Commit b67e830d38 ("serial: 8250: 8250_omap: Fix possible interrupt
storm on K3 SoCs") introduced fixup including a register read to
RX_LVL, however, we should be using word offset than byte offset
since our registers are on 4 byte boundary (port.regshift = 2) for
8250_omap.
Fixes: b67e830d38 ("serial: 8250: 8250_omap: Fix possible interrupt storm on K3 SoCs")
Cc: stable <stable@vger.kernel.org>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Vignesh Raghavendra <vigneshr@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20210903050550.29050-1-nm@ti.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This device does not support changing baud, parity, data bits, stop
bits, or detecting breaks. Disable "changing" these settings to prevent
their termios from diverging from the actual state of the uart.
In order to determine the correct parameters to enforce, we read the
various new devicetree parameters to discover how the uart was
configured when it was synthesized. The defaults match
ulite_console_setup. xlnx,use-parity, xlnx,odd-parity, and
xlnx,data-bits are optional since there were in-tree users (and
presumably out-of-tree users) who did not set them.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Link: https://lore.kernel.org/r/20210826192154.3202269-5-sean.anderson@seco.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
These properties are necessary for properly calculating the uart
timeout. I inspected the J2 source code, and believe these values to be
correct.
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Link: https://lore.kernel.org/r/20210826192154.3202269-4-sean.anderson@seco.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The uartlite device is a "soft" device. Many parameters, such as baud
rate, data bits, and the presence of a parity bit are configured before
synthesis and may not be changed (or discovered) at runtime. However, we
must know what these settings are in order to properly calculate the
uart timeout (and to inform the user about the actual baud of the uart).
These properties are present for out-of-tree bindings generated by
Xilinx's tools. However, they are also (mostly) present in in-tree
bindings. I chose current-speed over xlnx,baudrate primarily because it
seemed to be used by more existing bindings. Although these properties
are marked as "required", note that only current-speed is required by
the driver itself. Hopefully, this will allow for an easier transition.
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Link: https://lore.kernel.org/r/20210826192154.3202269-3-sean.anderson@seco.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This converts the existing documentation for the uartlite binding to
json-schema.
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Link: https://lore.kernel.org/r/20210826192154.3202269-2-sean.anderson@seco.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.
So, use the struct_size() helper to do the arithmetic instead of the
argument "size + size * count" in the kzalloc() function.
[1] https://www.kernel.org/doc/html/v5.14/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Signed-off-by: Len Baker <len.baker@gmx.com>
Link: https://lore.kernel.org/r/20210905155728.11147-1-len.baker@gmx.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This helps validating DTS files.
Redundant generic descriptions (e.g. "The base address of the UART
register bank") were dropped and periph_clk was dropped from example as
not required.
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Link: https://lore.kernel.org/r/20210829131927.13682-1-zajec5@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Seeing these errors when GT is likely in suspend state-
"RPM wakelock ref not held during HW access"
Ensure GT is awake before trying to access HW registers. Avoid
reading the register if that is not the case.
Signed-off-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Fixes: 41e5c17ebf ("drm/i915/guc/slpc: Sysfs hooks for SLPC")
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210907232704.12982-1-vinay.belgaumkar@intel.com
(cherry picked from commit f25e3908b9)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
gem context refcounting is another exercise in least locking design it
seems, where most things get destroyed upon context closure (which can
race with anything really). Only the actual memory allocation and the
locks survive while holding a reference.
This tripped up Jason when reimplementing the single timeline feature
in
commit 00dae4d3d3
Author: Jason Ekstrand <jason@jlekstrand.net>
Date: Thu Jul 8 10:48:12 2021 -0500
drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)
We could fix the bug by holding ctx->mutex in execbuf and clear the
pointer (again while holding the mutex) context_close, but it's
cleaner to just make the context object actually invariant over its
_entire_ lifetime. This way any other ioctl that's potentially racing,
but holding a full reference, can still rely on ctx->syncobj being
an immutable pointer. Which without this change, is not the case.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Fixes: 00dae4d3d3 ("drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)")
Cc: Jason Ekstrand <jason@jlekstrand.net>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: "Thomas Hellström" <thomas.hellstrom@intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-2-daniel.vetter@ffwll.ch
(cherry picked from commit c238980efd)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Using the I915_MMAP_TYPE_FIXED mmap type requires the TTM backend, so
for that mmap type, use __i915_gem_object_create_user() instead of
i915_gem_object_create_internal(), as we really want to tests objects
mmap-able by user-space.
This also means that the out-of-space error happens at object creation
and returns -ENXIO rather than -ENOSPC, so fix the code up to expect
that on out-of-offset-space errors.
Finally only use I915_MMAP_TYPE_FIXED for LMEM and SMEM for now if
testing on LMEM-capable devices. For stolen LMEM, we still take the
same path as for integrated, as that haven't been moved over to TTM yet,
and user-space should not be able to create out of stolen LMEM anyway.
v2:
- Check the presence of the obj->ops->mmap_offset callback rather than
hardcoding the supported mmap regions in can_mmap() (Maarten Lankhorst)
Fixes: 7961c5b60f ("drm/i915: Add TTM offset argument to mmap.")
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210831122931.157536-1-thomas.hellstrom@linux.intel.com
(cherry picked from commit 450cede7f3)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
The function is only used from within GEM_BUG_ON(), which is causing
warnings with Wunneeded-internal-declaration in some builds. Since the
function is a simple wrapper around a CT function, we can just call the
CT function directly instead.
Fixes: 1fb12c5871 ("drm/i915/guc: skip disabling CTBs before sanitizing the GuC")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210823163137.19770-1-daniele.ceraolospurio@intel.com
(cherry picked from commit 5db1856781)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Users reported that after commit 2bbd6dba84 ("drm/i915: Try to use
fast+narrow link on eDP again and fall back to the old max strategy on
failure"), the screen starts to have wobbly effect.
Commit a5c936add6 ("drm/i915/dp: Use slow and wide link training for
everything") doesn't help either, that means the affected eDP 1.2 panels
only work with max params.
So use max params for panels < eDP 1.4 as Windows does to solve the
issue.
v3:
- Do the eDP rev check in intel_edp_init_dpcd()
v2:
- Check eDP 1.4 instead of DPCD 1.1 to apply max params
Cc: stable@vger.kernel.org
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3714
Fixes: 2bbd6dba84 ("drm/i915: Try to use fast+narrow link on eDP again and fall back to the old max strategy on failure")
Fixes: a5c936add6 ("drm/i915/dp: Use slow and wide link training for everything")
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210820075301.693099-1-kai.heng.feng@canonical.com
(cherry picked from commit d7f213c131)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
After DPRX link training, intel_dp_link_train_phy() did not
return the training result properly. If link training failed,
i915 driver would not run into link train fallback function.
And no hotplug uevent would be received by user space application.
Fixes: b30edfd8d0 ("drm/i915: Switch to LTTPR non-transparent mode link training")
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Cooper Chiou <cooper.chiou@intel.com>
Cc: William Tseng <william.tseng@intel.com>
Signed-off-by: Lee Shawn C <shawn.c.lee@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210706152541.25021-1-shawn.c.lee@intel.com
(cherry picked from commit dab1b47e57)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
If there is no legacy RTC device, don't try to use it for storing trace
data across suspend/resume.
Cc: <stable@vger.kernel.org>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Rafael J. Wysocki <rafael@kernel.org>
Link: https://lore.kernel.org/r/20210903084937.19392-2-jgross@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Today the Xen ballooning is done via delayed work in a workqueue. This
might result in workqueue hangups being reported in case of large
amounts of memory are being ballooned in one go (here 16GB):
BUG: workqueue lockup - pool cpus=6 node=0 flags=0x0 nice=0 stuck for 64s!
Showing busy workqueues and worker pools:
workqueue events: flags=0x0
pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=2/256 refcnt=3
in-flight: 229:balloon_process
pending: cache_reap
workqueue events_freezable_power_: flags=0x84
pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
pending: disk_events_workfn
workqueue mm_percpu_wq: flags=0x8
pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=1/256 refcnt=2
pending: vmstat_update
pool 12: cpus=6 node=0 flags=0x0 nice=0 hung=64s workers=3 idle: 2222 43
This can easily be avoided by using a dedicated kernel thread for doing
the ballooning work.
Reported-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/20210827123206.15429-1-jgross@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>
User space can hold a tty open indefinitely and tty drivers must not
release the underlying structures until the last user is gone.
Switch to using the tty-port reference counter to manage the life time
of the greybus tty state to avoid use after free after a disconnect.
Fixes: a18e151757 ("greybus: more uart work")
Cc: stable@vger.kernel.org # 4.9
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210906124538.22358-1-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Comments after the last #endif of header files don't bring any
information and are redundant with the name of the file. Drop them.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-30-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Each file of the driver contains a short description of its purpose.
These description were a bit outdated.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-26-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The WF200 is the only representative of the WFxxx series and the
development of any successor is not expected. So, for clarity, replace
occurrences of "WFxxx" with "WF200".
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-25-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the early age, it was unexpected to access a VIF that did not exist.
With current code, this happens frequently. Having a trace associated on
this event brings absolutely no informations.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-23-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All the functions of hif_tx_mib.c format data to be sent to the
hardware. In this file, the struct to be sent is always named 'arg'.
Also applies this rule to hif_set_macaddr().
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-21-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The driver sometime use the term 'rate_id' to identify a retry policy
(which is in fact a series of rate IDs). This is misleading.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-16-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The device provide some internal statistic counters. However, the names
of counter were not very meaningful.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-15-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The 'channel' argument of hif_join() should never be NULL. hif_join()
does not have the responsibility to recover bug of caller.
In current code, if the argument channel is NULL, memory leaks. The new
code just emit a warning and does not give the illusion that it is
supported (and indeed a Oops will probably raise a few lines below).
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-13-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The PDS file contains antenna parameters. The file is specific to each
hardware design. Normally, the board designer should add a line in the
of_device_id table with his own antenna parameters.
Until, now the absence of PDS file is a hard fatal error. However,
during the development, in most of the cases, an empty PDS file is
sufficient to start WiFi communication.
This patch keep an error, but allow the user to play with the device.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-12-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The WF200 is able to filter beacons. However, it uses a positive filter:
any change to an IE not listed won't be reported.
In current code, the changes in Channel Switch Announcement (CSA) are
not reported to the host. Thus, it fixes the support for CSA in station
mode.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-11-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Checking if a skb_queue is empty is not an atomic operation. We should
take some precautions to do it.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-8-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the environment is noisy, the device may take time to send scan
requests. Thus, scan requests durations > 5s have already been observed.
During the scan, traffic is neither received, neither sent. From the
user point-of-view, the traffic is frozen for a long time.
This patch reworks the scan processing. It gives to the device a smaller
time budget than previously. However, it does not expect the scan to be
complete and it is able to send another scan request to finish the work.
A big part of the patch aims to avoid an infinite loop if the device
goes crazy.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-6-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the device has finished a scan request, it send a scan complete
("SCAN_COMPL") indication. It is also possible to abort a scan request
with a "SCAN_STOP" message. A SCAN_COMPL is also send in this case.
The driver limits the delay to make a scan request. A timeout happens
almost never but is theoretically possible. Currently, if it happens
the driver does not wait for the SCAN_COMPL. Then, when the driver
starts the next scan request, the device may return -EBUSY (scan
requests often occur back-to-back).
This patch give a chance to the device to send a SCAN_COMPL after a scan
timeout.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-5-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When multiple interface are in use. One is always AP while the other is
always station. When the two interface use the same channel, it makes no
sense to enabled Power Saving (PS) on the station. Indeed, because of
the AP, the device will be kept awake on this channel anyway.
In add, when multiple interface are in use, mac80211 does not update the
PS information and delegate to the driver responsibility to do the
right thing.
Thus, in the current code, when the user enable PS in this
configuration, the driver finally enable PS-Poll which is probably not
what the user expected.
This patch detect this case and applies a sane configuration in all
cases.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-4-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
During the scan requests, the Tx traffic is suspended. This lock is
shared by all the network interfaces. So, a scan request on one
interface will block the traffic on a second interface. This causes
trouble when the queued traffic contains CAB (Content After DTIM Beacon)
since this traffic cannot be delayed.
It could be possible to make the lock local to each interface. But It
would only push the problem further. The device won't be able to send
the CAB before the end of the scan.
So, this patch just ignore the DTIM indication when a scan is in
progress. The firmware will send another indication on the next DTIM and
this time the system will be able to send the traffic just behind the
beacon.
The only drawback of this solution is that the stations connected to
the AP will wait for traffic after the DTIM for nothing. But since the
case is really rare it is not a big deal.
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-3-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The wfx driver checks carefully the coherency of of the DTIM
notifications. We have noticed several times some small inconsistencies
from the firmware on these notification. They have never been critical.
However on the driver side they lead to big fat warnings. Worse, if
these warning are displayed on UART console, they can be long to display
(several hundreds of millisecs). Since, this warning is generated from a
work queue, it can delay all the workqueue users. Especially, it can
drastically slow down the frame management of the driver and then
generate errors that are serious this time (eg. an overflow of the
indication queue of the device).
Signed-off-by: Jérôme Pouiller <jerome.pouiller@silabs.com>
Link: https://lore.kernel.org/r/20210913130203.1903622-2-Jerome.Pouiller@silabs.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>