Similar to msm8916, msm8939 has (at least) 6 "General Purpose" clocks that
can be muxed to SoC pins. These clocks are:
GP_CLK{0, 1} : GPIO_{31, 32} (Belongs to CAMSS according to Linux)
GP_CLK_{1-3}{A, B} : GPIO_{49-51, 97, 12, 13} (Belongs to GCC itself)
GP_MN : GPIO_110 (Doesn't seem to be described in gcc,
ignored in this patch)
Those clocks may be used as e.g. PWM sources for external peripherals.
Add more frequencies to the table for those clocks so it's possible
for arbitrary peripherals to make use of them.
Reference: https://lore.kernel.org/r/20220612145955.385787-5-nikita@trvn.ru
Signed-off-by: Lin, Meng-Bo <linmengbo0689@protonmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221117171343.24216-1-linmengbo0689@protonmail.com
Modernize the krait-cc driver to parent-data API and refactor to drop
any use of parent_names. From Documentation all the required clocks should
be declared in DTS so fw_name can be correctly used to get the parents
for all the muxes. .name is also declared to save compatibility with old
DT.
While at it also drop some hardcoded index and introduce an enum to make
index values more clear.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221109005631.3189-5-ansuelsmth@gmail.com
Some bootloader may leave the system in an even more undefined state
with the secondary mux of L2 or other cores sourcing out of the acpu_aux
parent. This results in the clk set to the PXO rate or a PLL8 rate.
The current logic to reset the mux and set them to a defined state only
handle if the mux are configured to source out of QSB. Change this and
force a new and defined state if the current clk is lower than the aux
rate. This way we can handle any wrong configuration where the mux is
sourcing out of QSB (rate 225MHz, currently set to a virtual rate of 1),
PXO rate (rate 25MHz) or PLL8 (needs to be configured to run at 384Mhz).
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221109005631.3189-3-ansuelsmth@gmail.com
The secondary mux parent order is swapped.
This currently doesn't cause problems as the secondary mux is used for idle
clk and as a safe clk source while reprogramming the hfpll.
Each mux have 2 or more output but he always have a safe source to
switch while reprogramming the connected pll. We use a clk notifier to
switch to the correct parent before clk core can apply the correct rate.
The parent to switch is hardcoded in the mux struct.
For the secondary mux the safe source to use is the qsb parent as it's
the only fixed clk as the acpus_aux is a pll that can source from pxo or
from pll8.
The hardcoded safe parent for the secondary mux is set to index 0 that
in the secondary mux map is set to 2.
But the index 0 is actually acpu_aux in the parent list.
Fix the swapped parents to correctly handle idle frequency and output a
sane clk_summary report.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221109005631.3189-1-ansuelsmth@gmail.com
Currently div2 value is applied to the wrong bits. This is caused by a
bug in the code where the shift is done only for lpl, for anything
else the mask is not shifted to the correct bits.
Fix this by correctly shift if lpl is not supported.
Fixes: 4d7dc77bab ("clk: qcom: Add support for Krait clocks")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221108215625.30186-1-ansuelsmth@gmail.com
Downstream QCA 5.4 kernel defines networking resets which are not present
in the mainline kernel but are required for the networking drivers.
So, port the downstream resets and avoid using magic values for mask,
construct mask for resets which require multiple bits to be set/cleared.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221107132901.489240-3-robimarko@gmail.com
This patch adds the support for giving the complete bitmask
in reset structure and reset operation will use this bitmask
for all reset operations.
Currently, reset structure only takes a single bit for each reset
and then calculates the bitmask by using the BIT() macro.
However, this is not sufficient anymore for newer SoC-s like IPQ8074,
IPQ6018 and more, since their networking resets require multiple bits
to be asserted in order to properly reset the HW block completely.
So, in order to allow asserting multiple bits add "bitmask" field to
qcom_reset_map, and then use that bitmask value if its populated in the
driver, if its not populated, then we just default to existing behaviour
and calculate the bitmask on the fly.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221107132901.489240-1-robimarko@gmail.com
The sc7180 lpass clock controller's pm_runtime usage wasn't broken
quite as spectacularly as the sc7280's pm_runtime usage, but it was
still broken. Putting some printouts in at boot showed me this (with
serial console enabled, which makes the prints slow and thus changes
timing):
[ 3.109951] DOUG: my_pm_clk_resume, usage=1
[ 3.114767] DOUG: my_pm_clk_resume, usage=1
[ 3.664443] DOUG: my_pm_clk_suspend, usage=0
[ 3.897566] DOUG: my_pm_clk_suspend, usage=0
[ 3.910137] DOUG: my_pm_clk_resume, usage=1
[ 3.923217] DOUG: my_pm_clk_resume, usage=0
[ 4.440116] DOUG: my_pm_clk_suspend, usage=-1
[ 4.444982] DOUG: my_pm_clk_suspend, usage=0
[ 14.170501] DOUG: my_pm_clk_resume, usage=1
[ 14.176245] DOUG: my_pm_clk_resume, usage=0
...or this w/out serial console:
[ 0.556139] DOUG: my_pm_clk_resume, usage=1
[ 0.556279] DOUG: my_pm_clk_resume, usage=1
[ 1.058422] DOUG: my_pm_clk_suspend, usage=-1
[ 1.058464] DOUG: my_pm_clk_suspend, usage=0
[ 1.186250] DOUG: my_pm_clk_resume, usage=1
[ 1.186292] DOUG: my_pm_clk_resume, usage=0
[ 1.731536] DOUG: my_pm_clk_suspend, usage=-1
[ 1.731557] DOUG: my_pm_clk_suspend, usage=0
[ 10.288910] DOUG: my_pm_clk_resume, usage=1
[ 10.289496] DOUG: my_pm_clk_resume, usage=0
It seems to be doing roughly the right sequence of calls, but just
like with sc7280 this is more by luck than anything. Having a usage of
-1 is just not OK.
Let's fix this like we did with sc7280.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Fixes: ce8c195e65 ("clk: qcom: lpasscc: Introduce pm autosuspend for SC7180")
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221104064055.2.I49b25b9bda9430fc7ea21e5a708ca5a0aced2798@changeid
The pm_runtime usage in lpass-sc7280 was broken in quite a few
ways. Specifically:
1. At the end of probe it called "put" twice. This is a no-no and will
end us up with a negative usage count. Even worse than calling
"put" twice, it never called "get" once. Thus after bootup it could
be seen that the runtime usage of the devices managed by this
driver was -2.
2. In some error cases it manually called pm_runtime_disable() even
though it had previously used devm_add_action_or_reset() to set
this up to be called automatically. This meant that in these error
cases we'd double-call pm_runtime_disable().
3. It forgot to call undo pm_runtime_use_autosuspend(), which can
sometimes have subtle problems (and the docs specifically mention
that you need to undo this function).
Overall the above seriously calls into question how this driver is
working. It seems like a combination of "it doesn't", "by luck", and
"because of the weirdness of runtime_pm". Specifically I put a
printout to the serial console every time the runtime suspend/resume
was called for the two devices created by this driver (I wrapped the
pm_clk calls). When I had serial console enabled, I found that the
calls got resumed at bootup (when the clk core probed and before our
double-put) and then never touched again. That's no good.
[ 0.829997] DOUG: my_pm_clk_resume, usage=1
[ 0.835487] DOUG: my_pm_clk_resume, usage=1
When I disabled serial console (speeding up boot), I got a different
pattern, which I guess (?) is better:
[ 0.089767] DOUG: my_pm_clk_resume, usage=1
[ 0.090507] DOUG: my_pm_clk_resume, usage=1
[ 0.151885] DOUG: my_pm_clk_suspend, usage=-2
[ 0.151914] DOUG: my_pm_clk_suspend, usage=-2
[ 1.825747] DOUG: my_pm_clk_resume, usage=-1
[ 1.825774] DOUG: my_pm_clk_resume, usage=-1
[ 1.888269] DOUG: my_pm_clk_suspend, usage=-2
[ 1.888282] DOUG: my_pm_clk_suspend, usage=-2
These different patterns have to do with the fact that the core PM
Runtime code really isn't designed to be robust to negative usage
counts and sometimes may happen to stumble upon a behavior that
happens to "work". For instance, you can see that
__pm_runtime_suspend() will treat any non-zero value (including
negative numbers) as if the device is in use.
In any case, let's fix the driver to be correct. We'll hold a
pm_runtime reference for the whole probe and then drop it (once!) at
the end. We'll get rid of manual pm_runtime_disable() calls in the
error handling. We'll also switch to devm_pm_runtime_enable(), which
magically handles undoing pm_runtime_use_autosuspend() as of commit
b4060db925 ("PM: runtime: Have devm_pm_runtime_enable() handle
pm_runtime_dont_use_autosuspend()").
While we're at this, let's also use devm_pm_clk_create() instead of
rolling it ourselves.
Note that the above changes make it obvious that
lpassaudio_create_pm_clks() was doing more than just creating
clocks. It was also setting up pm_runtime parameters. Let's rename it.
All of these problems were found by code inspection. I started looking
at this driver because it was involved in a deadlock that I reported a
while ago [1]. Though I bisected the deadlock to commit 1b771839de
("clk: qcom: gdsc: enable optional power domain support"), it was
never really clear why that patch affected it other than a luck of
timing changes. I'll also note that by fixing the timing (as done in
this change) we also seem to aboid the deadlock, which is a nice
benefit.
Also note that some of the fixes here are much the same type of stuff
that Dmitry did in commit 72cfc73f46 ("clk: qcom: use
devm_pm_runtime_enable and devm_pm_clk_create"), but I guess
lpassaudiocc-sc7280.c didn't exist then.
[1] https://lore.kernel.org/r/20220922154354.2486595-1-dianders@chromium.org
Fixes: a9dd26639d ("clk: qcom: lpass: Add support for LPASS clock controller for SC7280")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221104064055.1.I00a0e4564a25489e85328ec41636497775627564@changeid
Pull more random number generator updates from Jason Donenfeld:
"This time with some large scale treewide cleanups.
The intent of this pull is to clean up the way callers fetch random
integers. The current rules for doing this right are:
- If you want a secure or an insecure random u64, use get_random_u64()
- If you want a secure or an insecure random u32, use get_random_u32()
The old function prandom_u32() has been deprecated for a while
now and is just a wrapper around get_random_u32(). Same for
get_random_int().
- If you want a secure or an insecure random u16, use get_random_u16()
- If you want a secure or an insecure random u8, use get_random_u8()
- If you want secure or insecure random bytes, use get_random_bytes().
The old function prandom_bytes() has been deprecated for a while
now and has long been a wrapper around get_random_bytes()
- If you want a non-uniform random u32, u16, or u8 bounded by a
certain open interval maximum, use prandom_u32_max()
I say "non-uniform", because it doesn't do any rejection sampling
or divisions. Hence, it stays within the prandom_*() namespace, not
the get_random_*() namespace.
I'm currently investigating a "uniform" function for 6.2. We'll see
what comes of that.
By applying these rules uniformly, we get several benefits:
- By using prandom_u32_max() with an upper-bound that the compiler
can prove at compile-time is ≤65536 or ≤256, internally
get_random_u16() or get_random_u8() is used, which wastes fewer
batched random bytes, and hence has higher throughput.
- By using prandom_u32_max() instead of %, when the upper-bound is
not a constant, division is still avoided, because
prandom_u32_max() uses a faster multiplication-based trick instead.
- By using get_random_u16() or get_random_u8() in cases where the
return value is intended to indeed be a u16 or a u8, we waste fewer
batched random bytes, and hence have higher throughput.
This series was originally done by hand while I was on an airplane
without Internet. Later, Kees and I worked on retroactively figuring
out what could be done with Coccinelle and what had to be done
manually, and then we split things up based on that.
So while this touches a lot of files, the actual amount of code that's
hand fiddled is comfortably small"
* tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
prandom: remove unused functions
treewide: use get_random_bytes() when possible
treewide: use get_random_u32() when possible
treewide: use get_random_{u8,u16}() when possible, part 2
treewide: use get_random_{u8,u16}() when possible, part 1
treewide: use prandom_u32_max() when possible, part 2
treewide: use prandom_u32_max() when possible, part 1
Pull more perf tools updates from Arnaldo Carvalho de Melo:
- Use BPF CO-RE (Compile Once, Run Everywhere) to support old kernels
when using bperf (perf BPF based counters) with cgroups.
- Support HiSilicon PCIe Performance Monitoring Unit (PMU), that
monitors bandwidth, latency, bus utilization and buffer occupancy.
Documented in Documentation/admin-guide/perf/hisi-pcie-pmu.rst.
- User space tasks can migrate between CPUs, so when tracing selected
CPUs, system-wide sideband is still needed, fix it in the setup of
Intel PT on hybrid systems.
- Fix metricgroups title message in 'perf list', it should state that
the metrics groups are to be used with the '-M' option, not '-e'.
- Sync the msr-index.h copy with the kernel sources, adding support for
using "AMD64_TSC_RATIO" in filter expressions in 'perf trace' as well
as decoding it when printing the MSR tracepoint arguments.
- Fix program header size and alignment when generating a JIT ELF in
'perf inject'.
- Add multiple new Intel PT 'perf test' entries, including a jitdump
one.
- Fix the 'perf test' entries for 'perf stat' CSV and JSON output when
running on PowerPC due to an invalid topology number in that arch.
- Fix the 'perf test' for arm_coresight failures on the ARM Juno
system.
- Fix the 'perf test' attr entry for PERF_FORMAT_LOST, adding this
option to the or expression expected in the intercepted
perf_event_open() syscall.
- Add missing condition flags ('hs', 'lo', 'vc', 'vs') for arm64 in the
'perf annotate' asm parser.
- Fix 'perf mem record -C' option processing, it was being chopped up
when preparing the underlying 'perf record -e mem-events' and thus
being ignored, requiring using '-- -C CPUs' as a workaround.
- Improvements and tidy ups for 'perf test' shell infra.
- Fix Intel PT information printing segfault in uClibc, where a NULL
format was being passed to fprintf.
* tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (23 commits)
tools arch x86: Sync the msr-index.h copy with the kernel sources
perf auxtrace arm64: Add support for parsing HiSilicon PCIe Trace packet
perf auxtrace arm64: Add support for HiSilicon PCIe Tune and Trace device driver
perf auxtrace arm: Refactor event list iteration in auxtrace_record__init()
perf tests stat+json_output: Include sanity check for topology
perf tests stat+csv_output: Include sanity check for topology
perf intel-pt: Fix system_wide dummy event for hybrid
perf intel-pt: Fix segfault in intel_pt_print_info() with uClibc
perf test: Fix attr tests for PERF_FORMAT_LOST
perf test: test_intel_pt.sh: Add 9 tests
perf inject: Fix GEN_ELF_TEXT_OFFSET for jit
perf test: test_intel_pt.sh: Add jitdump test
perf test: test_intel_pt.sh: Tidy some alignment
perf test: test_intel_pt.sh: Print a message when skipping kernel tracing
perf test: test_intel_pt.sh: Tidy some perf record options
perf test: test_intel_pt.sh: Fix return checking again
perf: Skip and warn on unknown format 'configN' attrs
perf list: Fix metricgroups title message
perf mem: Fix -C option behavior for perf mem record
perf annotate: Add missing condition flags for arm64
...
Pull Kbuild fixes from Masahiro Yamada:
- Fix CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y compile error for the
combination of Clang >= 14 and GAS <= 2.35.
- Drop vmlinux.bz2 from the rpm package as it just annoyingly increased
the package size.
- Fix modpost error under build environments using musl.
- Make *.ll files keep value names for easier debugging
- Fix single directory build
- Prevent RISC-V from selecting the broken DWARF5 support when Clang
and GAS are used together.
* tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
kbuild: fix single directory build
kbuild: add -fno-discard-value-names to cmd_cc_ll_c
scripts/clang-tools: Convert clang-tidy args to list
modpost: put modpost options before argument
kbuild: Stop including vmlinux.bz2 in the rpm's
Kconfig.debug: add toolchain checks for DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT
Kconfig.debug: simplify the dependency of DEBUG_INFO_DWARF4/5
