On Tue, Oct 28, 2014 at 08:00:45PM +0300, Max Filippov wrote:
On Tue, Oct 28, 2014 at 6:42 PM, Mark Brown broonie@kernel.org wrote:
Then this just shouldn't exist at all, adding Kconfig entries for all the simple-card devices would defeat the point of having simple-card which is why we don't do it for other systems.
sound/soc/sh/Kconfig does that as well. Not having single config item to enable at once all relevant drivers feels a bit strange... But ok, I'll drop that.
This is new code, not a replacement of old code.
So atomics don't work? Simple flags are one of the few cases where they might cover it... again, the fact that this isn't similar to other code doing the same thing is worrying.
tx_substream use pattern is the same as of a typical RCU variable. RCU wrappers combine ACCESS_ONCE and barriers that I'd otherwise have to write myself.
You *really* need to explain how it's supposed to work - right now it's not at all obvious, like I say the fact that this is a rarely used idiom is not helping. For example when we tear down the stream we just assign the pointer in _stop() but don't bother with a sync until the stream is closed - why? We also appear to be doing several different dereferences of the pointer inside a single rcu_read_lock() for some reason which is worrying.
I would also expect to see the data transfer and I2S bits more split out, presumably this IP can actually be used in designs with a DMA controller and one would expect that for practical systems this would be the common case?
I don't know of other designs that use this IP block. Can we do it when we get such users?
It's also about ensuring that the code is cleanly split up so that someone can actually go in and add the required support later (and TBH
Can you point me to an example of such split, so that I don't write it in an unusual way?
Essentially all drivers are split this way...
it'd be better to just add a DMA controller on the FPGA, everyone will be much happier).
Hardware people think different and it's been like that for more than 5 years.
They appear to be the only hardware people who think this way, while we do have some FIQ based PIO stuff in mainline all the examples I can think of are there because people didn't work out how to drive the DMA controller yet (and even there we're using FIQs not bashing things in from a regular interrupt).
Because this callback is said to be potentially called multiple times per initialization. Is it not?
It can but but I'm not seeing any connection between that and the idea of not keeping the clock enabled when the device isn't in use?
hw_params callback can change MCLK rate, so it has to disable and enable the clock anyway, and since enable can fail it does not guarantee that the clock will be left in the same state. Or should I adjust MCLK rate w/o disabling the clock?
So yet again: why not just enable the clock only when the device is in use? If it's being configured it stands to reason that the device isn't actively in use...
I'm not sure I find that terribly convincing, I'd like to be able to look at the code and tell that it's not going to blow up. This again comes back to the general comment about clarity - the code *looks* suspicious (strange indentation of the for loop with a comment in the middle of the statement itself for example).
The level field in the control register is 4 bit wide, so the allowed range of level is 0..15. FIFO size is 8192 entries, level = 1 corresponds to FIFO size / 2, level = 14 -- to FIFO size of 0. I guess this function won't get period_size of 0?
So if the IP gets changed and the code gets blown up this could well explode then... which doesn't seem entirely unlikely considering this is a FPGA platform so presumably this is easy to update. To repeat this is about clarity and the code looking like it's probably hiding bugs as much as if the code actually works if you really sit down and study it.