When qeth_iqd_tx_complete() detects that a TX buffer requires additional
async completion via QAOB, it might fail to replace the queue entry's
metadata (and ends up triggering recovery).
Assume now that the device gets torn down, overruling the recovery.
If the QAOB notification then arrives before the tear down has
sufficiently progressed, the buffer state is changed to
QETH_QDIO_BUF_HANDLED_DELAYED by qeth_qdio_handle_aob().
The tear down code calls qeth_drain_output_queue(), where
qeth_cleanup_handled_pending() will then attempt to replace such a
buffer _again_. If it succeeds this time, the buffer ends up dangling in
its replacement's ->next_pending list ... where it will never be freed,
since there's no further call to qeth_cleanup_handled_pending().
But the second attempt isn't actually needed, we can simply leave the
buffer on the queue and re-use it after a potential recovery has
completed. The qeth_clear_output_buffer() in qeth_drain_output_queue()
will ensure that it's in a clean state again.
Fixes: 72861ae792 ("qeth: recovery through asynchronous delivery")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The two expected notification sequences are
1. TX_NOTIFY_PENDING with a subsequent TX_NOTIFY_DELAYED_*, when
our TX completion code first observed the pending TX and the QAOB
then completes at a later time; or
2. TX_NOTIFY_OK, when qeth_qdio_handle_aob() picked up the QAOB
completion before our TX completion code even noticed that the TX
was pending.
But as qeth_iqd_tx_complete() and qeth_qdio_handle_aob() can run
concurrently, we may end up with a race that results in a sequence of
TX_NOTIFY_DELAYED_* followed by TX_NOTIFY_PENDING. Which would confuse
the af_iucv code in its tracking of pending transmits.
Rework the notification code, so that qeth_qdio_handle_aob() defers its
notification if the TX completion code is still active.
Fixes: b333293058 ("qeth: add support for af_iucv HiperSockets transport")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Calling into socket code is ugly already, at least check whether we are
dealing with the expected sk_family. Only looking at skb->protocol is
bound to cause troubles (consider eg. af_packet).
Fixes: b333293058 ("qeth: add support for af_iucv HiperSockets transport")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
- Add/delete some blanks, white spaces and braces.
- Fix misindentations.
- Adjust a deprecated header include, and htons() conversion.
- Remove extra 'return' statements.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace our custom version of netdev_name().
Once we started to allocate the netdev at probe time with
commit d3d1b205e8 ("s390/qeth: allocate netdevice early"), this
stopped working as intended anyway.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For OSA devices that are _not_ configured in prio-queue mode, give users
the option of selecting the number of active TX queues.
This requires setting up the HW queues with a reasonable default QoS
value in the QIB's PQUE parm area.
As with the other device types, we bring up the device with a minimal
number of TX queues for compatibility reasons.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use a proper struct, and only program the QIB extensions for devices
where they are supported.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When re-initializing a device, we can hit a situation where
qeth_osa_set_output_queues() detects that it supports more or less
HW TX queues than before. Right now we adjust dev->real_num_tx_queues
from right there, but
1. it's getting more & more complicated to cover all cases, and
2. we can't re-enable the actually expected number of TX queues later
because we lost the needed information.
So keep track of the wanted TX queues (on initial setup, and whenever
its changed via .set_channels), and later use that information when
re-enabling the netdevice.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Clarify which discipline-specific steps are needed to roll back after
error in qeth_l?_set_online(), and which are common to roll back
from qeth_hardsetup_card().
Some steps (cancelling the RX modeset, draining the TX queues) are only
necessary if the netdev was potentially UP before, so move them to the
common qeth_set_offline().
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move duplicated code from the disciplines into the core path.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Originators of cmd IO typically hold the rtnl or conf_mutex to protect
against a concurrent teardown.
Since qeth_set_offline() already holds the conf_mutex, the main reason
why we still care about cancelling pending cmds is so that they release
the rtnl when we need it ourselves.
So move this step a little earlier into the teardown sequence.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch detects whether device-to-bridge-notification, provided
by the Perform Network Subchannel Operation (PNSO) operation code
ADDR_INFO (OC3), is supported by this card. A following patch will
map this to the learning_sync bridgeport flag, so we store it in
priv->brport_hw_features in bridgeport flag format.
Only IQD cards provide PNSO.
There is a feature bit to indicate whether the machine provides OC3,
unfortunately it is not set on old machines.
So PNSO is called to find out. As this will disable notification
and is exclusive with bridgeport_notification, this must be done
during card initialisation before previous settings are restored.
PNSO functionality requires some configuration values that are added to
the qeth_card.info structure. Some helper functions are defined to fill
them out when the card is brought online and some other places are
adapted, that can also benefit from these fields.
Signed-off-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We got slightly different patches removing a double word
in a comment in net/ipv4/raw.c - picked the version from net.
Simple conflict in drivers/net/ethernet/ibm/ibmvnic.c. Use cached
values instead of VNIC login response buffer (following what
commit 507ebe6444 ("ibmvnic: Fix use-after-free of VNIC login
response buffer") did).
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
queue->state is a ternary spinlock in disguise, used by
OSA's TX completion path to lock the Output Queue and flush any pending
packets on it to the device. If the Queue is already locked by our TX
code, setting the lock word to QETH_OUT_Q_LOCKED_FLUSH lets the TX
completion code move on - the TX path will later take care of things
when it unlocks the Queue.
This sort of DIY locking is a non-starter of course, just let the
TX completion path block on the spinlock when necessary. If that ends up
causing additional latency due to lock contention, then converting
the OSA path to use xmit_more is the right way to go forward.
Also slightly expand the locked section and capture all of
qeth_do_send_packet(), so that the update for the 'bufs_pack' statistics
is done race-free.
While reworking the TX completion path's code, remove a barrier() that
doesn't make any sense.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Avoid poking around in the delayed_work struct's internals.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The (misplaced) comment doesn't make any sense, enforcing an
uninitialized RX buffer won't help with IRQ reduction.
So make the best use of all available RX buffers.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Running a RX refill outside of NAPI context is inherently racy, even
though the worker is only started for an entirely idle RX ring.
>From the moment that the worker has replenished parts of the RX ring,
the HW can use those RX buffers, raise an IRQ and cause our NAPI code to
run concurrently to the RX refill worker.
Instead let the worker schedule our NAPI instance, and refill the RX
ring from there. Keeping accurate count of how many buffers still need
to be refilled also removes some quirky arithmetic from the low-level
code.
Fixes: b333293058 ("qeth: add support for af_iucv HiperSockets transport")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When preparing a buffer for RX refill, tolerate that it already has a
pool_entry attached. Otherwise we could easily leak such a pool_entry
when re-driving the RX refill after an error (from eg. do_qdio()).
This needs some minor adjustment in the code that drains RX buffer(s)
prior to RX refill and during teardown, so that ->pool_entry is NULLed
accordingly.
Fixes: 4a71df5004 ("qeth: new qeth device driver")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We're not modifying these data blobs, so mark them as constant.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
To keep track of the addresses programmed from an RX modeset, we have
two separate hashtables (L2: mac_htable, L3: ip_mc_htable).
These are never used at the same time, so unify them into a single
rx_mode_addrs hashtable.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While initially just trying to fix up the indentation, condense a few
lines and get rid of a goto label.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the correct struct member instead of hardcoding its offset.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the correct helper for casting to a user pointer.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As the cmd IO path has learned to propagate errnos back to its callers,
let them deal with errors instead of trying to restore their previous
configuration from within the IO error path.
Also translate the HW error to a meaningful errno, instead of returning
-EIO for all cases (and don't map this to -EOPNOTSUPP later on...).
While at it, add a READ_ONCE() / WRITE_ONCE() pair to ensure that the
data path always sees a valid isolation mode during reconfiguration.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When qeth_set_access_ctrl_online() is called during the device's
initialization and discovers that isolation mode isn't supported, don't
clear the user's currently configured mode.
They intentionally choose to operate the device in this specific mode,
and degrading the isolation is not an option.
Only adjust the configuration when called via sysfs (ie. fallback = 1),
and here follow the common pattern and restore it from prev_isolation.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A newly initialized device defaults to ISOLATION_MODE_NONE, don't bother
with programming this a second time.
Then remove the OSD/OSX check, it's already done in the sysfs path
whenever the user actually changes the configuration.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we cancel all pending cmds (eg. when tearing down the device), don't
blame it on an IO error.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rather than delaying the decision until netdev setup, immediately reject
a device when we discover that it has an unsupported link type
(ie. Token Ring).
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a device is configured with ISOLATION_MODE_FWD, traffic never goes
through the internal switch. Don't apply the offload restrictions in
this case.
Fixes: c619e9a6f5 ("s390/qeth: don't use restricted offloads for local traffic")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current(?) OSA devices also store their cmd-specific return codes for
SET_ACCESS_CONTROL cmds into the top-level cmd->hdr.return_code.
So once we added stricter checking for the top-level field a while ago,
none of the error logic that rolls back the user's configuration to its
old state is applied any longer.
For this specific cmd, go back to the old model where we peek into the
cmd structure even though the top-level field indicated an error.
Fixes: 686c97ee29 ("s390/qeth: fix error handling in adapter command callbacks")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 5e1fb45ec8 ("s390/ccwgroup: remove pm support") removed power
management support from the ccwgroup bus driver. So remove the
associated callbacks from all ccwgroup drivers.
CC: Vineeth Vijayan <vneethv@linux.ibm.com>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When starting the reset worker via sysfs is unsuccessful, return an
error to the user.
Modernize the sysfs input parsing while at it.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When qeth_flush_buffers() gets called for a group of TX buffers
(currently up to 2 for OSA-style devices), the code iterates over each
buffer for some final processing.
During this processing, it sets the TX IRQ marker on the leading buffer
rather than the last one. This can result in delayed TX completion of
the trailing buffers. So pull the IRQ marker code out of the loop, and
apply it to the final buffer.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The TX path usually maps the full content of a page into a buffer
element. But there's specific skb layouts (ie. linearized TSO skbs)
where the HW header (1) requires a separate buffer element, and (2) is
page-contiguous with the packet data that's mapped into the next buffer
element.
Flag such buffer elements accordingly, so that HW can optimize its data
access for them.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Merge the __qeth_fill_buffer() helper into its only caller. This way all
mapping-related context is in one place, and we can make some more use
of it in a subsequent patch.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current OSA models don't support TSO for traffic to local next-hops, and
some old models didn't offer TX CSO for such packets either.
So as part of .ndo_features_check, check if a packet's next-hop resides
on the same OSA Adapter. Opt out from affected HW offloads accordingly.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For debugging purposes, provide read access to the local_addr caches
via debug/qeth/<dev_name>/local_addrs.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In configurations where specific HW offloads are in use, OSA adapters
will raise notifications to their virtual devices about the IP addresses
that currently reside on the same adapter.
Cache these addresses in two RCU-enabled hash tables, and flush the
tables once the relevant HW offload(s) get disabled.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When enabling TX CSO, make a note of whether the device has support for
LP2LP offloading. This will become relevant in subsequent patches.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With the introduction of TX coalescing, .ndo_start_xmit now potentially
starts the TX completion timer. So only kill the timer _after_ TX has
been disabled.
Fixes: ee1e52d1e4 ("s390/qeth: add TX IRQ coalescing support for IQD devices")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Upper-layer drivers allocate their SBALs by calling qdio_alloc_buffers()
for each individual queue. But when later passing the SBAL addresses to
qdio_establish(), they need to be in a single array of pointers.
So if the driver uses multiple Input or Output queues, it needs to
allocate a temporary array just to present all its SBAL pointers in this
layout.
This patch slightly changes the format of the QDIO initialization data,
so that drivers can pass a per-queue array where each element points to
a queue's SBAL array.
zfcp doesn't use multiple queues, so the impact there is trivial.
For qeth this brings a nice reduction in complexity, and removes
a page-sized allocation.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
All that qdio_allocate() actually uses from the init_data is the cdev,
and the number of Input and Output Queues. Have the driver pass those as
parameters, and defer the init_data processing into qdio_establish().
This includes writing per-device(!) trace entries, and most of the
sanity checks.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
The last machine generation that supports OSN is z13, and OSX is only
supported up to z14. Allow users and distros to decide whether they
still need support for these device types.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace list_for_each() with list_for_each_entry(), and
list_entry(head.next) with list_first_entry().
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a device is configured in prio-queue mode to pin all traffic onto
a specific HW queue, treat this as a distinct variant of prio-queueing
instead of QETH_NO_PRIO_QUEUEING.
This corrects an error message from qeth_osa_set_output_queues() for
devices configured in such a mode.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since IQD devices complete (most of) their transmissions synchronously,
they don't offer TX completion IRQs and have no HW coalescing controls.
But we can fake the easy parts in SW, and give the user some control wrt
to how often the TX NAPI code should be triggered to process the TX
completions.
Having per-queue controls can in particular help the dedicated mcast
queue, as it likely benefits from different fine-tuning than what the
ucast queues need.
CC: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
