The RX rings have a producer index owned by hardware, where newly
received frame buffers are placed, and a consumer index owned by
software, where newly allocated buffers are placed, in expectation of
hardware being able to place frame data in them.
Hardware increments the producer index when a frame is received, however
it is not allowed to increment the producer index to match the consumer
index (RBCIR) since the ring can hold at most RBLENR[LENGTH]-1 received
BDs. Whenever the producer index matches the value of the consumer
index, the ring has no unprocessed received frames and all BDs in the
ring have been initialized/prepared by software, i.e. hardware owns all
BDs in the ring.
The code uses the next_to_clean variable to keep track of the producer
index, and the next_to_use variable to keep track of the consumer index.
The RX rings are seeded from enetc_refill_rx_ring, which is called from
two places:
1. initially the ring is seeded until full with enetc_bd_unused(rx_ring),
i.e. with 511 buffers. This will make next_to_clean=0 and next_to_use=511:
.ndo_open
-> enetc_open
-> enetc_setup_bdrs
-> enetc_setup_rxbdr
-> enetc_refill_rx_ring
2. then during the data path processing, it is refilled with 16 buffers
at a time:
enetc_msix
-> napi_schedule
-> enetc_poll
-> enetc_clean_rx_ring
-> enetc_refill_rx_ring
There is just one problem: the initial seeding done during .ndo_open
updates just the producer index (ENETC_RBPIR) with 0, and the software
next_to_clean and next_to_use variables. Notably, it will not update the
consumer index to make the hardware aware of the newly added buffers.
Wait, what? So how does it work?
Well, the reset values of the producer index and of the consumer index
of a ring are both zero. As per the description in the second paragraph,
it means that the ring is full of buffers waiting for hardware to put
frames in them, which by coincidence is almost true, because we have in
fact seeded 511 buffers into the ring.
But will the hardware attempt to access the 512th entry of the ring,
which has an invalid BD in it? Well, no, because in order to do that, it
would have to first populate the first 511 entries, and the NAPI
enetc_poll will kick in by then. Eventually, after 16 processed slots
have become available in the RX ring, enetc_clean_rx_ring will call
enetc_refill_rx_ring and then will [ finally ] update the consumer index
with the new software next_to_use variable. From now on, the
next_to_clean and next_to_use variables are in sync with the producer
and consumer ring indices.
So the day is saved, right? Well, not quite. Freeing the memory
allocated for the rings is done in:
enetc_close
-> enetc_clear_bdrs
-> enetc_clear_rxbdr
-> this just disables the ring
-> enetc_free_rxtx_rings
-> enetc_free_rx_ring
-> sets next_to_clean and next_to_use to 0
but again, nothing is committed to the hardware producer and consumer
indices (yay!). The assumption is that the ring is disabled, so the
indices don't matter anyway, and it's the responsibility of the "open"
code path to set those up.
.. Except that the "open" code path does not set those up properly.
While initially, things almost work, during subsequent enetc_close ->
enetc_open sequences, we have problems. To be precise, the enetc_open
that is subsequent to enetc_close will again refill the ring with 511
entries, but it will leave the consumer index untouched. Untouched
means, of course, equal to the value it had before disabling the ring
and draining the old buffers in enetc_close.
But as mentioned, enetc_setup_rxbdr will at least update the producer
index though, through this line of code:
enetc_rxbdr_wr(hw, idx, ENETC_RBPIR, 0);
so at this stage we'll have:
next_to_clean=0 (in hardware 0)
next_to_use=511 (in hardware we'll have the refill index prior to enetc_close)
Again, the next_to_clean and producer index are in sync and set to
correct values, so the driver manages to limp on. Eventually, 16 ring
entries will be consumed by enetc_poll, and the savior
enetc_clean_rx_ring will come and call enetc_refill_rx_ring, and then
update the hardware consumer ring based upon the new next_to_use.
So.. it works?
Well, by coincidence, it almost does, but there's a circumstance where
enetc_clean_rx_ring won't be there to save us. If the previous value of
the consumer index was 15, there's a problem, because the NAPI poll
sequence will only issue a refill when 16 or more buffers have been
consumed.
It's easiest to illustrate this with an example:
ip link set eno0 up
ip addr add 192.168.100.1/24 dev eno0
ping 192.168.100.1 -c 20 # ping this port from another board
ip link set eno0 down
ip link set eno0 up
ping 192.168.100.1 -c 20 # ping it again from the same other board
One by one:
1. ip link set eno0 up
-> calls enetc_setup_rxbdr:
-> calls enetc_refill_rx_ring(511 buffers)
-> next_to_clean=0 (in hw 0)
-> next_to_use=511 (in hw 0)
2. ping 192.168.100.1 -c 20 # ping this port from another board
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=1 next_to_clean 0 (in hw 1) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=2 next_to_clean 1 (in hw 2) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=3 next_to_clean 2 (in hw 3) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=4 next_to_clean 3 (in hw 4) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=5 next_to_clean 4 (in hw 5) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=6 next_to_clean 5 (in hw 6) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=7 next_to_clean 6 (in hw 7) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=8 next_to_clean 7 (in hw 8) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=9 next_to_clean 8 (in hw 9) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=10 next_to_clean 9 (in hw 10) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=11 next_to_clean 10 (in hw 11) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=12 next_to_clean 11 (in hw 12) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=13 next_to_clean 12 (in hw 13) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=14 next_to_clean 13 (in hw 14) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=15 next_to_clean 14 (in hw 15) next_to_use 511 (in hw 0)
enetc_clean_rx_ring: enetc_refill_rx_ring(16) increments next_to_use by 16 (mod 512) and writes it to hw
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=0 next_to_clean 15 (in hw 16) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=1 next_to_clean 16 (in hw 17) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=2 next_to_clean 17 (in hw 18) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=3 next_to_clean 18 (in hw 19) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=4 next_to_clean 19 (in hw 20) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=5 next_to_clean 20 (in hw 21) next_to_use 15 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=6 next_to_clean 21 (in hw 22) next_to_use 15 (in hw 15)
20 packets transmitted, 20 packets received, 0% packet loss
3. ip link set eno0 down
enetc_free_rx_ring: next_to_clean 0 (in hw 22), next_to_use 0 (in hw 15)
4. ip link set eno0 up
-> calls enetc_setup_rxbdr:
-> calls enetc_refill_rx_ring(511 buffers)
-> next_to_clean=0 (in hw 0)
-> next_to_use=511 (in hw 15)
5. ping 192.168.100.1 -c 20 # ping it again from the same other board
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=1 next_to_clean 0 (in hw 1) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=2 next_to_clean 1 (in hw 2) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=3 next_to_clean 2 (in hw 3) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=4 next_to_clean 3 (in hw 4) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=5 next_to_clean 4 (in hw 5) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=6 next_to_clean 5 (in hw 6) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=7 next_to_clean 6 (in hw 7) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=8 next_to_clean 7 (in hw 8) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=9 next_to_clean 8 (in hw 9) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=10 next_to_clean 9 (in hw 10) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=11 next_to_clean 10 (in hw 11) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=12 next_to_clean 11 (in hw 12) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=13 next_to_clean 12 (in hw 13) next_to_use 511 (in hw 15)
enetc_clean_rx_ring: rx_frm_cnt=1 cleaned_cnt=14 next_to_clean 13 (in hw 14) next_to_use 511 (in hw 15)
20 packets transmitted, 12 packets received, 40% packet loss
And there it dies. No enetc_refill_rx_ring (because cleaned_cnt must be equal
to 15 for that to happen), no nothing. The hardware enters the condition where
the producer (14) + 1 is equal to the consumer (15) index, which makes it
believe it has no more free buffers to put packets in, so it starts discarding
them:
ip netns exec ns0 ethtool -S eno0 | grep -v ': 0'
NIC statistics:
Rx ring 0 discarded frames: 8
Summarized, if the interface receives between 16 and 32 (mod 512) frames
and then there is a link flap, then the port will eventually die with no
way to recover. If it receives less than 16 (mod 512) frames, then the
initial NAPI poll [ before the link flap ] will not update the consumer
index in hardware (it will remain zero) which will be ok when the buffers
are later reinitialized. If more than 32 (mod 512) frames are received,
the initial NAPI poll has the chance to refill the ring twice, updating
the consumer index to at least 32. So after the link flap, the consumer
index is still wrong, but the post-flap NAPI poll gets a chance to
refill the ring once (because it passes through cleaned_cnt=15) and
makes the consumer index be again back in sync with next_to_use.
The solution to this problem is actually simple, we just need to write
next_to_use into the hardware consumer index at enetc_open time, which
always brings it back in sync after an initial buffer seeding process.
The simpler thing would be to put the write to the consumer index into
enetc_refill_rx_ring directly, but there are issues with the MDIO
locking: in the NAPI poll code we have the enetc_lock_mdio() taken from
top-level and we use the unlocked enetc_wr_reg_hot, whereas in
enetc_open, the enetc_lock_mdio() is not taken at the top level, but
instead by each individual enetc_wr_reg, so we are forced to put an
additional enetc_wr_reg in enetc_setup_rxbdr. Better organization of
the code is left as a refactoring exercise.
Fixes: d4fd0404c1 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Station Interface Receive Interrupt Detect Register (SIRXIDR)
contains a 16-bit wide mask of 'interrupt detected' events for each ring
associated with a port. Bit i is write-1-to-clean for RX ring i.
I have no explanation whatsoever how this line of code came to be
inserted in the blamed commit. I checked the downstream versions of that
patch and none of them have it.
The somewhat comical aspect of it is that we're writing a binary number
to the SIRXIDR register, which is derived from enetc_bd_unused(rx_ring).
Since the RX rings have 512 buffer descriptors, we end up writing 511 to
this register, which is 0x1ff, so we are effectively clearing the
'interrupt detected' event for rings 0-8.
This register is not what is used for interrupt handling though - it
only provides a summary for the entire SI. The hardware provides one
separate Interrupt Detect Register per RX ring, which auto-clears upon
read. So there doesn't seem to be any adverse effect caused by this
bogus write.
There is, however, one reason why this should be handled as a bugfix:
next_to_clean _should_ be committed to hardware, just not to that
register, and this was obscuring the fact that it wasn't. This is fixed
in the next patch, and removing the bogus line now allows the fix patch
to be backported beyond that point.
Fixes: fd5736bf9f ("enetc: Workaround for MDIO register access issue")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The ENETC port 0 MAC supports in-band status signaling coming from a PHY
when operating in RGMII mode, and this feature is enabled by default.
It has been reported that RGMII is broken in fixed-link, and that is not
surprising considering the fact that no PHY is attached to the MAC in
that case, but a switch.
This brings us to the topic of the patch: the enetc driver should have
not enabled the optional in-band status signaling for RGMII unconditionally,
but should have forced the speed and duplex to what was resolved by
phylink.
Note that phylink does not accept the RGMII modes as valid for in-band
signaling, and these operate a bit differently than 1000base-x and SGMII
(notably there is no clause 37 state machine so no ACK required from the
MAC, instead the PHY sends extra code words on RXD[3:0] whenever it is
not transmitting something else, so it should be safe to leave a PHY
with this option unconditionally enabled even if we ignore it). The spec
talks about this here:
https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/138/RGMIIv1_5F00_3.pdf
Fixes: 71b77a7a27 ("enetc: Migrate to PHYLINK and PCS_LYNX")
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Quoting from the blamed commit:
In promiscuous mode, it is more intuitive that all traffic is received,
including VLAN tagged traffic. It appears that it is necessary to set
the flag in PSIPVMR for that to be the case, so VLAN promiscuous mode is
also temporarily enabled. On exit from promiscuous mode, the setting
made by ethtool is restored.
Intuitive or not, there isn't any definition issued by a standards body
which says that promiscuity has anything to do with VLAN filtering - it
only has to do with accepting packets regardless of destination MAC address.
In fact people are already trying to use this misunderstanding/bug of
the enetc driver as a justification to transform promiscuity into
something it never was about: accepting every packet (maybe that would
be the "rx-all" netdev feature?):
https://lore.kernel.org/netdev/20201110153958.ci5ekor3o2ekg3ky@ipetronik.com/
This is relevant because there are use cases in the kernel (such as
tc-flower rules with the protocol 802.1Q and a vlan_id key) which do not
(yet) use the vlan_vid_add API to be compatible with VLAN-filtering NICs
such as enetc, so for those, disabling rx-vlan-filter is currently the
only right solution to make these setups work:
https://lore.kernel.org/netdev/CA+h21hoxwRdhq4y+w8Kwgm74d4cA0xLeiHTrmT-VpSaM7obhkg@mail.gmail.com/
The blamed patch has unintentionally introduced one more way for this to
work, which is to enable IFF_PROMISC, however this is non-portable
because port promiscuity is not meant to disable VLAN filtering.
Therefore, it could invite people to write broken scripts for enetc, and
then wonder why they are broken when migrating to other drivers that
don't handle promiscuity in the same way.
Fixes: 7070eea5e9 ("enetc: permit configuration of rx-vlan-filter with ethtool")
Cc: Markus Blöchl <Markus.Bloechl@ipetronik.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When the enetc ports have rx-vlan-offload enabled, they report a TPID of
ETH_P_8021Q regardless of what was actually in the packet. When
rx-vlan-offload is disabled, packets have the proper TPID. Fix this
inconsistency by finishing the TODO left in the code.
Fixes: d4fd0404c1 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The workaround for the ENETC MDIO erratum caused a performance
degradation of 82 Kpps (seen with IP forwarding of two 1Gbps streams of
64B packets). This is due to excessive locking and unlocking in the fast
path, which can be avoided.
By taking the MDIO read-side lock only once per NAPI poll cycle, we are
able to regain 54 Kpps (65%) of the performance hit. The rest of the
performance degradation comes from the TX data path, but unfortunately
it doesn't look like we can optimize that away easily, even with
netdev_xmit_more(), there just isn't any skb batching done, to help with
taking the MDIO lock less often than once per packet.
We need to change the register accessor type for enetc_get_tx_tstamp,
because it now runs under the enetc_lock_mdio as per the new call path
detailed below:
enetc_msix
-> napi_schedule
-> enetc_poll
-> enetc_lock_mdio
-> enetc_clean_tx_ring
-> enetc_get_tx_tstamp
-> enetc_clean_rx_ring
-> enetc_unlock_mdio
Fixes: fd5736bf9f ("enetc: Workaround for MDIO register access issue")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael reports that since linux-next-20210211, the AER messages for ECC
errors have started reappearing, and this time they can be reliably
reproduced with the first ping on one of his LS1028A boards.
$ ping 1[ 33.258069] pcieport 0000:00:1f.0: AER: Multiple Corrected error received: 0000:00:00.0
72.16.0.1
PING [ 33.267050] pcieport 0000:00:1f.0: AER: can't find device of ID0000
172.16.0.1 (172.16.0.1): 56 data bytes
64 bytes from 172.16.0.1: seq=0 ttl=64 time=17.124 ms
64 bytes from 172.16.0.1: seq=1 ttl=64 time=0.273 ms
$ devmem 0x1f8010e10 32
0xC0000006
It isn't clear why this is necessary, but it seems that for the errors
to go away, we must clear the entire RFS and RSS memory, not just for
the ports in use.
Sadly the code is structured in such a way that we can't have unified
logic for the used and unused ports. For the minimal initialization of
an unused port, we need just to enable and ioremap the PF memory space,
and a control buffer descriptor ring. Unused ports must then free the
CBDR because the driver will exit, but used ports can not pick up from
where that code path left, since the CBDR API does not reinitialize a
ring when setting it up, so its producer and consumer indices are out of
sync between the software and hardware state. So a separate
enetc_init_unused_port function was created, and it gets called right
after the PF memory space is enabled.
Fixes: 07bf34a50e ("net: enetc: initialize the RFS and RSS memories")
Reported-by: Michael Walle <michael@walle.cc>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Michael Walle <michael@walle.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>
After the blamed patch, all RX traffic gets hashed to CPU 0 because the
hashing indirection table set up in:
enetc_pf_probe
-> enetc_alloc_si_resources
-> enetc_configure_si
-> enetc_setup_default_rss_table
is overwritten later in:
enetc_pf_probe
-> enetc_init_port_rss_memory
which zero-initializes the entire port RSS table in order to avoid ECC errors.
The trouble really is that enetc_init_port_rss_memory really neads
enetc_alloc_si_resources to be called, because it depends upon
enetc_alloc_cbdr and enetc_setup_cbdr. But that whole enetc_configure_si
thing could have been better thought out, it has nothing to do in a
function called "alloc_si_resources", especially since its counterpart,
"free_si_resources", does nothing to unwind the configuration of the SI.
The point is, we need to pull out enetc_configure_si out of
enetc_alloc_resources, and move it after enetc_init_port_rss_memory.
This allows us to set up the default RSS indirection table after
initializing the memory.
Fixes: 07bf34a50e ("net: enetc: initialize the RFS and RSS memories")
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The following call path suggests that calling unregister_netdev on an
interface that is up will first bring it down.
enetc_pf_remove
-> unregister_netdev
-> unregister_netdevice_queue
-> unregister_netdevice_many
-> dev_close_many
-> __dev_close_many
-> enetc_close
-> enetc_stop
-> phylink_stop
However, enetc first destroys the phylink instance, then calls
unregister_netdev. This is already dissimilar to the setup (and error
path teardown path) from enetc_pf_probe, but more than that, it is buggy
because it is invalid to call phylink_stop after phylink_destroy.
So let's first unregister the netdev (and let the .ndo_stop events
consume themselves), then destroy the phylink instance, then free the
netdev.
Fixes: 71b77a7a27 ("enetc: Migrate to PHYLINK and PCS_LYNX")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
FSL_ENETC_MDIO use symbols from PHYLIB (MDIO_BUS) and MDIO_DEVRES,
however there are no dependency specified in Kconfig
ERROR: modpost: "__mdiobus_register" [drivers/net/ethernet/freescale/enetc/fsl-enetc-mdio.ko] undefined!
ERROR: modpost: "mdiobus_unregister" [drivers/net/ethernet/freescale/enetc/fsl-enetc-mdio.ko] undefined!
ERROR: modpost: "devm_mdiobus_alloc_size" [drivers/net/ethernet/freescale/enetc/fsl-enetc-mdio.ko] undefined!
add depends on MDIO_DEVRES && MDIO_BUS
Signed-off-by: Tong Zhang <ztong0001@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Michael tried to enable Advanced Error Reporting through the ENETC's
Root Complex Event Collector, and the system started spitting out single
bit correctable ECC errors coming from the ENETC interfaces:
pcieport 0000:00:1f.0: AER: Multiple Corrected error received: 0000:00:00.0
fsl_enetc 0000:00:00.0: PCIe Bus Error: severity=Corrected, type=Transaction Layer, (Receiver ID)
fsl_enetc 0000:00:00.0: device [1957:e100] error status/mask=00004000/00000000
fsl_enetc 0000:00:00.0: [14] CorrIntErr
fsl_enetc 0000:00:00.1: PCIe Bus Error: severity=Corrected, type=Transaction Layer, (Receiver ID)
fsl_enetc 0000:00:00.1: device [1957:e100] error status/mask=00004000/00000000
fsl_enetc 0000:00:00.1: [14] CorrIntErr
Further investigating the port correctable memory error detect register
(PCMEDR) shows that these AER errors have an associated SOURCE_ID of 6
(RFS/RSS):
$ devmem 0x1f8010e10 32
0xC0000006
$ devmem 0x1f8050e10 32
0xC0000006
Discussion with the hardware design engineers reveals that on LS1028A,
the hardware does not do initialization of that RFS/RSS memory, and that
software should clear/initialize the entire table before starting to
operate. That comes as a bit of a surprise, since the driver does not do
initialization of the RFS memory. Also, the initialization of the
Receive Side Scaling is done only partially.
Even though the entire ENETC IP has a single shared flow steering
memory, the flow steering service should returns matches only for TCAM
entries that are within the range of the Station Interface that is doing
the search. Therefore, it should be sufficient for a Station Interface
to initialize all of its own entries in order to avoid any ECC errors,
and only the Station Interfaces in use should need initialization.
There are Physical Station Interfaces associated with PCIe PFs and
Virtual Station Interfaces associated with PCIe VFs. We let the PF
driver initialize the entire port's memory, which includes the RFS
entries which are going to be used by the VF.
Reported-by: Michael Walle <michael@walle.cc>
Fixes: d4fd0404c1 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Michael Walle <michael@walle.cc>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Link: https://lore.kernel.org/r/20210204134511.2640309-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Now that there aren't any more macros with parameters, move the macros
above any functions.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The macro enetc_mdio_rd_reg() is just used in that particular case and
has a hardcoded parameter name "mdio_priv". Define a specific function
to use for readx_poll_timeout() instead. Also drop the TIMEOUT macro
since it is used just once.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Before commit 6517798dd3 ("enetc: Make MDIO accessors more generic and
export to include/linux/fsl") these macros actually had some benefits.
But after the commit it just makes the code hard to read. Drop the macro
indirections.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
xdp_return_frame_bulk() needs to pass a xdp_buff
to __xdp_return().
strlcpy got converted to strscpy but here it makes no
functional difference, so just keep the right code.
Conflicts:
net/netfilter/nf_tables_api.c
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When CONFIG_OF is disabled, there is a harmless warning about
an unused variable:
enetc_pf.c: In function 'enetc_phylink_create':
enetc_pf.c:981:17: error: unused variable 'dev' [-Werror=unused-variable]
Slightly rearrange the code to pass around the of_node as a
function argument, which avoids the problem without hurting
readability.
Fixes: 71b77a7a27 ("enetc: Migrate to PHYLINK and PCS_LYNX")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Link: https://lore.kernel.org/r/20201204120800.17193-1-claudiu.manoil@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Trivial conflict in CAN, keep the net-next + the byteswap wrapper.
Conflicts:
drivers/net/can/usb/gs_usb.c
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The tc-taprio base time indicates the beginning of the tc-taprio
schedule, which is cyclic by definition (where the length of the cycle
in nanoseconds is called the cycle time). The base time is a 64-bit PTP
time in the TAI domain.
Logically, the base-time should be a future time. But that imposes some
restrictions to user space, which has to retrieve the current PTP time
from the NIC first, then calculate a base time that will still be larger
than the base time by the time the kernel driver programs this value
into the hardware. Actually ensuring that the programmed base time is in
the future is still a problem even if the kernel alone deals with this.
Luckily, the enetc hardware already advances a base-time that is in the
past into a congruent time in the immediate future, according to the
same formula that can be found in the software implementation of taprio
(in taprio_get_start_time):
/* Schedule the start time for the beginning of the next
* cycle.
*/
n = div64_s64(ktime_sub_ns(now, base), cycle);
*start = ktime_add_ns(base, (n + 1) * cycle);
There's only one problem: the driver doesn't let the hardware do that.
It interferes with the base-time passed from user space, by special-casing
the situation when the base-time is zero, and replaces that with the
current PTP time. This changes the intended effective base-time of the
schedule, which will in the end have a different phase offset than if
the base-time of 0.000000000 was to be advanced by an integer multiple
of the cycle-time.
Fixes: 34c6adf197 ("enetc: Configure the Time-Aware Scheduler via tc-taprio offload")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20201124220259.3027991-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently the control buffer descriptor (cbd) fields have endianness
restrictions while the commands passed into the control buffers
don't (with one exception). This patch fixes offending code,
by adding endianness accessors for cbd fields and removing the
unnecessary ones in case of data buffer fields. Currently there's
no need to convert all commands to little endian format, the patch
only focuses on fixing current endianness issues reported by sparse.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
These particular fields are specified in the H/W reference
manual as having network byte order format, so enforce big
endian annotation for them and clear the related sparse
warnings in the process.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Due to a hardware issue, an access to MDIO registers
that is concurrent with other ENETC register accesses
may lead to the MDIO access being dropped or corrupted.
The workaround introduces locking for all register accesses
to the ENETC register space. To reduce performance impact,
a readers-writers locking scheme has been implemented.
The writer in this case is the MDIO access code (irrelevant
whether that MDIO access is a register read or write), and
the reader is any access code to non-MDIO ENETC registers.
Also, the datapath functions acquire the read lock fewer times
and use _hot accessors. All the rest of the code uses the _wa
accessors which lock every register access.
The commit introducing MDIO support is -
commit ebfcb23d62 ("enetc: Add ENETC PF level external MDIO support")
but due to subsequent refactoring this patch is applicable on
top of a later commit.
Fixes: 6517798dd3 ("enetc: Make MDIO accessors more generic and export to include/linux/fsl")
Signed-off-by: Alex Marginean <alexandru.marginean@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Link: https://lore.kernel.org/r/20201112182608.26177-1-claudiu.manoil@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This is a methodical transition of the driver from phylib
to phylink, following the guidelines from sfp-phylink.rst.
The MAC register configurations based on interface mode
were moved from the probing path to the mac_config() hook.
MAC enable and disable commands (enabling Rx and Tx paths
at MAC level) were also extracted and assigned to their
corresponding phylink hooks.
As part of the migration to phylink, the serdes configuration
from the driver was offloaded to the PCS_LYNX module,
introduced in commit 0da4c3d393 ("net: phy: add Lynx PCS module"),
the PCS_LYNX module being a mandatory component required to
make the enetc driver work with phylink.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.cionei@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Decouple internal mdio bus creation from serdes
configuration, as a prerequisite to offloading
serdes configuration to a different module.
Group together mdio bus creation routines, cleanup.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Decouple level MAC configuration based on phy interface type
from general port configuration.
Group together MAC and link configuration code.
Decouple external mdio bus creation from interface type
parsing. No longer return an (unhandled) error code when
phy_node not found, use phy_node to indicate whether the
port has a phy or not. No longer fall-through when serdes
configuration fails for the link modes that require
internal link configuration.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Reviewed-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Two minor conflicts:
1) net/ipv4/route.c, adding a new local variable while
moving another local variable and removing it's
initial assignment.
2) drivers/net/dsa/microchip/ksz9477.c, overlapping changes.
One pretty prints the port mode differently, whilst another
changes the driver to try and obtain the port mode from
the port node rather than the switch node.
Signed-off-by: David S. Miller <davem@davemloft.net>
This is the correct resolution for the conflict from
merging the "net" tree fix:
commit 26cb7085c8 ("enetc: Remove the mdio bus on PF probe bailout")
with the "net-next" new work:
commit 07095c025a ("net: enetc: Use DT protocol information to set up the ports")
that moved mdio bus allocation to an ealier stage of
the PF probing routine.
Fixes: a57066b1a0 ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The driver calls napi_schedule_irqoff() from a context where, in RT,
hardirqs are not disabled, since the IRQ handler is force-threaded.
In the call path of this function, __raise_softirq_irqoff() is modifying
its per-CPU mask of pending softirqs that must be processed, using
or_softirq_pending(). The or_softirq_pending() function is not atomic,
but since interrupts are supposed to be disabled, nobody should be
preempting it, and the operation should be safe.
Nonetheless, when running with hardirqs on, as in the PREEMPT_RT case,
it isn't safe, and the pending softirqs mask can get corrupted,
resulting in softirqs being lost and never processed.
To have common code that works with PREEMPT_RT and with mainline Linux,
we can use plain napi_schedule() instead. The difference is that
napi_schedule() (via __napi_schedule) also calls local_irq_save, which
disables hardirqs if they aren't already. But, since they already are
disabled in non-RT, this means that in practice we don't see any
measurable difference in throughput or latency with this patch.
Signed-off-by: Jiafei Pan <Jiafei.Pan@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the case of invalid rule, a positive value EINVAL is returned here.
I think this is a typo error. It is necessary to return an error value.
Cc: Po Liu <Po.Liu@nxp.com>
Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The UDP reuseport conflict was a little bit tricky.
The net-next code, via bpf-next, extracted the reuseport handling
into a helper so that the BPF sk lookup code could invoke it.
At the same time, the logic for reuseport handling of unconnected
sockets changed via commit efc6b6f6c3
which changed the logic to carry on the reuseport result into the
rest of the lookup loop if we do not return immediately.
This requires moving the reuseport_has_conns() logic into the callers.
While we are here, get rid of inline directives as they do not belong
in foo.c files.
The other changes were cases of more straightforward overlapping
modifications.
Signed-off-by: David S. Miller <davem@davemloft.net>
For ENETC ports that register an external MDIO bus,
the bus doesn't get removed on the error bailout path
of enetc_pf_probe().
This issue became much more visible after recent:
commit 07095c025a ("net: enetc: Use DT protocol information to set up the ports")
Before this commit, one could make probing fail on the error
path only by having register_netdev() fail, which is unlikely.
But after this commit, because it moved the enetc_of_phy_get()
call up in the probing sequence, now we can trigger an mdiobus_free()
bug just by forcing enetc_alloc_msix() to return error, i.e. with the
'pci=nomsi' kernel bootarg (since ENETC relies on MSI support to work),
as the calltrace below shows:
kernel BUG at /home/eiz/work/enetc/net/drivers/net/phy/mdio_bus.c:648!
Internal error: Oops - BUG: 0 [#1] PREEMPT SMP
[...]
Hardware name: LS1028A RDB Board (DT)
pstate: 80000005 (Nzcv daif -PAN -UAO BTYPE=--)
pc : mdiobus_free+0x50/0x58
lr : devm_mdiobus_free+0x14/0x20
[...]
Call trace:
mdiobus_free+0x50/0x58
devm_mdiobus_free+0x14/0x20
release_nodes+0x138/0x228
devres_release_all+0x38/0x60
really_probe+0x1c8/0x368
driver_probe_device+0x5c/0xc0
device_driver_attach+0x74/0x80
__driver_attach+0x8c/0xd8
bus_for_each_dev+0x7c/0xd8
driver_attach+0x24/0x30
bus_add_driver+0x154/0x200
driver_register+0x64/0x120
__pci_register_driver+0x44/0x50
enetc_pf_driver_init+0x24/0x30
do_one_initcall+0x60/0x1c0
kernel_init_freeable+0x1fc/0x274
kernel_init+0x14/0x110
ret_from_fork+0x10/0x34
Fixes: ebfcb23d62 ("enetc: Add ENETC PF level external MDIO support")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
enetc_imdio_remove() is missing from the enetc_pf_probe()
bailout path. Not surprisingly because enetc_setup_serdes()
is registering the imdio bus for internal purposes, and it's
not obvious that enetc_imdio_remove() currently performs the
teardown of enetc_setup_serdes().
To fix this, define enetc_teardown_serdes() to wrap
enetc_imdio_remove() (improve code maintenance) and call it
on bailout and remove paths.
Fixes: 975d183ef0 ("net: enetc: Initialize SerDes for SGMII and USXGMII protocols")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the generic dynamic interrupt moderation (dim)
framework to implement adaptive interrupt coalescing
on Rx. With the per-packet interrupt scheme, a high
interrupt rate has been noted for moderate traffic flows
leading to high CPU utilization. The 'dim' scheme
implemented by the current patch addresses this issue
improving CPU utilization while using minimal coalescing
time thresholds in order to preserve a good latency.
On the Tx side use an optimal time threshold value by
default. This value has been optimized for Tx TCP
streams at a rate of around 85kpps on a 1G link,
at which rate half of the Tx ring size (128) gets filled
in 1500 usecs. Scaling this down to 2.5G links yields
the current value of 600 usecs, which is conservative
and gives good enough results for 1G links too (see
next).
Below are some measurement results for before and after
this patch (and related dependencies) basically, for a
2 ARM Cortex-A72 @1.3Ghz CPUs system (32 KB L1 data cache),
using 60secs log netperf TCP stream tests @ 1Gbit link
(maximum throughput):
1) 1 Rx TCP flow, both Rx and Tx processed by the same NAPI
thread on the same CPU:
CPU utilization int rate (ints/sec)
Before: 50%-60% (over 50%) 92k
After: 13%-22% 3.5k-12k
Comment: Major CPU utilization improvement for a single flow
Rx TCP flow (i.e. netperf -t TCP_MAERTS) on a single
CPU. Usually settles under 16% for longer tests.
2) 4 Rx TCP flows + 4 Tx TCP flows (+ pings to check the latency):
Total CPU utilization Total int rate (ints/sec)
Before: ~80% (spikes to 90%) ~100k
After: 60% (more steady) ~4k
Comment: Important improvement for this load test, while the
ping test outcome does not show any notable
difference compared to before.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Enable programming of the interrupt coalescing registers
and allow manual configuration of the coalescing time
thresholds via ethtool. Packet thresholds have been fixed
to predetermined values as there's no point in making them
run-time configurable, also anticipating the dynamic interrupt
moderation (DIM) algorithm which uses fixed packet thresholds
as well. If the interface is up when the operation mode of
traffic interrupt events is changed by the user (i.e. switching
from default per-packet interrupts to coalesced interrupts),
the traffic needs to be paused in the process.
This patch also prepares the ground for introducing DIM on Rx.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
'struct enetc_bdr' is already '____cacheline_aligned_in_smp'.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Interrupt coalescing registers naming in the current revision
of the Ref Man (RM) is ICR, deprecating the ICIR name used
in earlier (draft) versions of the RM.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A reliable traffic pause (and reconfiguration) procedure
is needed to be able to safely make h/w configuration
changes during run-time, like changing the mode in which the
interrupts are operating (i.e. with or without coalescing),
as opposed to making on-the-fly register updates that
may be subject to h/w or s/w concurrency issues.
To this end, the code responsible of the run-time device
configurations that basically starts resp. stops the traffic
flow through the device has been extracted from the
the enetc_open/_close procedures, to the separate standalone
enetc_start/_stop procedures. Traffic stop should be as
graceful as possible, it lets the executing napi threads to
to finish while the interrupts stay disabled. But since
the napi thread will try to re-enable interrupts by clearing
the device's unmask register, the enable_irq/ disable_irq
API has been used to avoid this potential concurrency issue
and make the traffic pause procedure more reliable.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It's time to differentiate between Rx and Tx ring sizes.
Not only Tx rings are processed differently than Rx rings,
but their default number also differs - i.e. up to 8 Tx rings
per device (8 traffic classes) vs. 2 Rx rings (one per CPU).
So let's set Tx rings sizes to half the size of the Rx rings
for now, to be conservative.
The default ring sizes were decreased as well (to the next
lower power of 2), to reduce the memory footprint, buffering
etc., since the measurements I've made so far show that the
rings are very unlikely to get full.
This change also anticipates the introduction of the
dynamic interrupt moderation (dim) algorithm which operates
on maximum packet thresholds of 256 packets for Rx and 128
packets for Tx.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use DT information rather than in-band information from bootloader to
set up MAC for XGMII. For RGMII use the DT indication in addition to
RGMII defaults in hardware.
However, this implies that PHY connection information needs to be
extracted before netdevice creation, when the ENETC Port MAC is
being configured.
Signed-off-by: Alex Marginean <alexandru.marginean@nxp.com>
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Michael Walle <michael@walle.cc>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ENETC has ethernet MACs capable of SGMII, 2500BaseX and USXGMII. But in
order to use these protocols some SerDes configurations need to be
performed. The SerDes is configurable via an internal PCS PHY which is
connected to an internal MDIO bus at address 0.
This patch basically removes the dependency on bootloader regarding
SerDes initialization.
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is to use eth_broadcast_addr() to assign broadcast address
insetad of memset().
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
