<qi-commits>
Maarten ter Huurne: jz4740-i2s: Declare that both the JZ4740 AIC and its internal codec are capable of 48 kHz output. http://qi-hw.com/p/qi-kernel/43d2cfc
<qi-commits>
Maarten ter Huurne: jz4740-i2s: Corrected DMA width constants. This driver relied on the wrong encoding in arch/mips/include/asm/mach-jz4740/dma.h, which were fixed recently. http://qi-hw.com/p/qi-kernel/64c7b5b
<qi-commits>
Maarten ter Huurne: jz4740_dma: Only set no-descriptor flag when enabling the channel. Otherwise, there can be data loss using the transfer (seen with SLCD). http://qi-hw.com/p/qi-kernel/cd0f502
<larsc>
mth: could you have a look and see if anything is missing
<mth>
flushing the dcache, or do you have an alternative commit that addresses that issue?
<larsc>
no. but it's not really neccessary. unless you want to use the bounce buffer
<mth>
there are other drivers using that call, afaik
<mth>
at least XFS, although probably no-one is going to use that on a mobile device
<mth>
but maybe more (either now or in the future)
<mth>
by the way, is there a plan to integrate the qi-kernel patches into the mainline kernel at some point?
<larsc>
yes
<larsc>
soon(tm)
<larsc>
i wanted to send them for review after the merge window
<mth>
in that case, maybe the i2c driver from OpenInkpot would be good to include as well
<mth>
it's a bit early for the SLCD driver, but that might be a candidate in the future
<mth>
speaking of the SLCD, some of the commits on that apply to the code that I copied from your LCD driver
<mth>
for example part of 558d011e3bbf77ae683d
<mth>
about the green offset in 15 and 16 bpp formats
<mth>
the board headers are not really needed anymore, I think, for example in 5afb3e8c011764410854 I removed board-a320.h
<mth>
default keyboards: I don't know if 28b9083b245ab4a3b1be is the right approach, but it would make sense not to select the AT keyboard by default on non-x86 systems
<mth>
that's all
<mth>
so you got all the commits that could be cherry picked, the other commits I mentioned are changes that could only be merged manually
<larsc>
good
<mth>
about the i2c driver: it does not give the same results as the GPIO I2C driver on Dingoo, but I don't know whether that is because of a driver problem or strange hardware
<mth>
for example the list of devices found with GPIO I2C is smaller, but far more likely to be true
<mth>
but accessing the Chrontel TV-out encoder works with either driver
<larsc>
hm
<xakh>
hey bitches
<xakh>
what's everybody doin?
<xakh>
hey guys
<xakh>
anyone wanna help me put PRBoom on my Gmenu2X home menu?