netdev - Re: r8169 rx descriptors

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20080615220410.GA10522@electric-eye.fr.zoreil.com>
Date:	Mon, 16 Jun 2008 00:04:10 +0200
From:	Francois Romieu <romieu@...zoreil.com>
To:	Denys Fedoryshchenko <denys@...p.net.lb>
Cc:	netdev@...r.kernel.org
Subject: Re: r8169 rx descriptors

Denys Fedoryshchenko <denys@...p.net.lb> :
[...]
> Why in Linux source:
> 
> #define NUM_TX_DESC     64      /* Number of Tx descriptor registers */
> #define NUM_RX_DESC     256     /* Number of Rx descriptor registers */

One do not necessarily gain much from a huge Rx / Tx ring. You can
increase NUM_RX_DESC though.

> Is it possible to do them changeable over ethtool -g ?

Yes.

> Maybe NetBSD developers signed NDA and have access to datasheets?

A NDA is not needed. You can get in touch with Edward Hsu
<edward_hsu@...ltek.com.tw> if you have some questions.

It can help for the moderately documented irq mitigation register.

> Also i seen in one patch from Realtek , something similar to
> -#define RX_DMA_BURST    6       /* Maximum PCI burst, '6' is 1024 */
> +#define RX_DMA_BURST    7       /* Maximum PCI burst, '6' is 1024 */
> 
> Why it is reverted or not applied?

I do not always apply Realtek's changes blindly. This one could make a
difference though.

> In sources from Realtek website it is 
> #define RX_DMA_BURST	7	/* Maximum PCI burst, '7' is unlimited */
> #define TX_DMA_BURST	7	/* Maximum PCI burst, '7' is unlimited */
> 
> I will try to change burst settings on my router, and waiting comments about 

Apparently 1) you are experiencing Rx fifo overflows 2) your mean packet
stands at 300 bytees. I'd try to lower RX_FIFO_THRESH first then lower
TX_DMA_BURST.

-- 
Ueimor
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html