linux-toradex.git/drivers/devfreq/rk3399_dmc.c, branch v6.3-rc1

PM / devfreq: rk3399_dmc: Block PMU during transitions

2022-05-08T18:38:07+00:00

See the previous patch ("soc: rockchip: power-domain: Manage resource
conflicts with firmware") for a thorough explanation of the conflicts.
While ARM Trusted Firmware may be modifying memory controller and
power-domain states, we need to block the kernel's power-domain driver.

If the power-domain driver is disabled, there is no resource conflict
and this becomes a no-op.

Signed-off-by: Brian Norris 
Reviewed-by: Douglas Anderson 
Acked-by: Chanwoo Choi 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Avoid static (reused) profile

2022-04-13T22:18:12+00:00

This static struct can get reused if the device gets removed/reprobed,
and that causes use-after-free in its ->freq_table.

Let's just move the struct to our dynamic allocation.

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Use devm_pm_opp_of_add_table()

2022-04-13T22:18:12+00:00

This simplifies error-cleanup and remove().

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Disable edev on remove()

2022-04-13T22:18:12+00:00

Otherwise we hit an unablanced enable-count when unbinding the DFI
device:

[ 1279.659119] ------------[ cut here ]------------
[ 1279.659179] WARNING: CPU: 2 PID: 5638 at drivers/devfreq/devfreq-event.c:360 devfreq_event_remove_edev+0x84/0x8c
...
[ 1279.659352] Hardware name: Google Kevin (DT)
[ 1279.659363] pstate: 80400005 (Nzcv daif +PAN -UAO -TCO BTYPE=--)
[ 1279.659371] pc : devfreq_event_remove_edev+0x84/0x8c
[ 1279.659380] lr : devm_devfreq_event_release+0x1c/0x28
...
[ 1279.659571] Call trace:
[ 1279.659582]  devfreq_event_remove_edev+0x84/0x8c
[ 1279.659590]  devm_devfreq_event_release+0x1c/0x28
[ 1279.659602]  release_nodes+0x1cc/0x244
[ 1279.659611]  devres_release_all+0x44/0x60
[ 1279.659621]  device_release_driver_internal+0x11c/0x1ac
[ 1279.659629]  device_driver_detach+0x20/0x2c
[ 1279.659641]  unbind_store+0x7c/0xb0
[ 1279.659650]  drv_attr_store+0x2c/0x40
[ 1279.659663]  sysfs_kf_write+0x44/0x58
[ 1279.659672]  kernfs_fop_write_iter+0xf4/0x190
[ 1279.659684]  vfs_write+0x2b0/0x2e4
[ 1279.659693]  ksys_write+0x80/0xec
[ 1279.659701]  __arm64_sys_write+0x24/0x30
[ 1279.659714]  el0_svc_common+0xf0/0x1d8
[ 1279.659724]  do_el0_svc_compat+0x28/0x3c
[ 1279.659738]  el0_svc_compat+0x10/0x1c
[ 1279.659746]  el0_sync_compat_handler+0xa8/0xcc
[ 1279.659758]  el0_sync_compat+0x188/0x1c0
[ 1279.659768] ---[ end trace cec200e5094155b4 ]---

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Support new *-ns properties

2022-04-13T22:18:12+00:00

We want to keep the idle time fixed, so compute based on the current DDR
frequency.

The old properties were deprecated and never used, so we can safely drop
them from the driver.

This is a rewritten version of work by Lin Huang .

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Support new disable-freq properties

2022-04-13T22:18:12+00:00

Implement the newly-defined properties to allow disabling certain
power-saving-at-idle features at higher frequencies.

This is a rewritten version of work by Lin Huang .

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Use bitfield macro definitions for ODT_PD

2022-04-13T22:18:12+00:00

We're going to add new usages, and it's cleaner to work with macros
instead of comments and magic numbers.

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Drop excess timing properties

2022-04-13T22:18:11+00:00

All of these properties are initialized by ARM Trusted Firmware, and
have been since the early days of this chip. It's redundant (and
possibly wrong) to do this here now. What's more, there seems to be some
confusion about the units and some of the definitions of this timing
struct: the DT docs say MHz for many of these, but downstream users were
in Hz (and therefore, the ATF interface was Hz). Also, the in-driver
usage for some of these (e.g., for comparing to target frequency) were
in Hz too. So doubly wrong.

We can avoid thinking about who got the right units by dropping the
unnecessary code and properties. They are marked deprecated in the
binding schema.

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Drop undocumented ondemand DT props

2022-04-13T22:18:11+00:00

These properties are:

* undocumented
* directly representing software properties, not hardware properties
* unused (no in-tree users, yet; this IP block has so far only been used
  in downstream kernels)

Let's just stick the values that downstream users have been using
directly in the driver and call it a day.

Signed-off-by: Brian Norris 
Signed-off-by: Chanwoo Choi

PM / devfreq: rk3399_dmc: Simplify with dev_err_probe()

2021-04-08T04:14:50+00:00

Common pattern of handling deferred probe can be simplified with
dev_err_probe().  Less code and the error value gets printed.

Signed-off-by: Krzysztof Kozlowski 
Signed-off-by: Chanwoo Choi