<Tenkawa>
I'm still so tempted to get a chromebook 2 13"
<javier__>
is a nice machine, and I try to keep it in good shape with mainline
<javier__>
the only missing bit is the broken audio
<Tenkawa>
yeah.. i use an altec lansing usb dev for that anyway
<Tenkawa>
even this dual core snow is a really decent box
<sjoerd>
javier__: how come that's still broken ooi?
<Tenkawa>
javier__: thanks again.. bbl. need to run some errands
Tenkawa has quit [Quit: leaving]
<javier__>
sjoerd: AFAIU the problem is that the snow ASoc machine driver lacks a .hw_params that should change the master clock rate
<javier__>
but maybe the machine could just use the simple-audio driver
<javier__>
sjoerd: but the symptom is that you get overruns when doing playback, I spent some days trying to fix it last year but failed
<javier__>
asked to some of the people that were involved with the audio support for the chromiumos tree and also the mainline driver but it seems they don't have time to look at it either
<sjoerd>
javier__: overrun on output? so the engine is just not consuming it at a fast enough rate then
<javier__>
sjoerd: yes, I believe the problem is that the clock rate is fixed while the downstream tree changes it
<javier__>
but the downstream tree has a complex clk tree that setups on probe and mainline has a single clock
<sjoerd>
fun fun
<javier__>
the fact that I'm pretty ignorant about audio in general doens't make it easy...
<javier__>
*easier
<LiquidAcid>
javier__, so we should send you some good mp3s then? ;)
<sjoerd>
javier__: in the end it'll just mostly be a dma engine copying data from memory onto e.g. an i2s bus at a given rate ;) everything else is pretty opaque
<sjoerd>
or well the receving end of a dma transfer for the socs side of it
<javier__>
sjoerd: yeah, I think the key is the given rate part which is fixed in the ASoC machine driver
<javier__>
LiquidAcid :D
<javier__>
but right now I'm debugging locking issues on media_device_ioctl()...
<LiquidAcid>
javier__, i just remembered, were you planning on adding some kconfig check for the fimc modules?
<javier__>
LiquidAcid: how so? what do you think that is missing?
<LiquidAcid>
javier__, iirc then there currently is nothing preventing that both exynos4-is and drm/fimc are build, which looks like an error since all other components are checked (gsc, g2d, s5p-tv, etc.)
<javier__>
LiquidAcid: I see, I didn't know that exynos4-is had conflicts with the DRM driver
<javier__>
LiquidAcid: I wonder if makes sense then to enable those as I do it with my patch
<LiquidAcid>
javier__, i haven't used exynos4-is myself, but it uses the fimc block so i assume it uses the same hw resources
<javier__>
I added more for build coverage since we are doing media wide changes in the API
<javier__>
LiquidAcid: I see, I haven't used it either
<LiquidAcid>
i've only used IPP, despite marek warning me about it :)
<javier__>
hehe
<javier__>
my test system is an OMAP3 board, but I try to make sure that we don't introduce regressions on other media platforms drivers
<LiquidAcid>
oh nice, i still have a pandaboard es lying around here, but haven't used it in ages
<javier__>
but if the Exynos DRM driver conflicts with exynos4-is, then probably that shouldn't be enabled in exynos_defconfig
<LiquidAcid>
yeah, at that time when i wrote the msg on the ml i was only sure about gsc and g2d
<LiquidAcid>
sry
<javier__>
LiquidAcid: also I wonder if the Kconfig check shouldn't be the opposite, to only enable the media platform ones if the Exynos DRM components are not enabled
<javier__>
after all IIUC the media drivers are the legacy ones and the DRM is what should be used instead
<LiquidAcid>
at least for the g2d driver that is only half-true, the media one supports both the old and new hw revision, while the drm one supports only the new revision
<javier__>
LiquidAcid: I see
<LiquidAcid>
but the media g2d is in bad shape, so yeah...
<javier__>
LiquidAcid: Ok, I commented in the ML what you mentioned and asked Krzysztof to drop the defconfig patch
<javier__>
LiquidAcid: I'll post a patch to check for !VIDEO_S5P_FIMC in DRM_EXYNOS_FIMC, probably later today
<LiquidAcid>
javier__, roger!
paulk-aldrin has quit [Remote host closed the connection]