Some xo require special value to set it to /4 divisor. Introduce
xo_div4_val to handle this special implementation.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Unsigned long have a different size in 32 bit system and this cause the
malfunction of debugcc returning wrong frequency. Use 64bit unsigned int
to correctly support 32 bit system and provide correct frequency.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Enforcing a primary+leaf configuration does not walk for all the cases.
Some clocks (msm8996 CPU) have more muxes to be powered on. Other (MCCC)
do not require GCC at all. Make clock measurement more versatile by
reworking GCC + optional leaf into a chain of muxes.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Having a template file with replacement macros is a bit cleaner than
encoding the setup of the file in `meson.build` and piping that through
`echo` with caputred variables.
Unfortunately there appears to be no support for repetition, to e.g.
expand a list of platform names into the various uses (array with comma-
separation, or individual `extern struct` declarations, forcing us to
have some form of string expansion within `meson.build`).
Fix the following error:
../debugcc.c: In function 'mmap_mux':
../debugcc.c:295:40: error: format '%zx' expects argument of type 'size_t', but argument 2 has type 'long unsigned int' [-Werror=format=]
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Switch to meson as a build system. Define a list of supported platforms
and use it to generate platforms list and the symlink targets.
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Add a README with a reminder to switch off CONFIG_IO_STRICT_DEVMEM. I
had debugcc working on 8939 and then tried it on 8280xp. It didn't work
but I remembered that there was a CONFIG option that needed to change.
It took a few minutes of menuconfig and Google to recall what that
config option was.
For future reference, lets just document it and whatever else in a
README.
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Hook MCCC up on SM6375 where it was validated and intentionally overreserve
the MCCC memory, as mmap expects we reserve at least one ALIGNED page
(which means we're aligned to 4K (0x1000) or whatever our pagesize is AND
that we have to reserve too much and stray from user error when adding
offsets to not try and read memory outside the actual MCCC region).
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
