Re: USB problem. 'irq 9: nobody cared!'

Re: USB problem. 'irq 9: nobody cared!'

[Milton Miller]

Yaroslav Halchenko wrote:
> Greg KH wrote:
> > Did you try David Brownell's patch for this issue?
> Can you please point which one exactly? I've tried to locate patch you
> meant but it is too much of USB staff is happening now seems to me.


I'm guessing this one: Re: irq 11: nobody cared! is back

http://marc.theaimsgroup.com/?l=linux-kernel&m=106399942523614&w=2


milton

Re: USB problem. 'irq 9: nobody cared!'

[Yaroslav Halchenko]

Nop - it didn't help  :-(

http://onerussian.com/Linux/bug.USB2/dmesg

which else usefull information I can provide?
--Yarik

On Wed, Sep 24, 2003 at 05:57:27PM -0500, Milton Miller wrote:
> 
> Yaroslav Halchenko wrote:
> > Greg KH wrote:
> > > Did you try David Brownell's patch for this issue?
> > Can you please point which one exactly? I've tried to locate patch you
> > meant but it is too much of USB staff is happening now seems to me.
> 
> 
> I'm guessing this one: Re: irq 11: nobody cared! is back
> 
> http://marc.theaimsgroup.com/?l=linux-kernel&m=106399942523614&w=2
> 
> 
> milton
                                  .-.
=------------------------------   /v\  ----------------------------=
Keep in touch                    // \\     (yoh@|www.)onerussian.com
Yaroslav Halchenko              /(   )\               ICQ#: 60653192
                   Linux User    ^^-^^    [175555]

Re: USB problem. 'irq 9: nobody cared!'

[Greg KH]

On Thu, Sep 25, 2003 at 12:23:26AM -0400, Yaroslav Halchenko wrote:
> Nop - it didn't help  :-(
> 
> http://onerussian.com/Linux/bug.USB2/dmesg
> 
> which else usefull information I can provide?

David, can you try to fix this up.  It all started with your uhci
patch...

thanks,

greg k-h

Re: USB problem. 'irq 9: nobody cared!'

[David Brownell]

[-- Attachment #1: Type: text/plain, Size: 888 bytes --]

Greg KH wrote:
> On Thu, Sep 25, 2003 at 12:23:26AM -0400, Yaroslav Halchenko wrote:
> 
>>Nop - it didn't help  :-(
>>
>>http://onerussian.com/Linux/bug.USB2/dmesg
>>
>>which else usefull information I can provide?
> 
> 
> David, can you try to fix this up.  It all started with your uhci
> patch...

I have tried, two or three times now ...

The problem is that nobody has ever reported back with results from
testing any updated patch (see attachment, the guts of this being
from Alan Stern).  Sort of makes trying be a moot point ... :)

It's OK with me if you just revert the patch that adds a uhci_reset()
entry, but based on what I saw with EHCI and OHCI that'll just turn
up a different set of problems with certain BIOS configurations (none
of which I have) ... which will need to be fixed by having a UHCI
reset sequence that works correctly from _all_ initial states.

- Dave



[-- Attachment #2: Diff --]
[-- Type: text/plain, Size: 729 bytes --]

--- 1.44/drivers/usb/host/uhci-hcd.c	Fri Jul 18 06:22:32 2003
+++ edited/drivers/usb/host/uhci-hcd.c	Fri Sep 19 12:23:54 2003
@@ -1960,8 +1960,9 @@
 {
 	unsigned int io_addr = uhci->io_addr;
 
-	/* Global reset for 50ms */
+	/* Global reset for 50ms, and don't interrupt me */
 	uhci->state = UHCI_RESET;
+	outw(0, io_addr + USBINTR);
 	outw(USBCMD_GRESET, io_addr + USBCMD);
 	set_current_state(TASK_UNINTERRUPTIBLE);
 	schedule_timeout((HZ*50+999) / 1000);
@@ -2187,6 +2188,7 @@
 	/* Maybe kick BIOS off this hardware.  Then reset, so we won't get
 	 * interrupts from any previous setup.
 	 */
+	outw(0, uhci->io_addr + USBINTR);
 	pci_write_config_word(hcd->pdev, USBLEGSUP, USBLEGSUP_DEFAULT);
 	reset_hc(uhci);
 	return 0;

Re: USB problem. 'irq 9: nobody cared!'

[Yaroslav Halchenko]

I've tried again with recent bk12 kernel and this patch worked... 

May be I was too tired and actually didn't patch the first time I
tried, and then reported to you that it didn't help - I'm sorry. Will
check later if it is the case.

Thanx for your help

Sincerely 
Yarik

On Wed, Sep 24, 2003 at 05:57:27PM -0500, Milton Miller wrote:
> 
> Yaroslav Halchenko wrote:
> > Greg KH wrote:
> > > Did you try David Brownell's patch for this issue?
> > Can you please point which one exactly? I've tried to locate patch you
> > meant but it is too much of USB staff is happening now seems to me.
> 
> 
> I'm guessing this one: Re: irq 11: nobody cared! is back
> 
> http://marc.theaimsgroup.com/?l=linux-kernel&m=106399942523614&w=2
> 
> 
> milton
                                  .-.
=------------------------------   /v\  ----------------------------=
Keep in touch                    // \\     (yoh@|www.)onerussian.com
Yaroslav Halchenko              /(   )\               ICQ#: 60653192
                   Linux User    ^^-^^    [175555]

Re: USB problem. 'irq 9: nobody cared!' (FIXED!)

[Arkadiusz Miskiewicz]

On Thursday 25 of September 2003 21:11, David Brownell wrote:

> The problem is that nobody has ever reported back with results from
> testing any updated patch (see attachment, the guts of this being
> from Alan Stern).  Sort of makes trying be a moot point ... :)
So I'm the first one? ;)

Your patch (http://www.ussg.iu.edu/hypermail/linux/kernel/0309.3/0341.html) 
FIXES things for me :-)

This time I've tried 2.6.0test6 + all acpi patches from 2.6.0-mm1 (ls *acpi* 
in akpm broken-out directory) + your patch and things work fine :)
Note that 2.6.0+all acpi patches from 2.6.0-mm1 without your patch doesn't 
work.

[arekm@mobarm arekm]$ cat /proc/interrupts
           CPU0
  0:     698946          XT-PIC  timer
  1:       1712          XT-PIC  i8042
  2:          0          XT-PIC  cascade
  3:       1561          XT-PIC  irda0
  5:          4          XT-PIC  yenta
  8:          0          XT-PIC  rtc
  9:       2657          XT-PIC  acpi, eth0
 10:      77708          XT-PIC  VIA686A, uhci-hcd, uhci-hcd
 12:         18          XT-PIC  i8042
 14:      12408          XT-PIC  ide0
 15:         20          XT-PIC  ide1
NMI:          0
ERR:          0

More details about my case in lkml archives and 
http://bugme.osdl.org/show_bug.cgi?id=905

> It's OK with me if you just revert the patch that adds a uhci_reset()
> entry, but based on what I saw with EHCI and OHCI that'll just turn
> up a different set of problems with certain BIOS configurations (none
> of which I have) ... which will need to be fixed by having a UHCI
> reset sequence that works correctly from _all_ initial states.
>
> - Dave

-- 
Arkadiusz Miśkiewicz    CS at FoE, Wroclaw University of Technology
arekm.pld-linux.org AM2-6BONE, 1024/3DB19BBD, arekm(at)ircnet, PLD/Linux

Re: USB problem. 'irq 9: nobody cared!' (FIXED!)

[David Brownell]

Arkadiusz Miskiewicz wrote:
> On Thursday 25 of September 2003 21:11, David Brownell wrote:
> 
> 
>>The problem is that nobody has ever reported back with results from
>>testing any updated patch (see attachment, the guts of this being
>>from Alan Stern).  Sort of makes trying be a moot point ... :)
> 
> So I'm the first one? ;)

Since then I got mixed reports -- not sure I trust them.


> Your patch (http://www.ussg.iu.edu/hypermail/linux/kernel/0309.3/0341.html) 
> FIXES things for me :-)

OK, that's good to hear.  There are two patches that seem
to work to various degrees.  This isn't the least conservative
fix (that'd be http://lkml.org/lkml/2003/9/29/74, returning
to that odd "reset before kicking bios off" sequence) but it's
good to know this also behaves.

- Dave



> This time I've tried 2.6.0test6 + all acpi patches from 2.6.0-mm1 (ls *acpi* 
> in akpm broken-out directory) + your patch and things work fine :)
> Note that 2.6.0+all acpi patches from 2.6.0-mm1 without your patch doesn't 
> work.
> 
> [arekm@mobarm arekm]$ cat /proc/interrupts
>            CPU0
>   0:     698946          XT-PIC  timer
>   1:       1712          XT-PIC  i8042
>   2:          0          XT-PIC  cascade
>   3:       1561          XT-PIC  irda0
>   5:          4          XT-PIC  yenta
>   8:          0          XT-PIC  rtc
>   9:       2657          XT-PIC  acpi, eth0
>  10:      77708          XT-PIC  VIA686A, uhci-hcd, uhci-hcd
>  12:         18          XT-PIC  i8042
>  14:      12408          XT-PIC  ide0
>  15:         20          XT-PIC  ide1
> NMI:          0
> ERR:          0
> 
> More details about my case in lkml archives and 
> http://bugme.osdl.org/show_bug.cgi?id=905
> 
> 
>>It's OK with me if you just revert the patch that adds a uhci_reset()
>>entry, but based on what I saw with EHCI and OHCI that'll just turn
>>up a different set of problems with certain BIOS configurations (none
>>of which I have) ... which will need to be fixed by having a UHCI
>>reset sequence that works correctly from _all_ initial states.
>>
>>- Dave
> 
>