linux-stable.git/drivers/opp/core.c, branch v6.5.2

OPP: Simplify the over-designed pstate <-> level dance

2023-06-19T04:20:09+00:00

While adding support for "performance states" in the OPP and genpd core,
it was decided to set the `pstate` field via genpd's
pm_genpd_opp_to_performance_state() helper, to allow platforms to set
`pstate` even if they don't have a corresponding `level` field in the DT
OPP tables (More details are present in commit 6e41766a6a50 ("PM /
Domain: Implement of_genpd_opp_to_performance_state()")).

Revisiting that five years later clearly suggests that it was
over-designed as all current users are eventually using the `level`
value only.

The previous commit already added necessary checks to make sure pstate
is only used for genpd tables. Lets now simplify this a little, and use
`level` directly and remove `pstate` field altogether.

Suggested-by: Ulf Hansson 
Signed-off-by: Viresh Kumar 
Reviewed-by: Ulf Hansson

OPP: pstate is only valid for genpd OPP tables

2023-06-19T04:20:02+00:00

It is not very clear right now that the `pstate` field is only valid for
genpd OPP tables and not consumer tables. And there is no checking for
the same at various places.

Add checks in place to verify that and make it clear to the reader.

Signed-off-by: Viresh Kumar 
Reviewed-by: Ulf Hansson 
Reviewed-by: Bjorn Andersson 
Tested-by: Bjorn Andersson

OPP: don't drop performance constraint on OPP table removal

2023-06-14T09:24:01+00:00

This code was added (long back) by commit 009acd196fc8 ("PM / OPP:
Support updating performance state of device's power domain") and at
that time the `opp->pstate` field was used to store the performance
state required by a device's OPP.

Over time that changed and the `->pstate` field is now used only for
genpd devices and consumer devices access that via the required-opps
instead.

Because of all these changes, _opp_table_kref_release() now drops the
constraint only when the genpd's OPP table gets freed and not the
device's. Which is definitely not what we wanted. And dropping the
constraint doesn't have much meaning as the genpd itself is going away.

Moreover, if we want to drop constraints here, then just dropping the
performance constraint alone isn't sufficient as there are other
resource constraints like clk, regulator, etc. too, which must be
handled.

Probably the right thing to do here is to leave this decision to the
consumers, which can call `dev_pm_opp_set_rate(dev, 0)` or similar APIs
to drop all constraints properly. Which many of the consumers already
do.

Remove the special code, which is broken anyway.

Signed-off-by: Viresh Kumar 
Reviewed-by: Ulf Hansson

OPP: Staticize `lazy_opp_tables` in of.c

2023-06-08T07:25:07+00:00

`lazy_opp_tables` is only used in of.c, move it there and mark it
`static`.

Signed-off-by: Viresh Kumar

opp: Fix use-after-free in lazy_opp_tables after probe deferral

2023-05-31T03:51:41+00:00

When dev_pm_opp_of_find_icc_paths() in _allocate_opp_table() returns
-EPROBE_DEFER, the opp_table is freed again, to wait until all the
interconnect paths are available.

However, if the OPP table is using required-opps then it may already
have been added to the global lazy_opp_tables list. The error path
does not remove the opp_table from the list again.

This can cause crashes later when the provider of the required-opps
is added, since we will iterate over OPP tables that have already been
freed. E.g.:

  Unable to handle kernel NULL pointer dereference when read
  CPU: 0 PID: 7 Comm: kworker/0:0 Not tainted 6.4.0-rc3
  PC is at _of_add_opp_table_v2 (include/linux/of.h:949
  drivers/opp/of.c:98 drivers/opp/of.c:344 drivers/opp/of.c:404
  drivers/opp/of.c:1032) -> lazy_link_required_opp_table()

Fix this by calling _of_clear_opp_table() to remove the opp_table from
the list and clear other allocated resources. While at it, also add the
missing mutex_destroy() calls in the error path.

Cc: stable@vger.kernel.org
Suggested-by: Viresh Kumar 
Fixes: 7eba0c7641b0 ("opp: Allow lazy-linking of required-opps")
Signed-off-by: Stephan Gerhold 
Signed-off-by: Viresh Kumar

OPP: rate-limit debug messages when no change in OPP is required

2023-05-22T03:41:40+00:00

Otherwise, when enabling a debug build and dynamic debug in the kernel, it
quickly floods the kernel ring buffer and makes debugging of other
subsystems almost impossible, unless manually disabled.

Signed-off-by: Adrián Larumbe 
Signed-off-by: Viresh Kumar

OPP: Move required opps configuration to specialized callback

2023-04-03T04:15:09+00:00

The required-opps configuration is closely tied to genpd and performance
states at the moment and it is not very obvious that required-opps can
live without genpds. Though we don't support configuring required-opps
for non-genpd cases currently.

This commit aims at separating these parts, where configuring genpds
would be a special case of configuring the required-opps.

Add a specialized callback, set_required_opps(), to the opp table and
set it to different callbacks accordingly.

This shouldn't result in any functional changes for now.

Signed-off-by: Viresh Kumar 
Reviewed-by: Ulf Hansson

OPP: Handle all genpd cases together in _set_required_opps()

2023-04-03T04:15:09+00:00

There is no real need of keeping separate code for single genpd case, it
can be made to work with a simple change.

Signed-off-by: Viresh Kumar 
Reviewed-by: Ulf Hansson

OPP: Fix an un-initialized variable usage

2022-08-16T05:18:08+00:00

smatch complains that 'ret' may be returned un-initialized.

Explicitly return 0 if we reach the end of the function (should
'opp_table->clk_count' be 0).

Fixes: 8174a3a613af ("OPP: Provide a simple implementation to configure multiple clocks")
Signed-off-by: Christophe JAILLET 
Signed-off-by: Viresh Kumar

OPP: Don't drop opp->np reference while it is still in use

2022-07-19T05:37:02+00:00

The struct dev_pm_opp contains a reference of the DT node, opp->np,
throughout its lifetime. We should increase the refcount for the same
from _opp_add_static_v2(), and drop it while removing the OPP finally.

Signed-off-by: Liang He 
[ Viresh: Updated subject / commit log, create _of_clear_opp() and drop
	  reference from it]
Signed-off-by: Viresh Kumar