Re: [Gumstix-users] Occasional boot failures

[William H. <wi...@ne...>]

Joe,

Thanks for sharing this.  I'm sure Phil Lutz will be interested in seeing if this applies to his situation.

	William


On 11/18/2011 6:38 AM, Joe Woodward wrote:
> I have noticed problems booting an Overo Fire (R3118) reliably, maybe once in every half dozen boots fails.
>
> This is seen when the "normal" set of boot messages out of the serial port appear but nothing else.
> "£t?b$U
>           @D¨±  à"
>
> This shows that the OMAP is able to start just fine, but somewhere in the loading or execution of x-loader something goes wrong.
> I'm booting from an MLO (x-loader version 1.5.1) on an SDCard. Older versions (such as 1.4ss) work 100% of the time.
>
> I've found a "hacky" work around is to force the check for the chip revision in x-loader to an ES1.0 chip (in fact the chips are ES3.1, see overo.c within the x-
> loader sources). I believe this then applies slower timings for the various clocks (including the SDRAM). Since doing this I've never had a boot failure, which
> makes me believe that some of the timings in x-loader may be marginal.
>
> (overo.c:prcm_init()):
> 	/* The DPLL tables are defined according to sysclk value and
> 	 * silicon revision. The clk_index value will be used to get
> 	 * the values for that input sysclk from the DPLL param table
> 	 * and sil_index will get the values for that SysClk for the
> 	 * appropriate silicon rev.
> 	 */
> //	sil_index = (get_cpu_rev() == CPU_3XX_ES10) ? 0 : 1;
> 	sil_index = 0;
>
>
> I believe the kernel will re-program these clocks correctly when things get going anyway, so the issue is only seen at boot time.
>
> I've seen the odd post about reliable booting, and would be interested to see if this change makes a difference for those people.
>
> Does anyone else see any of these problems?
>
> Cheers,
> Joe
>
>
>
>
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure
> contains a definitive record of customers, application performance,
> security threats, fraudulent activity, and more. Splunk takes this
> data and makes sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-novd2d
> _______________________________________________
> gumstix-users mailing list
> gum...@li...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users