Re: [PATCH RFC 10/12] mmc: mmci: Add clock support for Qualcomm.

2014-04-22 Thread Srinivas Kandagatla
Thankyou Stephen, On 21/04/14 23:20, Stephen Boyd wrote: Please turn on lockdep (PROVE_LOCKING) and sleeping while atomic checks (DEBUG_ATOMIC_SLEEP). You cannot call clk_set_rate() in atomic context. You are correct, there is a spinlock taken just before entering this code. I will fix this in

[PATCH RFC 10/12] mmc: mmci: Add clock support for Qualcomm.

2014-04-21 Thread srinivas . kandagatla
From: Srinivas Kandagatla srinivas.kandaga...@linaro.org MCICLK going to card bus is directly driven by the clock controller, so the driver has to set the required rates depending on the state of the card. This bit of support is very much similar to bypass mode but there is no such thing called

Re: [PATCH RFC 10/12] mmc: mmci: Add clock support for Qualcomm.

2014-04-21 Thread Stephen Boyd
On 04/21/14 14:49, srinivas.kandaga...@linaro.org wrote: From: Srinivas Kandagatla srinivas.kandaga...@linaro.org MCICLK going to card bus is directly driven by the clock controller, so the driver has to set the required rates depending on the state of the card. This bit of support is very