On Sat, Aug 29, 2015 at 1:05 AM, Matwey V. Kornilov <[email protected]> wrote: > 2015-08-29 1:00 GMT+03:00 Misha Komarovskiy <[email protected]>: >> Hello Matwey, >> >> On Sat, Aug 29, 2015 at 12:50 AM, Matwey V. Kornilov >> <[email protected]> wrote: >>> 2015-08-28 18:55 GMT+03:00 Misha Komarovskiy <[email protected]>: >>>> Hello Andreas, >>>> >>>> On Fri, Aug 28, 2015 at 6:36 PM, Andreas Färber <[email protected]> wrote: >>>>> Hi, >>>>> >>>>> Am 28.08.2015 um 09:22 schrieb Misha Komarovskiy: >>>>>> On Mon, Aug 24, 2015 at 8:01 PM, Matwey V. Kornilov >>>>>> <[email protected] <mailto:[email protected]>> wrote: >>>>>> >>>>>> Still no luck. >>>>>> >>>>>> >>>>>> I noticed that current default and vanilla configs contain some platform >>>>>> specific defines: >>>>>> >>>>>> grep -ni debug_ll * >>>>>> default:7724:# CONFIG_DEBUG_LL is not set >>>>>> default:7727:CONFIG_DEBUG_LL_INCLUDE="debug/bcm63xx.S" >>>>>> lpae:7335:# CONFIG_DEBUG_LL is not set >>>>>> lpae:7337:CONFIG_DEBUG_LL_INCLUDE="mach/debug-macro.S" >>>>>> vanilla:7719:# CONFIG_DEBUG_LL is not set >>>>>> vanilla:7722:CONFIG_DEBUG_LL_INCLUDE="debug/bcm63xx.S" >>>>>> >>>>>> maybe wrong debug macro may break some platform drivers? >>>>> >>>>> As you can see, CONFIG_DEBUG_LL is not enabled, so you can ignore >>>>> CONFIG_DEBUG_LL_* defaults. lpae simply doesn't enable bcm63xx. >>>>> >>>> >>>> I agree, but SoC bcm63xx support seems to forcibly enable alot of >>>> various defines even with DEBUG_LL disabled, maybe try to test with >>>> bcm63xx disabled? >>> >>> I am rebuilding kernel withot bcm63xx. I have no any other ideas, so lets >>> see. >> >> In case if it'll help, i already found fix which can make bcm63xx SoC >> support compatible with multiplatform >> config, without disabling itself: >> >> diff --git a/arch/arm/Kconfig.debug b/arch/arm/Kconfig.debug >> index 0c12ffb..304ca62 100644 >> --- a/arch/arm/Kconfig.debug >> +++ b/arch/arm/Kconfig.debug >> @@ -1322,7 +1322,7 @@ config DEBUG_UART_8250 >> >> # Compatibility options for BCM63xx >> config DEBUG_UART_BCM63XX >> - def_bool ARCH_BCM_63XX >> + bool >> >> config DEBUG_UART_PHYS >> hex "Physical base address of debug UART" >> > > Great, could you please send this patch to upstream? I also think that > def_bool should be replaced with 'depends on ARCH_BCM_63XX'
DEBUG_BCM63XX depends on ARCH_BCM_63XX and select DEBUG_UART_BCM63XX. Upstream does not include yet bcm63xx into multiplatform config, i'm not sure why, maybe some of this SoC requirements not fully migrated to device tree from platform data. Please test your serial output first, im not sure it is right root of any issue at all. > >>> >>>> >>>> >>>>> Regards, >>>>> Andreas >>>>> >>>>> P.S. No need for HTML here. :) >>>>> >>>>> -- >>>>> SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany >>>>> GF: Felix Imendörffer, Jane Smithard, Graham Norton; HRB 21284 (AG >>>>> Nürnberg) >>>> >>>> >>>> >>>> -- >>>> Best Regards, >>>> Misha Komarovskiy >>>> zombahatgmaildotcom >>> >>> >>> >>> -- >>> With best regards, >>> Matwey V. Kornilov >>> http://blog.matwey.name >>> xmpp://[email protected] >> >> >> >> -- >> Best Regards, >> Misha Komarovskiy >> zombahatgmaildotcom > > > > -- > With best regards, > Matwey V. Kornilov > http://blog.matwey.name > xmpp://[email protected] -- Best Regards, Misha Komarovskiy zombahatgmaildotcom -- To unsubscribe, e-mail: [email protected] To contact the owner, e-mail: [email protected]
