[PATCH AUTOSEL 5.4 6/8] clk: Don't hold prepare_lock when calling kref_put()

[PATCH AUTOSEL 5.4 6/8] clk: Don't hold prepare_lock when calling kref_put()

[Sasha Levin]

From: Stephen Boyd <sboyd@kernel.org>

[ Upstream commit 6f63af7511e7058f3fa4ad5b8102210741c9f947 ]

We don't need to hold the prepare_lock when dropping a ref on a struct
clk_core. The release function is only freeing memory and any code with
a pointer reference has already unlinked anything pointing to the
clk_core. This reduces the holding area of the prepare_lock a bit.

Note that we also don't call free_clk() with the prepare_lock held.
There isn't any reason to do that.

Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Link: https://lore.kernel.org/r/20240325184204.745706-3-sboyd@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/clk/clk.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
index 67051ca60920a..9dddf5f555ab5 100644
--- a/drivers/clk/clk.c
+++ b/drivers/clk/clk.c
@@ -3971,7 +3971,8 @@ void clk_unregister(struct clk *clk)
 	if (clk->core->ops == &clk_nodrv_ops) {
 		pr_err("%s: unregistered clock: %s\n", __func__,
 		       clk->core->name);
-		goto unlock;
+		clk_prepare_unlock();
+		return;
 	}
 	/*
 	 * Assign empty clock ops for consumers that might still hold
@@ -4002,11 +4003,10 @@ void clk_unregister(struct clk *clk)
 	if (clk->core->protect_count)
 		pr_warn("%s: unregistering protected clock: %s\n",
 					__func__, clk->core->name);
+	clk_prepare_unlock();
 
 	kref_put(&clk->core->ref, __clk_release);
 	free_clk(clk);
-unlock:
-	clk_prepare_unlock();
 }
 EXPORT_SYMBOL_GPL(clk_unregister);
 
@@ -4168,13 +4168,11 @@ void __clk_put(struct clk *clk)
 	    clk->max_rate < clk->core->req_rate)
 		clk_core_set_rate_nolock(clk->core, clk->core->req_rate);
 
-	owner = clk->core->owner;
-	kref_put(&clk->core->ref, __clk_release);
-
 	clk_prepare_unlock();
 
+	owner = clk->core->owner;
+	kref_put(&clk->core->ref, __clk_release);
 	module_put(owner);
-
 	free_clk(clk);
 }
 
-- 
2.43.0

Re: [PATCH AUTOSEL 5.4 6/8] clk: Don't hold prepare_lock when calling kref_put()

[Stephen Boyd]

Quoting Sasha Levin (2024-04-23 04:03:01)
> From: Stephen Boyd <sboyd@kernel.org>
> 
> [ Upstream commit 6f63af7511e7058f3fa4ad5b8102210741c9f947 ]
> 
> We don't need to hold the prepare_lock when dropping a ref on a struct
> clk_core. The release function is only freeing memory and any code with
> a pointer reference has already unlinked anything pointing to the
> clk_core. This reduces the holding area of the prepare_lock a bit.
> 
> Note that we also don't call free_clk() with the prepare_lock held.
> There isn't any reason to do that.

You'll want the patch before this, 8358a76cfb47 ("clk: Remove
prepare_lock hold assertion in __clk_release()"), to avoid lockdep
warnings. And it looks like the problem was reported on v5.15.y so all
5 patches from the series would need a backport.

 8358a76cfb47 clk: Remove prepare_lock hold assertion in __clk_release()
 6f63af7511e7 clk: Don't hold prepare_lock when calling kref_put()
 9d05ae531c2c clk: Initialize struct clk_core kref earlier
 e581cf5d2162 clk: Get runtime PM before walking tree during disable_unused
 9d1e795f754d clk: Get runtime PM before walking tree for clk_summary

Re: [PATCH AUTOSEL 5.4 6/8] clk: Don't hold prepare_lock when calling kref_put()

[Sasha Levin]

On Tue, Apr 23, 2024 at 12:24:51PM -0700, Stephen Boyd wrote:
>Quoting Sasha Levin (2024-04-23 04:03:01)
>> From: Stephen Boyd <sboyd@kernel.org>
>>
>> [ Upstream commit 6f63af7511e7058f3fa4ad5b8102210741c9f947 ]
>>
>> We don't need to hold the prepare_lock when dropping a ref on a struct
>> clk_core. The release function is only freeing memory and any code with
>> a pointer reference has already unlinked anything pointing to the
>> clk_core. This reduces the holding area of the prepare_lock a bit.
>>
>> Note that we also don't call free_clk() with the prepare_lock held.
>> There isn't any reason to do that.
>
>You'll want the patch before this, 8358a76cfb47 ("clk: Remove
>prepare_lock hold assertion in __clk_release()"), to avoid lockdep
>warnings. And it looks like the problem was reported on v5.15.y so all
>5 patches from the series would need a backport.
>
> 8358a76cfb47 clk: Remove prepare_lock hold assertion in __clk_release()
> 6f63af7511e7 clk: Don't hold prepare_lock when calling kref_put()
> 9d05ae531c2c clk: Initialize struct clk_core kref earlier
> e581cf5d2162 clk: Get runtime PM before walking tree during disable_unused
> 9d1e795f754d clk: Get runtime PM before walking tree for clk_summary

Ack, looks like its already the case. Thanks!

-- 
Thanks,
Sasha