Bugzilla is just used to track bugs - with something like this where there is a good chance of someone being able to fix it I'd also suggest posting a copy of the report to netdev@vger.kernel.org (you don't need to be subscribed)

i have already posted that to netdev, but as there was no response i thought it would be better to put it here so it won`t get lost. pardon, forgot to mention that - here is the reference http://marc.info/?l=linux-netdev&m=135574795027455&w=2

no response from netdev so far, so putting it here for reference. maybe someone has a clue about the proper fix.


> Gesendet: Mittwoch, 12. November 2014 um 22:46 Uhr
> Von: devzero@web.de
> An: netdev@vger.kernel.org
> Cc: thockin@hockin.org
> Betreff: re: bug? mac 00:00:00:00:00:00 with natsemi DP83815 after driver
> load
>
> hi,
> as i`m doing a little project with this older devices, i have come across
> this issue again and had some fun to dig deeper into it.
> 
> it`s a little bit academic, as i can do ifconfig eth0 hw ether..... as a
> workaround, but it sucks to hack that into startup scripts and i also have
> seen udev not playing nicely with it.
> 
> apparently the problem is being caused by a timing issue in the natsemi
> driver.
> 
> i added some debug printk`s in natsemi.c -> eeprom_read() after each
> occurrence of eeprom_delay(ee_addr); , and the problem went away. 
> 
> there is a hint about timing sensitivity in the code:
> 
> /* Delay between EEPROM clock transitions.
>    No extra delay is needed with 33Mhz PCI, but future 66Mhz access may need
>    a delay.  Note that pre-2.0.34 kernels had a cache-alignment bug that
>    made udelay() unreliable.
>    The old method of using an ISA access as a delay, __SLOW_DOWN_IO__, is
>    deprecated.
> */
> 
> looking at the source of natsemi-diag.c made me wonder why that utility is
> using 
> 
> #define eeprom_delay(ee_addr)   inl(ee_addr)
> 
> instead of 
> 
> #define eeprom_delay(ee_addr)   readl(ee_addr)
> 
> and apparently, that also fixes the problem (but gives a compile warning):
> 
> drivers/net/ethernet/natsemi/natsemi.c: In function âeeprom_readâ:
> drivers/net/ethernet/natsemi/natsemi.c:1019:3: warning: passing argument 1 of
> âinlâ makes integer from pointer without a cast [enabled by default]
> In file included from include/linux/io.h:22:0,
>                  from include/linux/pci.h:54,
>                  from drivers/net/ethernet/natsemi/natsemi.c:38:
> 
> 
> looking at a more recent version of natsemi-diag.c ,  i even found this one:
> 
> ftp://ftp.gwdg.de/pub/linux/misc/donald.becker/diag/natsemi-diag.c
> 
> /* Delay between EEPROM clock transitions.
>    This flushes the write buffer to prevent quick double-writes.
> */
> #define eeprom_delay(ee_addr) inl(ee_addr); inl(ee_addr)
> 
> The question is how to make a proper fix, as i don`t know what to pass to
> inl() , as it seems it should not get an mmapped adress but an i/o port
> instead !?
> 
> "The  in*() functions return data read from the specified I/O port"
> 
> "The read*() functions read data from device memory previously mapped by
> map_memory()"
> 
> regards
> roland
> 
> ps: CC driver maintainer from Kernel Maintainers file.
>

also sent to netdev ML.

This one should fix Bugzilla #51791 (details below).

Natsemi driver does not read MAC correctly from eeprom, while natsemi-diag from nictools-pci does. Apparently, tt`s a timing issue in the kernel driver.

According to ftp://ftp.gwdg.de/pub/linux/misc/donald.becker/diag/natsemi-diag.c , eeprom_delay(ee_addr) is defined as follows:

/* Delay between EEPROM clock transitions.
This flushes the write buffer to prevent quick double-writes.
*/
#define eeprom_delay(ee_addr) inl(ee_addr); inl(ee_addr)

while in the natsemi linux kernel driver, the delay is done this way :

#define eeprom_delay(ee_addr) readl(ee_addr)

, which results in the MAC being all zero`s.

So i simply added a second readl() to increase delay (instead of turning into inl() as proposed before). This may look a little bit ugly, but it`s fixing the problem for me.

I´m not sure how many natsemi users being left on this planet (probably few), but i guess this change does not do any harm on platforms where the driver does not behave buggy, so please consider adding it to mainline/stable/longterm.

Signed-off-by: Roland Kletzing <devzero@web.de>

---
drivers/net/ethernet/natsemi/natsemi.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/natsemi/natsemi.c b/drivers/net/ethernet/natsemi/natsemi.c
index b83f7c0..246bb91 100644
--- a/drivers/net/ethernet/natsemi/natsemi.c
+++ b/drivers/net/ethernet/natsemi/natsemi.c
@@ -987,7 +987,7 @@ static int natsemi_probe1(struct pci_dev *pdev, const struct pci_device_id *ent)
The old method of using an ISA access as a delay, __SLOW_DOWN_IO__, is
deprecated.
*/
-#define eeprom_delay(ee_addr) readl(ee_addr)
+#define eeprom_delay(ee_addr) readl(ee_addr); readl(ee_addr)

#define EE_Write0 (EE_ChipSelect)
#define EE_Write1 (EE_ChipSelect | EE_DataIn)
--