On 2016-06-06 21:26, Karl-Heinz Schneider wrote: > Am Montag, den 06.06.2016, 12:16 +0200 schrieb Peter Rosin: >> This patch has two fundamental problems. First, the patch is whitespace- >> damaged (your mailer probably destroyed it). Please fix that. Second, and >> more serious, the patch has not even been compile-tested when applied on >> top of v4.7-rc1. > Sorry mixed things up in the makefile. Was happy the compiler didn't > complain about anything... Ok, some kind of silly mistake... >> The i2c-muxing code saw an update in 4.6 and this driver needs adjustment. >> As a hint, I suspect that you can make this mux "mux-locked" and avoid >> the dead-lock workaround in ltc1760_write_select_bat_reg(). > Difficult to test for me, since I'm forced to use an old 3.14 Kernel for > the target device (SECO imx6 board). I feel your pain. I always feel trapped when working on vendor kernels. The conversion is not difficult. Have a look at the commits leading up to 23fe440c59b9 "i2c: mux: drop old unused i2c-mux-api" to see examples of how to make it work for v4.7 (I think cddcc40b1b15 "[media] rtl2832: convert to use an explicit i2c mux core" looks like it could be similar to your case). If you then think the mux should be mux-locked (see Documentation/i2c/i2c-topology) it should be a simple matter of specifying the flag I2C_MUX_LOCKED when calling i2c_mux_alloc() and getting rid of the dead-lock workaround. Cheers, Peter -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
