Remove the local variable "array" and all the memcpy function calls
because this copy operation from different arrays to this variable is
unnecessary.
The vnt_control_out function already does a kmemdup copy of its const
char *buffer argument and this was made unnecessary by:
commit 12ecd24ef9
("staging: vt6656: use off stack for out buffer USB transfers.")
Author: Malcolm Priestley <tvboxspy@gmail.com>
Date: Sat Apr 22 11:14:57 2017 +0100
staging: vt6656: use off stack for out buffer USB transfers.
Since 4.9 mandated USB buffers be heap allocated this causes the driver
to fail.
Since there is a wide range of buffer sizes use kmemdup to create
allocated buffer.
So, the same result can be achieved using the arrays directly.
Signed-off-by: Oscar Carter <oscar.carter@gmx.com>
Link: https://lore.kernel.org/r/20200425151747.8199-2-oscar.carter@gmx.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove debug print statements referring to non-existent fields
'lbq_clean_idx' and 'lbq_free_cnt' in the 'rx_ring' struct, which causes
a compilation failure when QL_DEV_DUMP is set.
These fields were initially removed as a part of commit aec626d209
("staging: qlge: Update buffer queue prod index despite oom") in 2019.
Their replacement fields ('next_to_use' and 'next_to_clean') are already
being printed, so this patch does not add new debug statements for them.
Signed-off-by: Rylan Dmello <mail@rylan.coffee>
Link: https://lore.kernel.org/r/aa7e0197f4e34cec0855124e45696e33dd9527e5.1587959245.git.mail@rylan.coffee
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove statement that tries to print the non-existent 'vlgrp' field
in the 'ql_adapter' struct, which causes a compilation failure when
QL_DEV_DUMP is set.
vlgrp seems to have been removed from ql_adapter as a part of
commit 18c49b9177 ("qlge: do vlan cleanup") in 2011.
vlgrp might be replaced by the 'active_vlans' array introduced in the
aforementioned commit. But I'm not sure if printing all 64 values of
that array would help with debugging this driver, so I'm leaving it
out of the debug code in this patch.
Signed-off-by: Rylan Dmello <mail@rylan.coffee>
Link: https://lore.kernel.org/r/51bae37a54d414491779e4a3329508cc864ab900.1587959245.git.mail@rylan.coffee
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Check the return value of vnt_control_out_* function calls. When
necessary modify the function prototype to be able to return the new
checked error code.
It's safe to modify all the function prototypes without fix the call
because the only change is the return value from void to int. If before
the call didn't check the return value, now neither.
Signed-off-by: Oscar Carter <oscar.carter@gmx.com>
Link: https://lore.kernel.org/r/20200425134257.4502-2-oscar.carter@gmx.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The device raises error indications when it thinks there is a bug in the
driver and it can't recover it (while it raises exception when a bug is
detected in the device).
The current list of of errors was a bit dated.
This patch cleans up the list of errors and the associated message. It
is also the right time to clean up the way the error indications are
handled. Replace the switch..case with a clean loop over an array.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-18-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The device keep up to date three series of stats. One for each
virtual interface and one for the whole device.
Until to now, the stats for the whole device were unavailable. Moreover,
it is interesting to retrieve counters for all interfaces even if they
are not awake.
Change the counters available in debugfs in order to retrieve stats
from all interfaces.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-13-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In some circumstances, Tx traffic is sent without associated station but
the station exists when the Tx status is received. Beside that, the
driver keep a counter associated to each station. So, in this case, the
counter is not incremented, but is decremented. In this case a warning
"inconsistent notification" appears:
------------[ cut here ]------------
WARNING: CPU: 3 PID: 82 at /home/jerome/wfx/data_tx.c:469 wfx_skb_dtor+0x1a4/0x1d4 [wfx]
inconsistent notification
Modules linked in: [...]
CPU: 3 PID: 82 Comm: kworker/3:1H Tainted: G C O 4.19.57-v7l+ #1244
Hardware name: BCM2835
Workqueue: events_highpri bh_work [wfx]
[<c0212c8c>] (unwind_backtrace) from [<c020d49c>] (show_stack+0x20/0x24)
[<c020d49c>] (show_stack) from [<c0976220>] (dump_stack+0xd4/0x118)
[<c0976220>] (dump_stack) from [<c0222270>] (__warn+0x104/0x11c)
[<c0222270>] (__warn) from [<c02222e0>] (warn_slowpath_fmt+0x58/0x74)
[<c02222e0>] (warn_slowpath_fmt) from [<bf497b48>] (wfx_skb_dtor+0x1a4/0x1d4 [wfx])
[<bf497b48>] (wfx_skb_dtor [wfx]) from [<bf4988b4>] (wfx_tx_confirm_cb+0x198/0x2f0 [wfx])
[<bf4988b4>] (wfx_tx_confirm_cb [wfx]) from [<bf49d054>] (hif_tx_confirm+0x50/0x70 [wfx])
[<bf49d054>] (hif_tx_confirm [wfx]) from [<bf49d42c>] (wfx_handle_rx+0x128/0x22c [wfx])
[<bf49d42c>] (wfx_handle_rx [wfx]) from [<bf4953cc>] (bh_work+0x3cc/0x964 [wfx])
[<bf4953cc>] (bh_work [wfx]) from [<c023dab8>] (process_one_work+0x170/0x458)
[<c023dab8>] (process_one_work) from [<c023ddfc>] (worker_thread+0x5c/0x5a4)
[<c023ddfc>] (worker_thread) from [<c02440e8>] (kthread+0x138/0x168)
[<c02440e8>] (kthread) from [<c02010ac>] (ret_from_fork+0x14/0x28)
Exception stack(0xee199fb0 to 0xee199ff8)
9fa0: 00000000 00000000 00000000 00000000
9fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
9fe0: 00000000 00000000 00000000 00000000 00000013 00000000
---[ end trace 64b9e754e12ef7de ]---
This patch fix this race between the station creation and the Tx data.
Fixes: 7d2d2bfdeb ("staging: wfx: relocate "buffered" information to sta_priv")
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-11-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It has been reported that trying to send small packets of data could
produce a "inconsistent notification" warning.
It seems that in some circumstances, the number of frame queued in the
driver could greatly increase and exceed UCHAR_MAX. So the field
"buffered" from struct sta_priv can overflow.
Just increase the size of "bueffered" to fix the problem.
Fixes: 7d2d2bfdeb ("staging: wfx: relocate "buffered" information to sta_priv")
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-10-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Device is able to detect a high temperature. In this case, the traffic
is not allowed to be sent until the temperature decrease.
This patch detects the warnings raised by the device and stop the
traffic accordingly. It also add a delayed task as safeguard in case the
chip would never send the indication that the temperature decrease.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-7-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
wfx_suspend_resume_mc() is called when the device is about to sent a
DTIM. This is the right moment to enqueue Content After DTIM Beacon
(CAB).
However, wfx_suspend_resume_mc() is also called when the DTIM period
ends. Until now, this event did also trig CAB.
Note this issue did not have too much impact since when a CAB is sent
outside of DTIM window, an error is reported by the firmware and
mac80211 retries to send the data.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-6-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When multiple virtual interfaces (on different channels) are in use, the
device ask to activate Power Save on station interfaces. The device
developers recommends to use legacy PS-Poll in this case since it is the
mode that disturb the less the other interface. However, some AP start
to not answer anymore to PS-Poll. The device is able to detect this case
and return a special warning in this case.
So, this commit catch the warning and force usage of FastPS in this
case.
In order to confuse the less possible the other interface a small FastPS
period is used (30ms).
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20200427134031.323403-5-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
buf_in_between() gets passed q->u.in.ack_start as 'bufnr' parameter.
The ack_start always ranges between 0 and QDIO_MAX_BUFFERS_PER_Q - 1,
so the subsequent check will always return true. Remove it.
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>
Except for some initial thinint-only steps, the processing is identical
to the non-thinint case. So re-use the existing helper.
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>
Knowing how many queues we initially allocated allows us to
1) sanity-check a subsequent qdio_establish() request, and
2) walk the queue arrays without further checks. Apply this while
cleanly splitting qdio_free_queues() into two separate helpers.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
When qdio_allocate_qs() fails, have it deal with its previous
allocations.
This way qdio_allocate() doesn't need to clean up afterwards.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Instead of having a catch-all qdio_release_memory() helper, free the
individual allocations from the respective error path.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
Wrap the init/exit steps for thinint into a single helper that follows
the established naming scheme.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
qdio_establish() calls qdio_setup_thinint() via qdio_setup_irq().
If the subsequent qdio_establish_thinint() fails, we miss to put the
DSCI again. Thus the DSCI isn't available for re-use. Given enough of
such errors, we could end up with having only the shared DSCI available.
Merge qdio_setup_thinint() into qdio_establish_thinint(), and deal with
such an error internally.
Fixes: 779e6e1c72 ("[S390] qdio: new qdio driver.")
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>
qdio_establish() calls qdio_establish_thinint(), but later has an error
exit path that doesn't roll this call back. Fix it.
Fixes: 779e6e1c72 ("[S390] qdio: new qdio driver.")
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>
For rolling back after an error, qdio_establish() calls qdio_shutdown().
If the error occurs early enough, then the qdio_irq's state still is
QDIO_IRQ_STATE_INACTIVE and qdio_shutdown() does nothing.
But at _any_ point where qdio_establish() bails out in this way,
qdio_setup_irq() will have already replaced the IRQ handler. This then
won't be restored after an early error, and the device can end up being
returned to the device driver with qdio's IRQ handler still installed.
Slightly reorder qdio_setup_irq() so we can be 100% sure that the IRQ
handler was replaced. Then fix the bug in qdio_establish() by calling a
helper that rolls back only the IRQ handler modification.
Also use the new helper in qdio_shutdown() to keep things in sync, and
slightly clean up the locking while doing so.
This makes minor semantical changes, but holding setup_mutex gives us
sufficient leeway to eg. pull qdio_shutdown_thinint() outside of the
ccwdev lock's scope.
Fixes: 779e6e1c72 ("[S390] qdio: new qdio driver.")
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>
There are changes in the usage of PCI for the user:
- new kernel parameter
- modification of the way functions are enumerated
Let's document these.
Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
The Physical function should not be disabled until no virtual
functions depends on it.
Let's force the user to first use echo 0 > sriov_numfs before
allowing to disable the PF with echo 0 > power.
Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
