S+mmtM9sLQ7Ia5dL+ZyE7tFIEGzuCSdAbbOIkygw+rhLy9
BO0rsTbXRuErRpUJgwKc56sFrvBL2pm4Rx+4IV29jhXdyJ70MUvRAINcHLESeBcO
uI5dqapScRf/EnVYP2Zso6i2FLep5hoC0reBq2UQ5Y95Lv0cpMCbHQ1xqdcm9pI3
ZvLDpVyGpp75Eoxlt1Ubiv6ijsEnTaM3cI+Q5rRWt8g3JKP7yYT7HGbJ3fb0cWhq
bNNISs+XN67lnyU1Mx3c2pjyb+szJZ0o4OGbYNAqsquItl4B0RQc3fUBMlCUlYI=
=+2Hj
-----END PGP SIGNATURE-----
Merge tag 'wireless-drivers-next-2020-03-24' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next
Kalle Valo says:
====================
wireless-drivers-next patches for v5.7
Second set of patches for v5.7. Lots of cleanup patches this time, but
of course various new features as well fixes.
When merging with wireless-drivers this pull request has a conflict in:
drivers/net/wireless/intel/iwlwifi/pcie/drv.c
To solve that just drop the changes from commit cf52c8a776 in
wireless-drivers and take the hunk from wireless-drivers-next as is.
The list of specific subsystem device IDs are not necessary after
commit d6f2134a38 (in wireless-drivers-next) anymore, the detection
is based on other characteristics of the devices.
Major changes:
qtnfmac
* support WPA3 SAE and OWE in AP mode
ath10k
* support for getting btcoex settings from Device Tree
* support QCA9377 SDIO device
ath11k
* add HE rate accounting
* add thermal sensor and cooling devices
mt76
* MT7663 support for the MT7615 driver
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the software CBS does not consider the packet sending time
when depleting the credits. It caused the throughput to be
Idleslope[kbps] * (Port transmit rate[kbps] / |Sendslope[kbps]|) where
Idleslope * (Port transmit rate / (Idleslope + |Sendslope|)) = Idleslope
is expected. In order to fix the issue above, this patch takes the time
when the packet sending completes into account by moving the anchor time
variable "last" ahead to the send completion time upon transmission and
adding wait when the next dequeue request comes before the send
completion time of the previous packet.
changelog:
V2->V3:
- remove unnecessary whitespace cleanup
- add the checks if port_rate is 0 before division
V1->V2:
- combine variable "send_completed" into "last"
- add the comment for estimate of the packet sending
Fixes: 585d763af0 ("net/sched: Introduce Credit Based Shaper (CBS) qdisc")
Signed-off-by: Zh-yuan Ye <ye.zh-yuan@socionext.com>
Reviewed-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For non-fatal syndromes like LOCAL_LENGTH_ERR, recovery shouldn't be
triggered. In these scenarios, the RQ is not actually in ERR state.
This misleads the recovery flow which assumes that the RQ is really in
error state and no more completions arrive, causing crashes on bad page
state.
Fixes: 8276ea1353 ("net/mlx5e: Report and recover from CQE with error on RQ")
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
In striding RQ mode, the buffers of an RX WQE are first
prepared and posted to the HW using a UMR WQEs via the ICOSQ.
We maintain the state of these in-progress WQEs in the RQ
SW struct.
In the flow of ICOSQ recovery, the corresponding RQ is not
in error state, hence:
- The buffers of the in-progress WQEs must be released
and the RQ metadata should reflect it.
- Existing RX WQEs in the RQ should not be affected.
For this, wrap the dealloc of the in-progress WQEs in
a function, and use it in the ICOSQ recovery flow
instead of mlx5e_free_rx_descs().
Fixes: be5323c837 ("net/mlx5e: Report and recover from CQE error on ICOSQ")
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
When resetting the RQ (moving RQ state from RST to RDY), the driver
resets the WQ's SW metadata.
In striding RQ mode, we maintain a field that reflects the actual
expected WQ head (including in progress WQEs posted to the ICOSQ).
It was mistakenly not reset together with the WQ. Fix this here.
Fixes: 8276ea1353 ("net/mlx5e: Report and recover from CQE with error on RQ")
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Add number of WQEBBs (WQE's Basic Block) to WQE info struct. Set the
number of WQEBBs on WQE post, and increment the consumer counter (cc)
on completion.
In case of error completions, the cc was mistakenly not incremented,
keeping a gap between cc and pc (producer counter). This failed the
recovery flow on the ICOSQ from a CQE error which timed-out waiting for
the cc and pc to meet.
Fixes: be5323c837 ("net/mlx5e: Report and recover from CQE error on ICOSQ")
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The cap_mask1 isn't protected by field_select and not listed among RW
fields, but it is required to be written to properly initialize ports
in IB virtualization mode.
Link: https://lore.kernel.org/linux-rdma/88bab94d2fd72f3145835b4518bc63dda587add6.camel@redhat.com
Fixes: ab118da4c1 ("net/mlx5: Don't write read-only fields in MODIFY_HCA_VPORT_CONTEXT command")
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Add a test case to check nf queue infrastructure.
Could be extended in the future to also cover serialization of
conntrack, uid and secctx attributes in nfqueue.
For now, this checks that 'queue bypass' works, that a queue rule with
no bypass option blocks traffic and that userspace receives the expected
number of packets.
For this we add two queues and hook all of
prerouting/input/forward/output/postrouting.
Packets get queued twice with a dummy base chain in between:
This passes with current nf tree, but reverting
commit 946c0d8e6e ("netfilter: nf_queue: fix reinject verdict handling")
makes this trip (it processes 30 instead of expected 20 packets).
v2: update config file with queue and other options missing/needed for
other tests.
v3: also test with tcp, this reveals problem with commit
28f8bfd1ac ("netfilter: Support iif matches in POSTROUTING"), due to
skb->dev pointing at another skb in the retransmit rbtree (skb->dev
aliases to rbnode child).
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Set skb->tc_redirected to 1, otherwise the ifb driver drops the packet.
Set skb->tc_from_ingress to 1 to reinject the packet back to the ingress
path after leaving the ifb egress path.
This patch inconditionally sets on these two skb fields that are
meaningful to the ifb driver. The existing forward action is guaranteed
to run from ingress path.
Fixes: 39e6dea28a ("netfilter: nf_tables: add forward expression to the netdev family")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Make sure the forward action is only used from ingress.
Fixes: 39e6dea28a ("netfilter: nf_tables: add forward expression to the netdev family")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
...and return -ENOTEMPTY to the front-end in this case, instead of
proceeding. Currently, nft takes care of checking for these cases
and not sending them to the kernel, but if we drop the set_overlap()
call in nft we can end up in situations like:
# nft add table t
# nft add set t s '{ type inet_service ; flags interval ; }'
# nft add element t s '{ 1 - 5 }'
# nft add element t s '{ 6 - 10 }'
# nft add element t s '{ 4 - 7 }'
# nft list set t s
table ip t {
set s {
type inet_service
flags interval
elements = { 1-3, 4-5, 6-7 }
}
}
This change has the primary purpose of making the behaviour
consistent with nft_set_pipapo, but is also functional to avoid
inconsistent behaviour if userspace sends overlapping elements for
any reason.
v2: When we meet the same key data in the tree, as start element while
inserting an end element, or as end element while inserting a start
element, actually check that the existing element is active, before
resetting the overlap flag (Pablo Neira Ayuso)
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
