Re: [RFC PATCH 02/34] msm: clock: Always use an array to iterateover clocks

From: Stephen Boyd
Date: Wed Nov 02 2011 - 17:34:12 EST


On 11/02/11 12:45, Russell King - ARM Linux wrote:
> On Wed, Nov 02, 2011 at 11:35:59AM -0700, David Brown wrote:
>> If the array of clk_lookups contains aliases for the same struct
>> clk, msm_clock_init() will add the clock to the clocks list
>> twice. This would cause list corruption so let's just remove the
>> clocks list and any associated code and iterate over the array
>> instead.
> Hmm...
>
>> @@ -158,13 +152,13 @@ void __init msm_clock_init(struct clk_lookup *clock_tbl, unsigned num_clocks)
>> */
>> static int __init clock_late_init(void)
>> {
>> + unsigned i, count = 0;
>> unsigned long flags;
>> - struct clk *clk;
>> - unsigned count = 0;
>>
>> clock_debug_init();
>> - mutex_lock(&clocks_mutex);
>> - list_for_each_entry(clk, &clocks, list) {
>> + for (i = 0; i < msm_num_clocks; i++) {
>> + struct clk *clk = msm_clocks[i].clk;
>> +
>> clock_debug_add(clk);
> This means you'll end up calling clock_debug_add() twice for the same
> struct clk - this sounds like a bad idea in itself. It looks like
> there's no protection within that function against it being called
> twice with the same struct clk.
>
> Are you sure this is safe?

This hasn't proven to be a problem so far because debugfs returns an
error when you create a directory with the same name twice. If we ever
do something more in clock_debug_add() we would have a problem.

--
Sent by an employee of the Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/