Problems with LinkSys 10/100 cards
Steve Huang
steve@ltnlcc.com.tw
Thu Feb 4 03:54:32 1999
> -----Original Message-----
> From: Brian Denheyer [mailto:briand@deldotd.com]
> Sent: Thursday, February 04, 1999 3:28 PM
> To: steve@ltnlcc.com.tw
> Cc: doo@shroom.com; linux-tulip@cesdis1.gsfc.nasa.gov
> Subject: Re: Problems with LinkSys 10/100 cards
>
>
> >>>>> "Steve" == Steve Huang <steve@ltnlcc.com.tw> writes:
>
> Steve> STOP transmit Interrupt occured and show the
> message. We can remove the
> Steve> Stop Tx code in the pnic_timer routine to reduce
> the message show.
>
> Steve> 2058a2059
> Steve> outl(tp->csr6 | 0x0002, ioaddr +
> CSR6); /* Restart Tx
> Steve> */
> Steve> outl(tp->csr6 | 0x2002, ioaddr +
> CSR6); (orginal)
> Steve> ---
> Steve> outl(tp->csr6 | 0x2002, ioaddr +
> CSR6); (new)
>
>
> This patch was a little confusing, here is what I used for the code :
>
> tp->csr6 = new_csr6;
> outl(tp->csr6 | 0x2002, ioaddr + CSR6);
> dev->trans_start = jiffies;
>
> Is this what you intended ?
>
YES. The system will never display the warning message while system
boot up or
media changed.
> I installed the rest of the patches for the different registers.
> There is no performance improvement. I STILL see performance of about
> 1Mbyte/s. I just can't seem to do any better, it's almost as if the
> chip keeps configuring for 10base instead of 100base.
Do you modified the following lines in tulip.c(v.90f)?
1372c1372
< tp->csr6 = 0x814C0000 | (tp->full_duplex ? 0x0200 : 0);
(new)
---
> tp->csr6 = 0x816C0000 | (tp->full_duplex ? 0x0200 : 0);
(original)
1542c1542
< new_csr6 = 0x810C0000; (new)
---
> new_csr6 = 0x812C0000; (original)
2004c2004
< new_csr6 |= 0x810C0000; (new)
---
> new_csr6 |= 0x812E0000; (original)
2006c2006
< new_csr6 |= 0x814C0000; (new)
---
> new_csr6 |= 0x816E0000; (original)
Could you tell me how do you measure the performance and your testing
environment ?
I would like to simulate it. Thanks.
Steve
>
> I also tried running one of my boxes under win98 so that I was using
> the "official" linksys driver and still do NOT see anything better
> than 1.5Mbyte/s.
>
> ??
>
>
> Brian
>