[PATCH 1/6] clk: tegra: don't abort clk init on error
From: pdeschrijver@nvidia.com (Peter De Schrijver)
Date: 2014-08-15 22:45:20
Also in:
linux-devicetree, linux-tegra, lkml
From: pdeschrijver@nvidia.com (Peter De Schrijver)
Date: 2014-08-15 22:45:20
Also in:
linux-devicetree, linux-tegra, lkml
On Tue, Jul 22, 2014 at 07:16:15PM +0200, Stephen Warren wrote:
On 07/15/2014 09:24 AM, Peter De Schrijver wrote:quoted
Just continue initializing clocks if there's an error on one of them. This is useful if there's a mistake in the inittable, because the system could hang if clk_disable_unused() disables some of the critical clocks in this table.If there's a problem in the init table, we should simply fix it instead of working around it.
Yes, ofcourse. However today we silently stop processing the init_table if a clock cannot be found. That doesn't sound right either to me and makes detecting wrong entries in the table more complex than it should be.
At the very least, we need to WARN on this rather than just ignoring problems.
Cheers, Peter.