RE: [PATCH] r8169: RxConfig hack for the 8168evl.

[hayeswang <hayeswang@realtek.com>]

> From: Francois Romieu [mailto:romieu@fr.zoreil.com] 
[...]
> > The definition of the IO 0x44 bit 14 is opposite for new chips.
> > For 8111C, 0 means fetching one Rx descriptor, and 1 means fetching
> > multi-descriptors.
> > For 8111D and the later chips, 0 means fetching 
> multi-descriptors, and 1 means
> > fetching one Rx descriptor.
> 
> Ok. Is there much point fetching one Rx descriptor versus several ?

I just know fetching several Rx descriptor has better performance if the
decsriptors are enough. However, I have no idea about how the hardware
implements it.

> 
> [...]
> > The CFG_METHOD_16 is the internal test chip. We don't have 
> mass production for
> > it. Even it could be removed from driver. I don't think the 
> kernel have to
> > support it.
> 
> Ok.
> 
> There seem to be a few differences for the CFG_METHOD_16 
> chipset between
> the kernel driver and Realtek's own. I have noticed the points below.
> Should some of those be included ?

I think you should reference the CFG_METHOD_17, not CFG_METHOD_16.

[...] 
> - 0x7cf00000 / 0x2c900000 is a Realtek internal, test-only chipset
Should be 0x7cf00000 / 0x2c800000

[...] 
> rtl8169_get_mac_version(struct rtl8169_private *tp,
>  		{ 0x7cf00000, 0x48000000,	RTL_GIGA_MAC_VER_35 },
>  
>  		/* 8168E family. */
> -		{ 0x7c800000, 0x2c800000,	RTL_GIGA_MAC_VER_34 },
> +		{ 0x7cf00000, 0x2c800000,	RTL_GIGA_MAC_VER_34 },

Should be	{ 0x7cf00000, 0x2c900000,	RTL_GIGA_MAC_VER_34 },

[...] 
> @@ -4797,6 +4826,9 @@ static void rtl_hw_start_8168e_2(struct 
> rtl8169_private *tp)
>  
>  	RTL_W8(MaxTxPacketSize, EarlySize);
>  
> +	RTL_W8(0xf2, (RTL_R8(0xf2) & ~0x02) | 0x05);

Please remove this line. We verify it would case kernel panic.

[...] 

 
Best Regards,
Hayes