Re: Panda: USB crash with today's linux-next

[Felipe Balbi <balbi@ti.com>]

[-- Attachment #1: Type: text/plain, Size: 4148 bytes --]

Hi,

On Mon, May 14, 2012 at 07:06:25PM +0200, Cousson, Benoit wrote:
> On 5/14/2012 2:47 PM, Felipe Balbi wrote:
> >Hi,
> >
> >On Mon, May 14, 2012 at 03:29:21PM +0300, Felipe Balbi wrote:
> >>On Mon, May 14, 2012 at 03:24:11PM +0300, Tomi Valkeinen wrote:
> >>>On Mon, 2012-05-14 at 15:15 +0300, Felipe Balbi wrote:
> >>>
> >>>>looks like MUSB is probing before transceiver driver... could it be ?
> >>>>Can you check transceiver has actually probed ? I guess panda's using
> >>>>twl6030-usb.c
> >>>
> >>>Ah. Perhaps it's then about this?
> >>>
> >>>[    0.352905] Skipping twl internal clock init and using bootloader value (unknown osc rate)
> >>>[    0.354034] twl 1-0048: PIH (irq 39) chaining IRQs 352..372
> >>>[    0.356079] VUSB: 3300 mV normal standby
> >>>[    0.358123] genirq: Threaded irq requested with handler=NULL and !ONESHOT for irq 356
> >>>[    0.358215] twl6030_usb twl6030_usb: can't get IRQ 356, err -22
> >>>[    0.358398] twl6030_usb: probe of twl6030_usb failed with error -22
> >>
> >>sounds about right. Now, why can't it get the IRQ... Benoit, is this
> >>related to your sparse irq/irq_domain changes ?
> >
> >Looks like twl6030-irq still missed conversion to threaded IRQ. It still
> >has that ugly ass kthread to handle the IRQs. Oh well, yet another
> >broken OMAP driver...
> 
> Well, yeah, I did not clean all that mess.
> 
> That being said, we did have some issue recently as well, but due to
> the increase of IRQ number and the fact the NR_IRQS is still used
> since SPARSE_IRQ migration is not completed.
> 
> At least we saw similar issue with OMAP5.
> 
> Maybe increasing NR_IRQS will fix that, but in this case, it looks
> like the IRQ might already been used by someone else.
> This is probably because something is still using the hard coded IRQ
> BASE number from the irqs.h define.
> 
> I was planning to get rid of them to highlight the broken driver /
> board that might still used them. But this is on my TODO list :-(

another thing you might want to add to your TODO list is implementing
irq_chip properly on twl6030-irq:

$ git grep -e EXPORT_SYMBOL drivers/mfd/twl6030-irq.c
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_interrupt_unmask);
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_interrupt_mask);
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_mmc_card_detect_config);
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_mmc_card_detect);

$ git grep -e "twl6030_interrupt_\(mask\|unmask\)" drivers/
drivers/mfd/twl6030-irq.c:int twl6030_interrupt_unmask(u8 bit_mask, u8 offset)
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_interrupt_unmask);
drivers/mfd/twl6030-irq.c:int twl6030_interrupt_mask(u8 bit_mask, u8 offset)
drivers/mfd/twl6030-irq.c:EXPORT_SYMBOL(twl6030_interrupt_mask);
drivers/mfd/twl6030-irq.c:	twl6030_interrupt_unmask(TWL6030_MMCDETECT_INT_MASK,
drivers/mfd/twl6030-irq.c:	twl6030_interrupt_unmask(TWL6030_MMCDETECT_INT_MASK,
drivers/rtc/rtc-twl.c:		twl6030_interrupt_unmask(TWL6030_RTC_INT_MASK,
drivers/rtc/rtc-twl.c:		twl6030_interrupt_unmask(TWL6030_RTC_INT_MASK,
drivers/rtc/rtc-twl.c:		twl6030_interrupt_mask(TWL6030_RTC_INT_MASK,
drivers/rtc/rtc-twl.c:		twl6030_interrupt_mask(TWL6030_RTC_INT_MASK,
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_unmask(0x05, REG_INT_MSK_LINE_C);
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_unmask(0x05, REG_INT_MSK_STS_C);
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_unmask(TWL6030_CHARGER_CTRL_INT_MASK,
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_unmask(TWL6030_CHARGER_CTRL_INT_MASK,
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_mask(TWL6030_USBOTG_INT_MASK,
drivers/usb/otg/twl6030-usb.c:	twl6030_interrupt_mask(TWL6030_USBOTG_INT_MASK,

That whole MMC card detection is also pretty screwed up. Balaji/Venkat,
can you guys look into that ? Probably making something generic using a
threaded IRQ handler ?

I mean, all the MMC core should need is an IRQ number (through GPIOs or
not doesn't/shouldn't matter) and it should be able to use a threaded
IRQ handler to kick the card detection/initialization.

-- 
balbi

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]