[s3c6410 BSP] Funnies and questions

Ben Dooks ben at simtec.co.uk
Mon Nov 3 10:46:34 CET 2008


On Mon, Nov 03, 2008 at 09:35:53AM +0000, Andy Green wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Somebody in the thread at some point said:
> 
> |> 1) enable_irq_wake(S3C_EINT(4)); fails, it should be OK to enable EINT4
> |> as a wake source
> |
> | Without any of the power management there is no set_wake methods set
> on the
> | interrupts.
> 
> Ah OK then.
> 
> |> 2) s3c6400-uart: probe of s3c6400-uart.3 failed with error -22
> |>
> |> But there are 4 UARTs on S3C6410
> |
> | I'll update the Samsung serial driver to support 4 UARTs instead of 3.
> 
> Great.  On GTA03 debug console happens to be on the fourth one.
> 
> |> 3) We did not fit the 27MHz clock on S3C6410, so we need to disable it
> |> system-wide as a clk source... eg for MMC base clock... what's the best
> |> way to do that?
> |
> | This information will be available in the platform data registered with
> | the system.
> 
> OK I take that it means it will be exposed to the machine-specific stuff
> at some point later... in the meanwhile I will patch
> arch/arm/plat-s3c64xx/s3c6400-clock.c or stuff around there since it
> seems to select 27MHz clock as the default source for the HS_MMC0[1].

That's something your bootloader should really setup, all this code
is doing is reading the current settings for where the clocks are
sourced from. Either set this in the bootloader, or it should be
allowable to set these registers early in the machine specific
setup.
 
> [    0.000000] mmc_bus: source is clk_27m (3), rate is 27000000
> 
> |> 4) HS_MMC is a valid platform device and is initialized, but it doesn't
> |> see the SDHC card.  Assuming this is missing somewhat, what's the status
> |> on it?
> |
> | Furhter testing is being done on the SDHCI, will release a set of what
> | I consider stable later today.
> 
> Great it will be very handy to be able to mount a rootfs from there and
> do SDIO.
> 
> Thanks,
> 
> - -Andy
> 
> [1] This nearly caused a heart attack when I noticed it yesterday since
> it looked like the iROM was using the (NC on GTA03) 27MHz source, but I
> verified in bootloader it's EPll.

Hmm, iirc the EPLL is selected via the mmc_bus clock, it might be using
the HCLK fed into the unit which is another possible clock source?

-- 
Ben

Q:      What's a light-year?
A:      One-third less calories than a regular year.




More information about the openmoko-kernel mailing list