* ARM/DT: registering clks from DT in .init_early
@ 2011-06-09 16:45 Jamie Iles
2011-06-09 19:28 ` Jamie Iles
0 siblings, 1 reply; 2+ messages in thread
From: Jamie Iles @ 2011-06-09 16:45 UTC (permalink / raw)
To: linux-arm-kernel
I'm currently working on moving our platform to device tree with the aim
of mainlining it. Before device tree we registered the clks in the
.init_early machine callback which is fine as the clks are static data.
I'm now trying to dynamically register these clks from the device tree
but the slab allocator isn't available at this stage so I can't allocate
them dynamically.
Am I heading in the wrong direction here? We currently get the
clocksource and clock event frequency from a clk so I don't think this
can be made later (though I guess this could come from device tree
instead).
Jamie
^ permalink raw reply [flat|nested] 2+ messages in thread
* ARM/DT: registering clks from DT in .init_early
2011-06-09 16:45 ARM/DT: registering clks from DT in .init_early Jamie Iles
@ 2011-06-09 19:28 ` Jamie Iles
0 siblings, 0 replies; 2+ messages in thread
From: Jamie Iles @ 2011-06-09 19:28 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Jun 09, 2011 at 05:45:57PM +0100, Jamie Iles wrote:
> I'm currently working on moving our platform to device tree with the aim
> of mainlining it. Before device tree we registered the clks in the
> .init_early machine callback which is fine as the clks are static data.
> I'm now trying to dynamically register these clks from the device tree
> but the slab allocator isn't available at this stage so I can't allocate
> them dynamically.
>
> Am I heading in the wrong direction here? We currently get the
> clocksource and clock event frequency from a clk so I don't think this
> can be made later (though I guess this could come from device tree
> instead).
OK, I found http://permalink.gmane.org/gmane.linux.ports.arm.kernel/105261
which describes the issue. I've changed to alloc_bootmem() which Grant
suggested in that thread and that does resolve the crash, however
of_clk_add_provider() uses kzalloc() and I can't help but feel that it's
going to get a bit too messy doing it this way.
Jamie
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-06-09 19:28 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-06-09 16:45 ARM/DT: registering clks from DT in .init_early Jamie Iles
2011-06-09 19:28 ` Jamie Iles
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).