From mboxrd@z Thu Jan 1 00:00:00 1970 From: sebastian.hesselbarth@gmail.com (Sebastian Hesselbarth) Date: Tue, 20 Aug 2013 22:01:29 +0200 Subject: [RFC 05/17] ARM: highbank: remove custom .init_time hook In-Reply-To: <5213BF91.3030300@calxeda.com> References: <1376964271-22715-1-git-send-email-sebastian.hesselbarth@gmail.com> <1376964271-22715-6-git-send-email-sebastian.hesselbarth@gmail.com> <5213BF91.3030300@calxeda.com> Message-ID: <5213CB19.6000905@gmail.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 08/20/2013 09:12 PM, Rob Herring wrote: > On 08/19/2013 09:04 PM, Sebastian Hesselbarth wrote: >> With arch/arm calling of_clk_init(NULL) from time_init(), we can now >> remove custom .init_time hooks. Highbank clock provider need a reference >> to system registers, as a workaround current clk driver maps those >> independent of arch code now. >> >> Signed-off-by: Sebastian Hesselbarth >> --- >> Notes: >> - I suggest to consolidate the four single clock providers into a single >> provider with one node if possible. > > And how do you suggest doing that while maintaining the DT ABI? Rob, I am aware of ABI status of the DT bindings and the discussion about deprecating/cleanup of what went in in the past. Maybe "suggest" is a little bit too strong and should have been "would be nice to have". But looking at clk-highbank, all providers depend on the same register. If you'd introduce e.g. a new "calxeda,core-clocks" that sets up the four providers and also CLK_OF_DECLARE the same init with one out of the four existing compatibles shouldn't that maintain backwards compatibility? Anyway, I was just puting the note above to raise some comment on it. You actually gave a good one to leave it as is. >> --- >> arch/arm/mach-highbank/highbank.c | 23 +++++++---------------- >> drivers/clk/clk-highbank.c | 10 +++++++--- >> 2 files changed, 14 insertions(+), 19 deletions(-) > > Acked-by: Rob Herring Ok, I keep that in mind until all other issues are resolved. Thanks, Sebastian