Merge tag 'misc-habanalabs-next-2020-07-24' of git://people.freedesktop.org/~gabbayo/linux into char-misc-next

Oded writes:

This tag contains the following changes for kernel 5.9-rc1:

- Remove rate limiters from GAUDI configuration (no longer needed).
- Set maximum amount of in-flight CS per ASIC type and increase
  the maximum amount for GAUDI.
- Refactor signal/wait command submissions code
- Calculate trace frequency from PLLs to show accurate profiling data
- Rephrase error messages to make them more clear to the common user
- Add statistics of dropped CS (counter per possible reason for drop)
- Get ECC information from firmware
- Remove support for partial SoC reset in Gaudi
- Halt device CPU only when reset is certain to happen. Sometimes we abort
  the reset procedure and in that case we can't leave device CPU in halt
  mode.
- set each CQ to its own work queue to prevent a race between completions
  on different CQs.
- Use queue pi/ci in order to determine queue occupancy. This is done to
  make the code reusable between current and future ASICs.
- Add more validations for user inputs.
- Refactor PCIe controller configuration to make the code reusable between
  current and future ASICs.
- Update firmware interface headers to latest version
- Move all common code to a dedicated common sub-folder

* tag 'misc-habanalabs-next-2020-07-24' of git://people.freedesktop.org/~gabbayo/linux: (28 commits)
  habanalabs: Fix memory leak in error flow of context initialization
  habanalabs: use no flags on MMU cache invalidation
  habanalabs: enable device before hw_init()
  habanalabs: create internal CB pool
  habanalabs: update hl_boot_if.h from firmware
  habanalabs: create common folder
  habanalabs: check for DMA errors when clearing memory
  habanalabs: verify queue can contain all cs jobs
  habanalabs: Assign each CQ with its own work queue
  habanalabs: halt device CPU only upon certain reset
  habanalabs: remove unused hash
  habanalabs: use queue pi/ci in order to determine queue occupancy
  habanalabs: configure maximum queues per asic
  habanalabs: remove soft-reset support from GAUDI
  habanalabs: PCIe iATU refactoring
  habanalabs: Extract ECC information from FW
  habanalabs: Add dropped cs statistics info struct
  habanalabs: extract cpu boot status lookup
  habanalabs: rephrase error messages
  habanalabs: Increase queues depth
  ...

habanalabs: Fix memory leak in error flow of context initialization

Add a missing free of the cs_pending array in the error flow of context
initialization.

Fixes: c16d45f42b64 ("habanalabs: Use pending CS amount per ASIC")
Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: use no flags on MMU cache invalidation

gaudi_mmu_invalidate_cache() doesn't use the flags parameter, and thus
it can be set to 0 when the function is called in the gaudi only files.

Signed-off-by: Tomer Tayar <ttayar@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: enable device before hw_init()

Device is now enabled before the hw_init() because part of the
initialization requires communication with the device firmware to get
information that is required for the initialization itself

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Tomer Tayar <ttayar@habana.ai>

habanalabs: create internal CB pool

Create a device MMU-mapped internal command buffer pool, in order to allow
the driver to allocate CBs for the signal/wait operations
that are fetched by the queues when they are configured with the user's
address space ID.

We must pre-map this internal pool due to performance issues.

This pool is needed for future ASIC support and it is currently unused in
GOYA and GAUDI.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: update hl_boot_if.h from firmware

Update the boot interface file from the latest version from firmware.
Defines for secure boot were added.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Omer Shpigelman <oshpigelman@habana.ai>

habanalabs: create common folder

For internal needs of our CI we need to move all the common code into a
common folder instead of putting them in the root folder of the driver.

Same applies to the common header files under include/

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Omer Shpigelman <oshpigelman@habana.ai>

habanalabs: check for DMA errors when clearing memory

In GAUDI we use QMAN0 DMA for clearing the MMU memory region
at initialization. if this operation fails it places the DMA in an error
state and then when trying to initialize QMAN0 we fail and erroneously
assume its the QMAN that failed.

This commit adds a check and clear of such DMA errors at initialization so
we will have a better understanding of what went wrong.

Signed-off-by: Moti Haimovski <mhaimovski@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: verify queue can contain all cs jobs

In order for the user to be aware of wrong inputs, we must return
error in case the amount of jobs per cs exceeds the corresponding
queue size.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: Assign each CQ with its own work queue

We identified a possible race during job completion when working
with a single multi-threaded work queue. In order to overcome this
race we suggest using a single threaded work queue per completion
queue, hence we guarantee jobs completion in order.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: halt device CPU only upon certain reset

Currently the driver halts the device CPU in the halt engines function,
which halts all the engines of the ASIC. The problem is that if later on we
stop the reset process (due to inability to clean memory mappings in time),
the CPU will remain in halt mode. This creates many issues, such as
thermal/power control and FLR handling.

Therefore, move the halting of the device CPU to the very end of the reset
process, just before writing to the registers to initiate the reset. In
addition, the driver now needs to send a message to the device F/W to
disable it from sending interrupts to the host machine because during halt
engines function the driver disables the MSI/MSI-X interrupts.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Tomer Tayar <ttayar@habana.ai>

habanalabs: remove unused hash

Remove an old hash that is not in use anymore.

Signed-off-by: Omer Shpigelman <oshpigelman@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: use queue pi/ci in order to determine queue occupancy

Instead of using the free slots amount on the compute CQ to determine
whether we can submit work to queues, use the queues pi/ci.

This is needed in future ASICs where we don't have CQ per queue.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: configure maximum queues per asic

Currently the amount of maximum queues is statically configured.
Using a static value is causing redundunt cycles when traversing
all queues and consumes more memory than actually needed.
In this patch we configure each asic with the exact number of
queues needed.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: remove soft-reset support from GAUDI

Soft-reset isn't supported in GAUDI. Remove the code that performs it and
print error in case the user wants to do it via sysfs.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Tomer Tayar <ttayar@habana.ai>

habanalabs: PCIe iATU refactoring

Divide iATU initialization into inbound/outbound methods.
We must separate it in order to enable different match mode
per PCIe region.
In addition, added support for PCI address match mode.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: Extract ECC information from FW

ECC (Error Correcting Code) interrupts are going to be handled
by the FW. Hence, we define an interface in which the driver can
obtain the relevant ECC information.
This information is needed for monitoring and can also lead
to a hard reset if ECC error is not correctable.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: Add dropped cs statistics info struct

Add command submission statistics structure which can be obtained
through the info ioctl. Each drop counter describes the reason for
which the command submission was dropped.
This information is needed for the user to be aware of the specific
reason for which the submitted work was dropped. The user can then
utilize the driver more efficiently.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: extract cpu boot status lookup

Extract detection of the cpu boot status to a function
to allow code reuse

Signed-off-by: Christine Gharzuzi <cgharzuzi@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: rephrase error messages

rephrase some error/warning/notice messages to make them more accessible to
ordinary users.

There is no need to print context ASID as the driver currently doesn't
support multiple contexts.

Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Reviewed-by: Tomer Tayar <ttayar@habana.ai>

habanalabs: Increase queues depth

After recent concurrent cs amount increase, we must also
increase queues depth since much more concurrent work can be done.
All external queue depths were increased to 4096 as gaudi's
internal queue depths were also increased to 1024.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: rephrase error message

Rephrase F/W error message to make it more understandable to ordinary
users.

Signed-off-by: Omer Shpigelman <oshpigelman@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: calculate trace frequency from PLL

The profiler needs to know the PLL values for correctly showing the
profiling data. Because our firmware can use different PLL configurations,
we need to read the PLL values from the ASIC to pass them to the profiler.

Signed-off-by: Adam Aharon <aaharon@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: align armcp_packet structure to 8 bytes

Once there is a 64-bit field in a structure, GCC compiler for ARM aligns
the structure to 8 bytes. In order to avoid confusion when these
structures are being passed between CPUs from different architectures, we
explicitly align the structure to 8 bytes.

Reviewed-by: Omer Shpigelman <oshpigelman@habana.ai>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

uapi/habanalabs: fix some comments

MAP/UNMAP are done also for device memory.

Reviewed-by: Omer Shpigelman <oshpigelman@habana.ai>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: Use mask instead of shift in sync stream registers

Use proper bitfield masks instead of shifting values when configuring
packets sent to device.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: sync stream generic functionality

Currently sync stream is limited only for external queues. We want to
remove this constraint by adding a new queue property dedicated for sync
stream. In addition we move the initialization and reset methods to the
common code since we can re-use them with slight changes.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: Use pending CS amount per ASIC

Training schemes requires much more concurrent command submissions than
inference does. In addition, training command submissions can be completed
in a non serialized manner. Hence, we add support in which each ASIC will
be able to configure the amount of concurrent pending command submissions,
rather than use a predefined amount. This change will enhance performance
by allowing the user to add more concurrent work without waiting for the
previous work to be completed.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

habanalabs: remove rate limiters from GAUDI

We no longer need to initialize the rate limiters in GAUDI A1.

Reviewed-by: Omer Shpigelman <oshpigelman@habana.ai>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>

Merge tag 'icc-5.9-rc1' of https://git.linaro.org/people/georgi.djakov/linux into char-misc-next

Georgi writes:

interconnect changes for 5.9

Here are the interconnect changes for the 5.9-rc1 merge window
consisting mostly of changes that give the core more flexibility
in order to support some new provider drivers.

Core changes:
- Export of_icc_get_from_provider()
- Relax requirement in of_icc_get_from_provider()
- Allow inter-provider pairs to be configured
- Mark all dummy functions as static inline

Signed-off-by: Georgi Djakov <georgi.djakov@linaro.org>
* tag 'icc-5.9-rc1' of https://git.linaro.org/people/georgi.djakov/linux:
  interconnect: Mark all dummy functions as static inline
  interconnect: Allow inter-provider pairs to be configured
  interconnect: Relax requirement in of_icc_get_from_provider()
  interconnect: Export of_icc_get_from_provider()

dyndbg: export ddebug_exec_queries

Export ddebug_exec_queries() for use by modules.

This will allow module authors to control all their *pr_debug*s
dynamically.  And since ddebug_exec_queries() is what implements
"echo $query >control", it gives the same per-callsite control.

Virtues of this:
- simplicity. just an export.
- full control over any/all subsets of callsites.
- same "query/command-string" in code and console
- full callsite selectivity with module file line format

Format in particular deserves special attention; it is where
low-hanging fruit will be found.

Consider: drivers/gpu/drm/amd/display/include/logger_types.h:

  #define DC_LOG_SURFACE(...)          pr_debug("[SURFACE]:"__VA_ARGS__)
  #define DC_LOG_HW_LINK_TRAINING(...) pr_debug("[HW_LINK_TRAINING]:"__VA_ARGS__)
  .. 9 more ..

Thats 11 string prefixes, used in 804 places in drivers/gpu/**
Clearly this is a systematized classification of those callsites.
And one I'd expect to see repeated often.

Using ddebug_exec_queries(), authors can select on those prefixes
as a unitary set, equivalent to:

    echo "module=MODULE_NAME format=^[SURFACE]: +p" >control

Trivially, those sets can be subsected with the other query terms too,
say file=foo, should the author see fit.

Perhaps as important, users can modify the set of enabled callsites,
presumably to aid debugging by enabling helpful debug callsites, and
disabling those that just clutter the info.

Authors could even alter [fmlt] flags, though I dont see a good reason
why they would.  Perhaps harnessed by bug-logging automation to get
fuller, or more minimal bug-reports.

DRM

drm has both drm.debug, which defines 32 categories of drm_printk
logging, and entirely separate uses of pr_debug, which are dynamic on
this i915 laptop, running mainline.  So I can observe and report on
both.

The i915 driver has 118 dyndbg callsites, with following
"classifications" defined in drivers/gpu/drm/i915/gvt/**

$ grep 915 /proc/dynamic_debug/control | cut -d= -f2 | cut -d: -f1,2 | sort -u
_ "gvt: cmd
_ "gvt: core
_ "gvt: dpy
_ "gvt: el
_ "gvt: irq
_ "gvt: mm
_ "gvt: mmio
_ "gvt: render
_ "gvt: sched
_ "%s for root hub!\012"
_ "Vendor defined info completion code %u\012"

This classification is entirely out-of-band for control by drm.debug,
and is only available to root user at the console.  But module authors
can activate them with ddebug_exec_queries(sprintf("format=^%s +p")),
and then decide how to expose the groups to the user for max utility.

drm.debug

drm.debug has 32 bit-flags, and matching enum drm_debug_category
values to classify the ~2943 DRM_DEBUG*() callsites in drivers/gpu

The drm.debug callback could invoke ddebug_exec_queries() with 32
different hardcoded query strings, needing only (bit) ? " +p" : " -p"
added.

I briefly enabled drm.debug=0xff on my i915 laptop, which yielded
these unique prefixes: (dmesg | cut -c17- | cut -d\] -f1 | sort -u)

[drm:drm_atomic_check_only [drm
[drm:drm_atomic_get_crtc_state [drm
[drm:drm_atomic_get_plane_state [drm
[drm:drm_atomic_nonblocking_commit [drm
[drm:drm_atomic_set_fb_for_plane [drm
[drm:drm_atomic_state_default_clear [drm
[drm:__drm_atomic_state_free [drm
[drm:drm_atomic_state_init [drm
[drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm
[drm:drm_handle_vblank [drm
[drm:drm_ioctl [drm
[drm:drm_mode_addfb2 [drm
[drm:drm_mode_object_get [drm
[drm:drm_mode_object_put.part.0 [drm
[drm:drm_update_vblank_count [drm
[drm:drm_vblank_enable [drm
[drm:drm_vblank_restore [drm
[drm:vblank_disable_fn [drm
i915 0000:00:02.0: [drm:gen9_set_dc_state [i915
i915 0000:00:02.0: [drm:intel_atomic_get_global_obj_state [i915
i915 0000:00:02.0: [drm:__intel_display_power_get_domain.part.0 [i915
i915 0000:00:02.0: [drm:__intel_display_power_put_domain [i915
i915 0000:00:02.0: [drm:intel_plane_atomic_calc_changes [i915
i915 0000:00:02.0: [drm:skl_enable_dc6 [i915

Several good format=^prefixes are apparent there, and some misses.

 ^[drm:drm_atomic_ # misses: [drm:__drm_atomic_state_free [drm
 ^[drm:drm_ioctl
 ^[drm:drm_mode
 ^[drm:drm_vblank_ # misses: [drm:drm_update_vblank_count & [drm:vblank_disable_fn

Its not a perfect 1:1 single format-match per class, but the misses
above can be covered with 1 & 2 additional queries, which can be
concatenated together with ";" separators and submitted with 1 call.

Benefits:

For drm, adapting DRM_DEBUG to use dynamic-debug inside could
replicate (and thereby obsolete) lots of bit-checking in current
DRM_DEBUG callsites, at least with JUMP_LABEL optimized code.
ddebug_exec_queries() and a handful of fixed query-strings can select
and thereby control the already classified callsites.

With the classes mapped to queries, the enum type and parameter can be
eliminated (folded away with macro magic), at least for DYNAMIC_DEBUG
& JUMP_LABEL builds.

Is it safe ?

ddebug_exec_queries() is currently exposed to user space in
several limited ways;

1 it is called from module-load callback, where it implements the
  $modname.dyndbg=+p "fake" parameter provided to all modules.

2 it handles query input via >control directly

IOW, it is "fully" exposed to local root user; exposing the same
functionality to other kernel modules is no additional risk.

The other standard issue to check is locking:

dyndbg has a single mutex, taken by ddebug_change to handle >control,
and by ddebug_proc_(start|stop) to span `cat control`.  Queries
submitted via export will typically have module specified, which
dramatically cuts the scan by ddebug_change vs "module=* +p".
ISTM this proposed export presents no locking problems.

TLDR;

It would be interesting to see how drm.dyndbg=$QUERY and
drm.debug=$HEXY would interact; it might be order dependent, as
if given as modprobe args or in /etc/modprobe.d/

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-19-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: shorten our logging prefix, drop __func__

For log-message output, reduce column space consumed by current
pr_fmt by dropping __func__ and shortening "dynamic_debug" to
"dyndbg".  This improves readability on narrow consoles, and better
matches other kernel boot info messages.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-18-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: allow anchored match on format query term

This should work:

  echo module=amd* format=^[IF_TRACE]: +p  >/proc/dynamic_debug/control

consider drivers/gpu/drm/amd/display/include/logger_types.h:
It has 11 defines like:

  #define DC_LOG_IF_TRACE(...) pr_debug("[IF_TRACE]:"__VA_ARGS__)

These defines are used 804 times at recent count; they are a good use
case to evaluate existing format-message based classifications of
*pr_debug*.  Those macros prefix the supplied format with a fixed
string, I'd expect most existing message classification schemes to do
something similar.

Hence we want to be able to anchor our match to the beginning of the
format string, allowing easy construction of clear and precise
queries, leveraging the existing classification scheme to enable and
disable those callsites.

Note that unlike other search terms, formats are implicitly floating
substring matches, without the need for explicit wildcards.

This makes no attempt at wider regex features, just the one we need.

TLDR: Using the anchor also means the []s are less helpful for
disamiguating the prefix from a random in-message occurrence, allowing
shorter prefixes.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-17-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: combine flags & mask into a struct, simplify with it

flags & mask are used together everywhere, and are passed around
together between multiple functions; they belong together in a struct,
call that struct flag_settings.

Use struct flag_settings to rework 3 functions:
 - ddebug_exec_query - declares query and flag-settings,
         calls other 2, passing flags
 - ddebug_parse_flags - fills flag_settings and returns
 - ddebug_change - test all callsites against query,
       modify passing sites.

benefits:
 - bit-banging always needs flags & mask, best together.
 - simpler function signatures
 - 1 less parameter, less stack overhead

no functional changes

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-16-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: accept query terms like file=bar and module=foo

Current code expects "keyword" "arg" as 2 words, space separated.
Change to also accept "keyword=arg" form as well, and drop !(nwords%2)
requirement.  Then in rest of function, use new keyword, arg variables
instead of word[i], word[i+1]

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-15-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: accept 'file foo.c:func1' and 'file foo.c:10-100'

Accept these additional query forms:

   echo "file $filestr +_" > control

       path/to/file.c:100 # as from control, column 1
       path/to/file.c:1-100 # or any legal line-range
       path/to/file.c:func_A # as from an editor/browser
       path/to/file.c:drm_* # wildcards still work
       path/to/file.c:*_foo # lead wildcard too

1st 2 examples are treated as line-ranges, 3-5 are treated as func's

Doc these changes, and sprinkle in a few extra wild-card examples and
trailing # explanation texts.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-14-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: refactor parse_linerange out of ddebug_parse_query

Make the code-block reusable to later handle "file foo.c:101-200" etc.
This is a 99% code move, with reindent, function wrap&call, +pr_debug.

no functional changes.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-13-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: use gcc ?: to reduce word count

reduce word count via gcc ?: extension, no actual code change.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-12-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: make ddebug_tables list LIFO for add/remove_module

loadable modules are the last in on this list, and are the only
modules that could be removed.  ddebug_remove_module() searches from
head, but ddebug_add_module() uses list_add_tail().  Change it to
list_add() for a micro-optimization.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-11-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: prefer declarative init in caller, to memset in callee

ddebug_exec_query declares an auto var, and passes it to
ddebug_parse_query, which memsets it before using it.  Drop that
memset, instead initialize the variable in the caller; let the
compiler decide how to do it.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-10-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: fix pr_err with empty string

this pr_err attempts to print the string after the OP, but the string
has been parsed and chopped up, so looks empty.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-9-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: fix a BUG_ON in ddebug_describe_flags

ddebug_describe_flags() currently fills a caller provided string buffer,
after testing its size (also passed) in a BUG_ON.  Fix this by
replacing them with a known-big-enough string buffer wrapped in a
struct, and passing that instead.

Also simplify ddebug_describe_flags() flags parameter from a struct to
a member in that struct, and hoist the member deref up to the caller.
This makes the function reusable (soon) where flags are unpacked.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-8-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: fix overcounting of ram used by dyndbg

during dyndbg init, verbose logging prints its ram overhead.  It
counted strlens of struct _ddebug's 4 string members, in all callsite
entries, which would be approximately correct if each had been
mallocd.  But they are pointers into shared .rodata; for example, all
10 kobject callsites have identical filename, module values.

Its best not to count that memory at all, since we cannot know they
were linked in because of CONFIG_DYNAMIC_DEBUG=y, and we want to
report a number that reflects what ram is saved by deconfiguring it.

Also fix wording and size under-reporting of the __dyndbg section.

Heres my overhead, on a virtme-run VM on a fedora-31 laptop:

  dynamic_debug:dynamic_debug_init: 260 modules, 2479 entries \
    and 10400 bytes in ddebug tables, 138824 bytes in __dyndbg section

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-7-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: rename __verbose section to __dyndbg

dyndbg populates its callsite info into __verbose section, change that
to a more specific and descriptive name, __dyndbg.

Also, per checkpatch:
  simplify __attribute(..) to __section(__dyndbg) declaration.

and 1 spelling fix, decriptor

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-6-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: refine debug verbosity; 1 is basic, 2 more chatty

The verbose/debug logging done for `cat $MNT/dynamic_debug/control` is
voluminous (2 per control file entry + 2 per PAGE).  Moreover, it just
prints pointer and sequence, which is not useful to a dyndbg user.
So just drop them.

Also require verbose>=2 for several other debug printks that are a bit
too chatty for typical needs;

ddebug_change() prints changes, once per modified callsite.  Since
queries like "+p" will enable ~2300 callsites in a typical laptop, a
user probably doesn't need to see them often.  ddebug_exec_queries()
still summarizes with verbose=1.

ddebug_(add|remove)_module() also print 1 line per action on a module,
not needed by typical modprobe user.

This leaves verbose=1 better focussed on the >control parsing process.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-5-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg: drop obsolete comment on ddebug_proc_open

commit 4bad78c55002 ("lib/dynamic_debug.c: use seq_open_private() instead of seq_open()")'

The commit was one of a tree-wide set which replaced open-coded
boilerplate with a single tail-call.  It therefore obsoleted the
comment about that boilerplate, clean that up now.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-4-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg-docs: initialization is done early, not arch

since cf964976484 in 2012, initialization is done with early_initcall,
update the Docs, which still say arch_initcall.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-3-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dyndbg-docs: eschew file /full/path query in docs

Regarding:
commit 2b6783191da7 ("dynamic_debug: add trim_prefix() to provide source-root relative paths")
commit a73619a845d5 ("kbuild: use -fmacro-prefix-map to make __FILE__ a relative path")

2nd commit broke dynamic-debug's "file $fullpath" query form, but
nobody noticed because 1st commit had trimmed prefixes from
control-file output, so the click-copy-pasting of fullpaths into new
queries had ceased; that query form became unused.

Removing the function is cleanest, but it could be useful in
old-compiler corner cases, where __FILE__ still has /full/path,
and it safely does nothing otherwize.

So instead, quietly deprecate "file /full/path" query form, by
removing all /full/paths examples in the docs.  I skipped adding a
back-compat note.

Acked-by: <jbaron@akamai.com>
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Link: https://lore.kernel.org/r/20200719231058.1586423-2-jim.cromie@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: hw: don't use one element arrays

Replace the single element arrays with a simple value type u8 reserved,
even thought is is not used for dynamically sized trailing elements
it confuses the effort of replacing one-element arrays with
flexible arrays for that purpose.

Link: https://github.com/KSPP/linux/issues/79
Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-7-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: hw: use sizeof of variable instead of struct type

Use sizeof(*dev) + sizeof(*hw) instead of
sizeof(struct mei_device) + sizeof(struct mei_me_hw)

There is a possibility of bug when variable type has changed but
corresponding struct passed to the sizeof has not.

Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-6-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: client: use sizeof of variable instead of struct type

There is a possibility of bug when variable type has changed but
corresponding struct passed to the sizeof has not.

Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-5-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: bus: use sizeof of variable instead of struct type

There is a possibility of bug when variable type has changed but
corresponding struct passed to the sizeof has not.

Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-4-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: ioctl: use sizeof of variable instead of struct type

Use sizeof(connect_data))) instead of
sizeof(struct mei_connect_client_data) when copying data
between user space and kernel.

There is a possibility of bug when variable type has changed but
corresponding struct passed to the sizeof has not.

Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-3-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: hbm: use sizeof of variable instead of struct type

There is a possibility of bug when variable type has changed but
corresponding struct passed to the sizeof has not.

Reviewed-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200723145927.882743-2-tomas.winkler@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Revert "mei: Avoid the use of one-element arrays"

This reverts commit 3c3b7ddef7879abb2c42422e898145826c79e5f0, as it
turns out Tomas made a better series of patches for this same issue.

Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Cc: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

bus: fsl-mc: probe the allocatable objects first

Because the DPNIs are probed before DPMCPs and other objects that need
to be allocated, messages like "No more resources of type X left" are
printed by the fsl-mc bus driver. This patch resolves the issue by probing
the allocatable objects first and then any other object that may use
them.

Signed-off-by: Grigore Popescu <grigore.popescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Link: https://lore.kernel.org/r/20200717154800.17169-4-ioana.ciornei@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

bus: fsl-mc: use raw spin lock to serialize mc cmds

Replace the spinlock that serializes the MC commands with a raw
spinlock. This is needed for the RT kernel because there are MC
commands sent in interrupt context.

Signed-off-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Link: https://lore.kernel.org/r/20200717154800.17169-3-ioana.ciornei@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

bus: fsl-mc: add missing device types

The MC bus has different types of devices that can be discovered on the
bus. Add the missing device types.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Reviewed-by: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Link: https://lore.kernel.org/r/20200717154800.17169-2-ioana.ciornei@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cardreader/rtsx_pcr.c: use generic power management

Drivers should not use legacy power management as they have to manage power
states and related operations, for the device, themselves. This driver was
handling them with the help of PCI helper functions like
pci_save/restore_state(), pci_enable/disable_device(), etc.

With generic PM, all essentials will be handled by the  PCI core. Driver
needs to do only device-specific operations.

The driver was also using pci_enable_wake(...,..., 0) to disable wake. Use
device_wakeup_disable() instead.

Compile-tested only.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Link: https://lore.kernel.org/r/20200720101722.145211-1-vaibhavgupta40@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: hpilo: avoid a useless memset

Avoid a memset after a call to 'dma_alloc_coherent()'.
This is useless since
commit 518a2f1925c3 ("dma-mapping: zero memory returned from dma_alloc_*")

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/20200718070246.338016-1-christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: hpilo: switch from 'pci_' to 'dma_' API

The wrappers in include/linux/pci-dma-compat.h should go away.

The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.

When memory is allocated in 'ilo_ccb_setup()' GFP_ATOMIC must be used
because a spin_lock is hold in 'ilo_open()' before calling
'ilo_ccb_setup()'

@@
@@
-    PCI_DMA_BIDIRECTIONAL
+    DMA_BIDIRECTIONAL

@@
@@
-    PCI_DMA_TODEVICE
+    DMA_TO_DEVICE

@@
@@
-    PCI_DMA_FROMDEVICE
+    DMA_FROM_DEVICE

@@
@@
-    PCI_DMA_NONE
+    DMA_NONE

@@
expression e1, e2, e3;
@@
-    pci_alloc_consistent(e1, e2, e3)
+    dma_alloc_coherent(&e1->dev, e2, e3, GFP_)

@@
expression e1, e2, e3;
@@
-    pci_zalloc_consistent(e1, e2, e3)
+    dma_alloc_coherent(&e1->dev, e2, e3, GFP_)

@@
expression e1, e2, e3, e4;
@@
-    pci_free_consistent(e1, e2, e3, e4)
+    dma_free_coherent(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_map_single(e1, e2, e3, e4)
+    dma_map_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_single(e1, e2, e3, e4)
+    dma_unmap_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4, e5;
@@
-    pci_map_page(e1, e2, e3, e4, e5)
+    dma_map_page(&e1->dev, e2, e3, e4, e5)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_page(e1, e2, e3, e4)
+    dma_unmap_page(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_map_sg(e1, e2, e3, e4)
+    dma_map_sg(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_sg(e1, e2, e3, e4)
+    dma_unmap_sg(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_single_for_cpu(e1, e2, e3, e4)
+    dma_sync_single_for_cpu(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_single_for_device(e1, e2, e3, e4)
+    dma_sync_single_for_device(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_sg_for_cpu(e1, e2, e3, e4)
+    dma_sync_sg_for_cpu(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_sg_for_device(e1, e2, e3, e4)
+    dma_sync_sg_for_device(&e1->dev, e2, e3, e4)

@@
expression e1, e2;
@@
-    pci_dma_mapping_error(e1, e2)
+    dma_mapping_error(&e1->dev, e2)

@@
expression e1, e2;
@@
-    pci_set_dma_mask(e1, e2)
+    dma_set_mask(&e1->dev, e2)

@@
expression e1, e2;
@@
-    pci_set_consistent_dma_mask(e1, e2)
+    dma_set_coherent_mask(&e1->dev, e2)

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/20200718070224.337964-1-christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'phy-for-5.9' of git://git./linux/kernel/git/phy/linux-phy into char-misc-next

Vinod writes:

phy for 5.9

 - New PHY Drivers:
   - Samsung UFS
   - Qcom USB DWC for ipq806x
   - Xilinx ZynqMP Gigabit Transceiver
   - Qcom USB QMP for IPQ8074
   - BCM63xx USBH

 - Removed:
   - Qcom ufs qmp phy driver

 - Updates:
   - Support for Qcom SM8250 QMP V4 USB3 UNIPHY
   - qcom-snps runtime pm support
   - Cleanup of W=1 warns in the subsystem

* tag 'phy-for-5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy: (46 commits)
  phy: qualcomm: fix setting of tx_deamp_3_5db when device property read fails
  phy: bcm63xx-usbh: Add BCM63xx USBH driver
  dt-bindings: phy: add bcm63xx-usbh bindings
  phy: armada-38x: fix NETA lockup when repeatedly switching speeds
  dt: update Marvell Armada 38x COMPHY binding
  phy: samsung-ufs: Fix IS_ERR argument
  dt-bindings: phy: renesas,usb3-phy: Add r8a774e1 support
  dt-bindings: phy: renesas,usb2-phy: Add r8a774e1 support
  phy: renesas: rcar-gen3-usb2: exit if request_irq() failed
  phy: renesas: rcar-gen3-usb2: move irq registration to init
  devicetree: bindings: phy: Document ipq806x dwc3 qcom phy
  phy: qualcomm: add qcom ipq806x dwc usb phy driver
  phy: samsung-ufs: add UFS PHY driver for samsung SoC
  dt-bindings: phy: Document Samsung UFS PHY bindings
  phy: sun4i-usb: explicitly include gpio/consumer.h
  phy: stm32: use NULL instead of zero
  phy: exynos5-usbdrd: use correct format for structure description
  phy: rockchip-typec: use correct format for structure description
  phy: xgene: remove unsigned integer comparison with less than zero
  phy: mapphone-mdm6600: Add missing description for some structure fields
  ...

char: Replace HTTP links with HTTPS ones

Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.

Deterministic algorithm:
For each file:
  If not .svg:
    For each line:
      If doesn't contain `\bxmlns\b`:
        For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
  If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
            If both the HTTP and HTTPS versions
            return 200 OK and serve the same content:
              Replace HTTP with HTTPS.

Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Link: https://lore.kernel.org/r/20200713104453.33414-1-grandmaster@al2klimov.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: Replace HTTP links with HTTPS ones

Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.

Deterministic algorithm:
For each file:
  If not .svg:
    For each line:
      If doesn't contain `\bxmlns\b`:
        For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
  If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
            If both the HTTP and HTTPS versions
            return 200 OK and serve the same content:
              Replace HTTP with HTTPS.

Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Link: https://lore.kernel.org/r/20200713164024.35988-1-grandmaster@al2klimov.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

android: binder.h: drop a duplicated word

Drop the repeated word "the" in a comment.

Cc: Arve Hjønnevåg <arve@android.com>
Cc: Todd Kjos <tkjos@android.com>
Cc: Martijn Coenen <maco@android.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Hridya Valsaraju <hridya@google.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: devel@driverdev.osuosl.org
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Link: https://lore.kernel.org/r/20200719002738.20210-1-rdunlap@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: mic: <linux/mic_bus.h>: drop a duplicated word

Drop the repeated word "the" in a comment.

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Sudeep Dutt <sudeep.dutt@intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit@intel.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20200719002943.20624-1-rdunlap@infradead.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: Avoid the use of one-element arrays

One-element arrays are being deprecated[1]. Replace the one-element
arrays with a simple value type u8 reserved, once this is just a
placeholder for alignment.

Also, while there, use the preferred form for passing a size of a struct.
The alternative form where struct name is spelled out hurts readability
and introduces an opportunity for a bug when the variable type is changed
but the corresponding sizeof that is passed as argument is not.

[1] https://github.com/KSPP/linux/issues/79

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20200714214516.GA1040@embeddedor
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: Replace HTTP links with HTTPS ones

Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.

Deterministic algorithm:
For each file:
  If not .svg:
    For each line:
      If doesn't contain `\bxmlns\b`:
        For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
  If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
            If both the HTTP and HTTPS versions
            return 200 OK and serve the same content:
              Replace HTTP with HTTPS.

Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Acked-by: Tomas Winkler <tomas.winkler@intel.com>
Link: https://lore.kernel.org/r/20200717185925.84102-1-grandmaster@al2klimov.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

mei: hdcp: Replace one-element array with flexible-array member

There is a regular need in the kernel to provide a way to declare
having a dynamically sized set of trailing elements in a structure.
Kernel code should always use “flexible array members”[1] for these
cases. The older style of one-element or zero-length arrays should
no longer be used[2].

Also, make use of the array_size() helper instead of the open-coded
version in memcpy(). These sorts of multiplication factors need to
be wrapped in array_size().

And while there, use the preferred form for passing a size of a struct.
The alternative form where struct name is spelled out hurts readability
and introduces an opportunity for a bug when the pointer variable type is
changed but the corresponding sizeof that is passed as argument is not.

[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://github.com/KSPP/linux/issues/79

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20200722181534.GA31357@embeddedor
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'fpga-for-5.9' of git://git./linux/kernel/git/mdf/linux-fpga into char-misc-next

Moritz writes:

FPGA Manager changes for 5.9-rc1

Here is the (slightly larger than usual) patch set for the 5.9-rc1 merge
window.

DFL:
- Xu's changes add support for AFU interrupt handling and puts them to
  use for error handling.
- Xu's other change also adds another device-id for the Intel FPGA PAC N3000.
- John's change converts from using get_user_pages() to
  pin_user_pages().
- Gustavo's patch cleans up some of the allocation by using
  struct_size().

Xilinx:
- Luca's changes clean up the xilinx-spi and xilinx-slave-serial drivers
  and updates the comments and dt-bindings to reflect the fact it also
  supports 7 series devices.

Core:
- Tom cleaned up the fpga-bridge / fpga-mgr core by removing some
  dead-stores.

All patches have been reviewed on the mailing list, and have been in the
last few linux-next releases (as part of my for-next branch) without issues.

Signed-off-by: Moritz Fischer <mdf@kernel.org>
* tag 'fpga-for-5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/mdf/linux-fpga:
  fpga: dfl: pci: add device id for Intel FPGA PAC N3000
  Documentation: fpga: dfl: add descriptions for interrupt related interfaces.
  fpga: dfl: afu: add AFU interrupt support
  fpga: dfl: fme: add interrupt support for global error reporting
  fpga: dfl: afu: add interrupt support for port error reporting
  fpga: dfl: introduce interrupt trigger setting API
  fpga: dfl: pci: add irq info for feature devices enumeration
  fpga: dfl: parse interrupt info for feature devices on enumeration
  fpga manager: xilinx-spi: check INIT_B pin during write_init
  dt-bindings: fpga: xilinx-slave-serial: add optional INIT_B GPIO
  fpga: Fix dead store in fpga-bridge.c
  fpga: Fix dead store fpga-mgr.c
  fpga: dfl: Use struct_size() in kzalloc()
  fpga manager: xilinx-spi: remove unneeded, mistyped variables
  fpga manager: xilinx-spi: valid for the 7 Series too
  dt-bindings: fpga: xilinx-slave-serial: valid for the 7 Series too
  fpga: dfl: afu: convert get_user_pages() --> pin_user_pages()

Merge tag 'soundwire-5.9-rc1' of git://git./linux/kernel/git/vkoul/soundwire into char-misc-next

Vinod writes:

soundwire updates for 5.9-rc1

This contains few core changes and bunch of Intel driver updates:

 - Adds definitions for 1.2 spec
 - Sanyog left as a MAINTAINER and Bard took his place while Sanyog
   is a reviewer now.
 - Intel: Lots of updates to stream/dai handling, wake support and link
   synchronization.

* tag 'soundwire-5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire: (31 commits)
  Soundwire: intel_init: save Slave(s) _ADR info in sdw_intel_ctx
  soundwire: intel: add wake interrupt support
  soundwire: intel/cadence: merge Soundwire interrupt handlers/threads
  soundwire: intel_init: use EXPORT_SYMBOL_NS
  soundwire: intel_init: add implementation of sdw_intel_enable_irq()
  soundwire: intel: introduce helper for link synchronization
  soundwire: intel: introduce a helper to arm link synchronization
  soundwire: intel: revisit SHIM programming sequences.
  soundwire: intel: reuse code for wait loops to set/clear bits
  soundwire: fix the kernel-doc comment
  soundwire: sdw.h: fix indentation
  soundwire: sdw.h: fix PRBS/Static_1 swapped definitions
  soundwire: intel: don't free dma_data in DAI shutdown
  soundwire: cadence: allocate/free dma_data in set_sdw_stream
  soundwire: intel: remove stream allocation/free
  soundwire: stream: add helper to startup/shutdown streams
  soundwire: intel: implement get_sdw_stream() operations
  MAINTAINERS: change SoundWire maintainer
  soundwire: bus: initialize bus clock base and scale registers
  soundwire: extend SDW_SLAVE_ENTRY
  ...

misc: rtsx: Use standard PCI definitions

When reading registers defined by the PCIe spec, use the names already
defined by the PCI core.  This makes maintenance of the PCI core and
drivers easier.  No functional change intended.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200721212336.1159079-6-helgaas@kernel.org
[ additional replacements due to changes in my tree - gregkh ]
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: rtsx: Find L1 PM Substates capability instead of hard-coding

Instead of hard-coding the location of the L1 PM Substates capability based
on the Device ID, search for it in the extended capabilities list.  This
works for any device, as long as it implements the L1 PM Substates
capability correctly, so it doesn't require maintenance as new devices are
added.  No functional change intended.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200721212336.1159079-5-helgaas@kernel.org
[ minor addition due to differences in my tree - gregkh]
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: rtsx: Remove rtsx_pci_read/write_config() wrappers

rtsx_pci_read_config_dword() and similar wrappers around the PCI config
accessors add very little value, and they obscure the fact that often we
are accessing standard PCI registers that should be coordinated with the
PCI core.

Remove the wrappers and use the PCI config accessors directly.  No
functional change intended.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200721212336.1159079-4-helgaas@kernel.org
[ fixed up some other instances as original patch was based on old tree - gregkh
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: rtsx: Remove unused pcie_cap

There are no more uses of struct rtsx_pcr.pcie_cap.  Remove it.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200721212336.1159079-3-helgaas@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

misc: rtsx: Use pcie_capability_clear_and_set_word() for PCI_EXP_LNKCTL

Instead of using the driver-specific rtsx_pci_write_config_byte() to update
the PCIe Link Control Register, use pcie_capability_write_word() like the
rest of the kernel does.  This makes it easier to maintain ASPM across the
PCI core and drivers.

No functional change intended.  I missed this when doing 3d1e7aa80d1c
("misc: rtsx: Use pcie_capability_clear_and_set_word() for
PCI_EXP_LNKCTL").

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20200721212336.1159079-2-helgaas@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

phy: qualcomm: fix setting of tx_deamp_3_5db when device property read fails

Currently when reading of the device property for "qcom,tx-deamp_3_5db"
fails the default is being assigned incorrectly to phy_dwc3->rx_eq. This
looks like a copy-n-paste error and in fact should be assigning the
default instead to phy_dwc3->tx_deamp_3_5db

Addresses-Coverity: ("Copy-paste error")
Fixes: ef19b117b834 ("phy: qualcomm: add qcom ipq806x dwc usb phy driver")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20200721150613.416876-1-colin.king@canonical.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>

phy: bcm63xx-usbh: Add BCM63xx USBH driver

Add BCM63xx USBH PHY driver for BMIPS.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Signed-off-by: Simon Arlott <simon@octiron.net>
Link: https://lore.kernel.org/r/20200720131209.1236590-3-noltari@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dt-bindings: phy: add bcm63xx-usbh bindings

Document BCM63xx USBH PHY bindings.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20200720131209.1236590-2-noltari@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>

phy: armada-38x: fix NETA lockup when repeatedly switching speeds

The mvneta hardware appears to lock up in various random ways when
repeatedly switching speeds between 1G and 2.5G, which involves
reprogramming the COMPHY.  It is not entirely clear why this happens,
but best guess is that reprogramming the COMPHY glitches mvneta clocks
causing the hardware to fail.  It seems that rebooting resolves the
failure, but not down/up cycling the interface alone.

Various other approaches have been tried, such as trying to cleanly
power down the COMPHY and then take it back through the power up
initialisation, but this does not seem to help.

It was finally noticed that u-boot's last step when configuring a
COMPHY for "SGMII" mode was to poke at a register described as
"GBE_CONFIGURATION_REG", which is undocumented in any external
documentation.  All that we have is the fact that u-boot sets a bit
corresponding to the "SGMII" lane at the end of COMPHY initialisation.

Experimentation shows that if we clear this bit prior to changing the
speed, and then set it afterwards, mvneta does not suffer this problem
on the SolidRun Clearfog when switching speeds between 1G and 2.5G.

This problem was found while script-testing phylink.

This fix also requires the corresponding change to DT to be effective.
See "ARM: dts: armada-38x: fix NETA lockup when repeatedly switching
speeds".

Fixes: 14dc100b4411 ("phy: armada38x: add common phy support")
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/E1jxtRj-0003Tz-CG@rmk-PC.armlinux.org.uk
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dt: update Marvell Armada 38x COMPHY binding

Update the Marvell Armada 38x COMPHY binding with an additional
optional register pair describing the location of an undocumented
system register controlling something to do with the Gigabit Ethernet
and COMPHY.  There is one bit for each COMPHY lane that may be using
the serdes, but exactly what this register does is completely unknown.

This register only appears to exist on Armada 38x devices, and not
other SoCs using the NETA ethernet block, so it seems logical that it
should be part of the COMPHY.

This is also how u-boot groups this register; it is dealt with as part
of the COMPHY initialisation there.

However, at the end of the day, due to the undocumented nature of this
register, we can only guess.

Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/E1jxtRZ-0003Ta-4h@rmk-PC.armlinux.org.uk
Signed-off-by: Vinod Koul <vkoul@kernel.org>

coresight: etm4x: Fix save/restore during cpu idle

The ETM state save/restore incorrectly reads/writes some of the 64bit
registers (e.g, address comparators, vmid/cid comparators etc.) using
32bit accesses. Ensure we use the appropriate width accessors for
the registers.

Fixes: f188b5e76aae ("coresight: etm4x: Save/restore state across CPU low power states")
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Cc: Mike Leach <mike.leach@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-18-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: etm: perf: Add default sink selection to etm perf

Add default sink selection to the perf trace handling in the etm driver.
Uses the select default sink infrastructure to select a sink for the perf
session, if no other sink is specified.

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-17-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: tmc: Update sink types for default selection

An additional sink subtype is added to differentiate ETB/ETF buffer
sinks and ETR type system memory sinks.

This allows the prioritised selection of default sinks.

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-16-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: Add default sink selection to CoreSight base

Adds a method to select a suitable sink connected to a given source.

In cases where no sink is defined, the coresight_find_default_sink
routine can search from a given source, through the child connections
until a suitable sink is found.

The suitability is defined in by the sink coresight_dev_subtype on the
CoreSight device, and the distance from the source by counting
connections.

Higher value subtype is preferred - where these are equal, shorter
distance from source is used as a tie-break.

This allows for default sink to be discovered were none is specified
(e.g. perf command line)

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-15-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: tmc: Fix TMC mode read in tmc_read_unprepare_etb()

Reading TMC mode register without proper coresight power
management can lead to exceptions like the one in the call
trace below in tmc_read_unprepare_etb() when the trace data
is read after the sink is disabled. So fix this by having
a check for coresight sysfs mode before reading TMC mode
management register in tmc_read_unprepare_etb() similar to
tmc_read_prepare_etb().

  SError Interrupt on CPU6, code 0xbe000411 -- SError
  pstate: 80400089 (Nzcv daIf +PAN -UAO)
  pc : tmc_read_unprepare_etb+0x74/0x108
  lr : tmc_read_unprepare_etb+0x54/0x108
  sp : ffffff80d9507c30
  x29: ffffff80d9507c30 x28: ffffff80b3569a0c
  x27: 0000000000000000 x26: 00000000000a0001
  x25: ffffff80cbae9550 x24: 0000000000000010
  x23: ffffffd07296b0f0 x22: ffffffd0109ee028
  x21: 0000000000000000 x20: ffffff80d19e70e0
  x19: ffffff80d19e7080 x18: 0000000000000000
  x17: 0000000000000000 x16: 0000000000000000
  x15: 0000000000000000 x14: 0000000000000000
  x13: 0000000000000000 x12: 0000000000000000
  x11: 0000000000000000 x10: dfffffd000000001
  x9 : 0000000000000000 x8 : 0000000000000002
  x7 : ffffffd071d0fe78 x6 : 0000000000000000
  x5 : 0000000000000080 x4 : 0000000000000001
  x3 : ffffffd071d0fe98 x2 : 0000000000000000
  x1 : 0000000000000004 x0 : 0000000000000001
  Kernel panic - not syncing: Asynchronous SError Interrupt

Fixes: 4525412a5046 ("coresight: tmc: making prepare/unprepare functions generic")
Reported-by: Mike Leach <mike.leach@linaro.org>
Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Tested-by: Mike Leach <mike.leach@linaro.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-14-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: tmc: Add shutdown callback for TMC ETR

Implement a shutdown callback to ensure ETR hardware is
properly shutdown in reboot/shutdown path. This is required
for ETR which has SMMU address translation enabled like on
SC7180 SoC and few others. If the hardware is still accessing
memory after SMMU translation is disabled as part of SMMU
shutdown callback in system reboot or shutdown path, then
IOVAs(I/O virtual address) which it was using will go on the
bus as the physical addresses which might result in unknown
crashes (NoC/interconnect errors). So we make sure from this
shutdown callback that the ETR is shutdown before SMMU translation
is disabled and device_link in SMMU driver will take care of
ordering of shutdown callbacks such that SMMU shutdown callback
is not called before any of its consumer shutdown callbacks.

Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-13-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: Fix comment in main header file

Comment for an elemnt in the coresight_device structure appears to have
been corrupted and makes no sense. Fix this before making further changes.

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-12-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: etmv4: Counter values not saved on disable

The counter value registers change during operation, however this change
is not reflected in the values seen by the user in sysfs.

This fixes the issue by reading back the values on disable.

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Fixes: 2e1cdfe184b52 ("coresight-etm4x: Adding CoreSight ETM4x driver")
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-11-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: etmv4: Fix resource selector constant

ETMv4 max resource selector constant incorrectly set to 16. Updated to the
correct 32 value, and adjustments made to limited code using it.

Signed-off-by: Mike Leach <mike.leach@linaro.org>
Fixes: 2e1cdfe184b52 ("coresight-etm4x: Adding CoreSight ETM4x driver")
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-10-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: Drop double check for ACPI companion device

acpi_dev_get_resources() does perform the NULL pointer check against
ACPI companion device which is given as function parameter. Thus,
there is no need to duplicate this check in the caller.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-9-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: Use devm_kcalloc() in coresight_alloc_conns()

A multiplication for the size determination of a memory allocation
indicated that an array data structure should be processed.
Thus use the corresponding function "devm_kcalloc".

Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-8-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dt-bindings: arm: coresight: Add optional property to replicators

Add an optional boolean property "qcom,replicator-loses-context" to
identify replicators which loses context when AMBA clocks are removed
in certain configurable replicator designs.

Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-7-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: replicator: Reset replicator if context is lost

On some QCOM SoCs, replicators in Always-On domain loses its
context as soon as the clock is disabled. Currently as a part
of pm_runtime workqueue, clock is disabled after the replicator
is initialized by amba_pm_runtime_suspend assuming that context
is not lost which is not true for replicators with such
limitations. So add a new property "qcom,replicator-loses-context"
to identify such replicators and reset them.

Suggested-by: Mike Leach <mike.leach@linaro.org>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-6-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dt-bindings: arm: coresight: Add support to skip trace unit power up

Add "qcom,skip-power-up" property to identify systems which can
skip powering up of trace unit since they share the same power
domain as their CPU core. This is required to identify such
systems with hardware errata which stops the CPU watchdog counter
when the power up bit is set (TRCPDCR.PU).

Signed-off-by: Tingwei Zhang <tingwei@codeaurora.org>
Co-developed-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-5-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: etm4x: Add support to skip trace unit power up

On some Qualcomm Technologies Inc. SoCs like SC7180, there
exists a hardware errata where the APSS (Application Processor
SubSystem)/CPU watchdog counter is stopped when the trace unit
power up ETM register is set (TRCPDCR.PU = 1). Since the ETMs
share the same power domain as that of respective CPU cores,
they are powered on when the CPU core is powered on. So we can
skip powering up of trace unit after checking for this errata
via new property called "qcom,skip-power-up".

Signed-off-by: Tingwei Zhang <tingwei@codeaurora.org>
Co-developed-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-4-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: catu: Use CS_AMBA_ID macro for id table

Use CS_AMBA_ID macro for coresight catu AMBA id table
instead of open coding.

Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-3-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

coresight: replicator: Use CS_AMBA_ID macro for id table

Use CS_AMBA_ID macro for dynamic replicator AMBA id table
instead of open coding.

Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Link: https://lore.kernel.org/r/20200716175746.3338735-2-mathieu.poirier@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

phy: samsung-ufs: Fix IS_ERR argument

Fix IS_ERR argument in samsung_ufs_phy_symbol_clk_init(). The proper
argument to be passed to IS_ERR() is phy->rx1_symbol_clk.

This bug was detected with the help of Coccinelle.

Fixes: bca21e930451 ("phy: samsung-ufs: add UFS PHY driver for samsung SoC")
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Link: https://lore.kernel.org/r/20200720132718.GA13413@embeddedor
Signed-off-by: Vinod Koul <vkoul@kernel.org>

Soundwire: intel_init: save Slave(s) _ADR info in sdw_intel_ctx

Save ACPI information in context so that we can match machine driver
with sdw _ADR matching tables.

Suggested-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200716150947.22119-10-yung-chuan.liao@linux.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>