Re: 2.6.12 breaks 8139cp

Re: 2.6.12 breaks 8139cp

[Nick Warne]

> It seems that whatever is causing the bug i'm seeing was a change that 
> involved the network driver/interrupt subsystem and not actually the 
> 8139too driver, this was done at around 2.6.3-2.6.4.   A NAPI patch that 
> was introduced at that time basically reverted the interrupt function 
> and removed the poll functions from the driver (via not enabling napi in 
> Kconfig).

Ummm.  I was involved in this.  Here is the 'final' post after Mr Hirofumi 
found the cause of my issues:

http://www.ussg.iu.edu/hypermail/linux/kernel/0402.3/1709.html

But I looked at what he said and found the real problem on my system (after 
all that):

http://www.ussg.iu.edu/hypermail/linux/kernel/0403.1/1537.html

Nick
-- 
"When you're chewing on life's gristle,
Don't grumble, Give a whistle..."

8139too PCI IRQ issues WAS Re: 2.6.12 breaks 8139cp

[Ed Sweetman]

What are you supposed to do though if you dont have that bios option.  
And if the bios wasn't changed between kernel version upgrades, what is 
the PCI irq subsystem doing now that requires such a change? And what 
makes it possible to remove the problem with reverting just the 8139too 
driver ...   There is a quirk here, but the fix should be in the kernel, 
since not all bios' allow you to make the fix yourself. 

Nick Warne wrote:

>>It seems that whatever is causing the bug i'm seeing was a change that 
>>involved the network driver/interrupt subsystem and not actually the 
>>8139too driver, this was done at around 2.6.3-2.6.4.   A NAPI patch that 
>>was introduced at that time basically reverted the interrupt function 
>>and removed the poll functions from the driver (via not enabling napi in 
>>Kconfig).
>>    
>>
>
>Ummm.  I was involved in this.  Here is the 'final' post after Mr Hirofumi 
>found the cause of my issues:
>
>http://www.ussg.iu.edu/hypermail/linux/kernel/0402.3/1709.html
>
>But I looked at what he said and found the real problem on my system (after 
>all that):
>
>http://www.ussg.iu.edu/hypermail/linux/kernel/0403.1/1537.html
>
>Nick
>  
>

Re: 8139too PCI IRQ issues WAS Re: 2.6.12 breaks 8139cp

[Nick Warne]

On Sunday 26 June 2005 15:59, you wrote:
> What are you supposed to do though if you dont have that bios option.
> And if the bios wasn't changed between kernel version upgrades, what is
> the PCI irq subsystem doing now that requires such a change? And what
> makes it possible to remove the problem with reverting just the 8139too
> driver ...   There is a quirk here, but the fix should be in the kernel,
> since not all bios' allow you to make the fix yourself.

Well, I wasn't saying _this_ is the solution, I was just pointing out the 
threads involved from last year for reference during this issue.

Nick
-- 
"When you're chewing on life's gristle,
Don't grumble, Give a whistle..."

Re: 8139too PCI IRQ issues WAS Re: 2.6.12 breaks 8139cp

[OGAWA Hirofumi]

[-- Attachment #1: Type: text/plain, Size: 595 bytes --]

Ed Sweetman <safemode@comcast.net> writes:

> What are you supposed to do though if you dont have that bios option.
> And if the bios wasn't changed between kernel version upgrades, what
> is the PCI irq subsystem doing now that requires such a change? And
> what makes it possible to remove the problem with reverting just the
> 8139too driver ...   There is a quirk here, but the fix should be in
> the kernel, since not all bios' allow you to make the fix yourself. 

Can you post the detail? (/proc/interrupt, mptable, 8259A.pl, lspci -vvv)
-- 
OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 8259A.pl --]
[-- Type: text/x-perl, Size: 1145 bytes --]

#!/usr/bin/perl
#
# 8259A
#

use strict;
use warnings;

use Getopt::Std;

sub outb
{
    my $data = shift;
    my $port = shift;

    open(IO, "> /dev/port") || die;
    sysseek(IO, $port, 0) || die;
    my $x = pack("C", $data);
    syswrite(IO, $x, 1);
    close(IO);
}

sub inb
{
    my $port = shift;
    my $data;

    open(IO, "/dev/port") || die;
    sysseek(IO, $port, 0) || die;
    sysread(IO, $data, 1);
    close(IO);

    return unpack("C", $data);
}

sub set_edge_level
{
    my $irq = shift;
    my $to = shift;
    my $mask = 1 << ($irq & 7);
    my $port = 0x4d0 + ($irq >> 3);
    my $val = inb($port);

    if ($to eq "edge") {
	outb($val & ~$mask, $port);
    } else {
	outb($val | $mask, $port);
    }
    print "irq $irq: -> $to\n";
}

our($opt_e, $opt_l);
getopts('e:l:');

if ($opt_e and ($opt_e < 15)) {
    set_edge_level($opt_e, "edge");
} elsif ($opt_l and ($opt_l < 15)) {
    set_edge_level($opt_l, "level");
}

foreach my $irq (0..15) {
    my $mask = 1 << ($irq & 7);
    my $port = 0x4d0 + ($irq >> 3);
    my $val = inb($port);

    printf "irq %d: %02x, %s\n", $irq, $val, ($val & $mask) ? "level" : "edge";
}