L.S.
Mmm.... I might be onto something about this weird MTU problem with Zydas
1211b based devices on my embedded box.
This box has an RDC i486 SOC processor.
Within the vendor driver of the Zydas 1211b there is an extra compiler flag
called "ZDCONF_R8610_FRAG_FIX" (r8610 is an RDC i486 compatible SOC), it is
only tested in the vendor module zd1205c as follows:
(small snippet of code, arrow is mine)
//copy packet for IP header is located on 4-bytes alignment
#if ZDCONF_R8610_FRAG_FIX == 1
<<<<<<<<<<<<<<<<<<-------------------
if (1)
#else
if (length < RX_COPY_BREAK)
#endif
{
dev_kfree_skb_any(skb);
skb = dev_alloc_skb(length+2);
if (skb){
skb->dev = g_dev;
skb_reserve(skb, 2);
skb_copy_from_linear_data(skb,pData,length);
skb_put(skb, length);
}
}
else{
#ifdef NET_SKBUFF_DATA_USES_OFFSET
skb->tail = 0;
skb->data = pData;
#else
skb->tail = skb->data = pData;
#endif
skb_put(skb, length);
}
i.e. if the flag is set in the receiver code a number of statements is
always executed on an R8610...
I reckon this certainly might be the reason of my MTU > 500 problem... seems
plausible.
This weird patch is of course not present in the community driver.
Any comments?
TIA
Regards,
Ernst
----- Original Message -----
From: "Hin-Tak Leung" <[email protected]>
To: <[email protected]>; "Ernst J. Oud"
<[email protected]>
Sent: Sunday, February 14, 2010 11:41 PM
Subject: Re: [zd1211-devs] Problem with MTU > 500
Hmm, what's the hardware (the zd1211rw: ... line in your dmesg, you can
probably just grep zd1211 /var/log/message for it as well) may be all the
same. There are quite a few variants.
I don't know if putting compat-wireless on will help, but you could give
that a try. FWIW, the driver development has gone over to
linux-wireless.org and changes goes directly into wireless-testing git; the
sourceforge zd1211 site isn't much more than the mailing list these days,
and even the mailing list isn't very active. (try the linux-wireless list
may be a good idea).
--- On Sun, 14/2/10, Ernst J. Oud <[email protected]> wrote:
> Ok.
>
> Did some further experimenting. I remembered having an
> Thomson st121g USB WLAN stick and tried that (Prism
> chipset). With the same distro and config it works fine
> using minstrel or pid; i.e. with an MTU of 1500 no problems
> whatsoever. If I bring the box down and use my zd1211 stick
> instead (some distro/config) and boot it, it doesn't work
> with any MTU > 500 using pid or minstrel.
>
> So this must be some zd1211rw driver issue. It isn't the
> slackware setup and not the hardware; i.e. not the embedded
> box since the st121g works fine on it and not the zd1211
> stick since I used three separate sticks (from three
> different manufacturers: icidu, msi and cc&c) and they
> all have the same problem.
>
> Any further clues?
>
> BTW; the svn for the zd1211rw driver shows that the latest
> version is 3 years old?
>
> TIA
>
> Ernst
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Zd1211-devs mailing list - http://zd1211.wiki.sourceforge.net/
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/zd1211-devs
