3c59x.c Linux driver patch

3c59x.c Linux driver patch

[Angelo Archie Amoruso]

Hi All

I'm sending you a diff file for a patch to the Linux
3c59x.c driver (from the 2.4.18 vanilla source tree), 
in order to support some strange 3C905-B
cards which map themselves on PCI device 00B7:9055.

Here come's the diff output:
============================

--- 3c59x.c.old Thu May 23 18:22:47 2002
+++ 3c59x.c     Thu May 23 21:23:14 2002
@@ -166,6 +166,10 @@
     - Rename wait_for_completion() to issue_and_wait() to avoid completion.h
       clash.

+    LK1.1.17 23 May 2002 Angelo Archie Amoruso (aamoruso@libero.it)
+    - Added entry on the vortex_pci_tbl[] in order to support 3c905b cards
+      which map to PCI device 00B7:9055
+
     - See http://www.uow.edu.au/~andrewm/linux/#3c59x-2.3 for more details.
     - Also see Documentation/networking/vortex.txt
 */
@@ -181,8 +185,8 @@


 #define DRV_NAME       "3c59x"
-#define DRV_VERSION    "LK1.1.16"
-#define DRV_RELDATE    "19 July 2001"
+#define DRV_VERSION    "LK1.1.17"
+#define DRV_RELDATE    "22 May 2002"



@@ -554,6 +558,7 @@
        { 0x10B7, 0x9050, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905_1 },
        { 0x10B7, 0x9051, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905_2 },
        { 0x10B7, 0x9055, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_1 },
+        { 0x00B7, 0x9055, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_1 },

        { 0x10B7, 0x9058, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_2 },
        { 0x10B7, 0x905A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_FX },


-- 

       /|
      /=|rchie 

Re: 3c59x.c Linux driver patch

[Andrew Morton]

Angelo Archie Amoruso wrote:
> 
> ...
>         { 0x10B7, 0x9050, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905_1 },
>         { 0x10B7, 0x9051, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905_2 },
>         { 0x10B7, 0x9055, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_1 },
> +        { 0x00B7, 0x9055, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CH_3C905B_1 },
> 

erm..  0x10B7 is 3Com.  I suspect you've dropped a bit in the
device's EEPROM and it's loading the wrong manufacturer ID
into the PCI config registers.

Grab vortex-diag from http://www.scyld.com/diag/index.html and
run `vortex-diag -aaee'.  Look at the EEPROM section.  I get:

EEPROM contents (64 words, offset 0x30):
 0x000: 2978 6056 0040 8060 0040 0000 0000 0080
 0x008: 0000 0000 0000 0000 0000 0000 0000 0000
 0x010: 0000 0000 0000 0000 0000 0000 10b7 6356
 0x018: 0000 0000 0000 0000 0000 0000 0000 0a0a
 0x020: ff29 2829 0008 0000 0000 0000 0000 0000
 0x028: 0000 ff01 0000 0000 0000 0000 0000 0000
 0x030: 0000 8642 002e 6056 0061 0009 0000 6d50
 0x038: 2970 0009 0000 8642 002e 0010 0000 00aa

See the 10b7 at offset 0x016?  That's the manufacturer ID.

You _might_ be able to fix it with 3com's DOS-based setup
tool ftp://ftp.3com.com/pub/nic/3c90x/3c90xx2.exe or you
can just run with your patch and be happy ;)

-

Re: 3c59x.c Linux driver patch

[Alan Cox]

On Thu, 2002-05-30 at 22:03, Angelo Archie Amoruso wrote: 
> I'm sending you a diff file for a patch to the Linux
> 3c59x.c driver (from the 2.4.18 vanilla source tree), 
> in order to support some strange 3C905-B
> cards which map themselves on PCI device 00B7:9055.

That sounds like the eeprom is faulty on your card or the card isn't in
the slot properly. 

Alan