Hi,
I am trying to get both CSI2 interfaces up and running through the ISS
on the v3.16 kernel for the TI OMAP4 Blaze platform (omap4430 ES 2.3
revision).
Trough a omap_device_build() call (using the "iss" omap_hwmod) I call the
iss_probe function. It devm_clk_get's both the iss_ctrlclk and the iss_fck.
Since I am building the kernel with the omap4-sdp-es23plus device tree
appended, I figured I need to define the iss_fck in the omap44xx-clocks.dtsi
file, right after the iss_ctrlclk, as following:
iss_fck: iss_fck {
#clock-cells = <0>;
compatible = "ti,gate-clock";
clocks = <&ducati_clk_mux_ck>;
ti,bit-shift = <1>;
reg = <0x1020>;
};
For that, I used the information in [1], the TI clock tree tool and the
Linux documentation.
Now the omap4iss_get() call throws L3 Standard Errors, right after the first
time the interrupts, set in iss_enable_interrupts, occur. I am pretty sure the
cause for that is a wrong initialization of the iss_fck (since I haven't
changed much more), even though the kernel runs through and the
"cat clk_summary ¦ grep iss" command in /sys/kernel/debug/clk/ writes:
iss_ctrlclk 0 1 96000000 0
iss_fck 0 1 400000000 0
Could there be an error in the device tree entry stated above? Or might I be
missing something else? Has anyone ever enabled iss_fck through device tree?
BTW I've also added iss_fck as an opt_clk in omap_hwmod_44xx_data and added
the clock to the ti_dt_clk struct.
Thanks,
Martin
[1] http://www.ti.com/pdfs/wtbu/OMAP4430_ES2.x_PUBLIC_TRM_vAJ.zip
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
