On Sat 8 Aug 2009 05:50, Ben Warren pondered:
> Allesandro,
>
> Alessandro Rubini wrote:
> > I finally fixed the defrag code, testing with NFS as well.
> > Didn't take performance figures, tough, for lack of time.
> >
> > I wanted to do "config + environment" for the NFS case, like tftp, but
> > didnt' do the second step out of laziness (also, the source file has
> > long lines while I have 80 columns).
> >
> > For the record, I added the check on ip_src and ip_dst, but everything
> > stopped working. So I reverted that instead of entering a long
> > debugging session.
> >
> > The CONFIG_NET_MAXDEFRAG argument is the actual payload, so I add NFS
> > overhead to that figure, which is expected to a beautiful 4096 or
> > 8192. I feel this is better than other options, as the person writing
> > the config is not expected to know how much protocol overhead is
> > there.
> >
> > Alessandro Rubini (4):
> > net: defragment IP packets
> > tftp: get the tftp block size from config file and from the
> > environment
I noticed that while playing with this, that if you set the
"tftpblocksize" environment, do a transfer, and then clear it, it
does not go back to the default setting. I was not sure if this was
the intended or not, but this fixes it (and provides a small code size
reduction when this option is not activated).
Also wondering -- if the user sets the "tftpblocksize" to a number larger
than IP_MAXUDP, the transfer will never finish. Should this be restricted
here?
diff --git a/net/tftp.c b/net/tftp.c
index 9544691..56db247 100644
--- a/net/tftp.c
+++ b/net/tftp.c
+#ifdef CONFIG_TFTP_BLOCKSIZE
/* Allow the user to choose tftpblocksize */
if ((ep = getenv("tftpblocksize")) != NULL)
TftpBlkSizeOption = simple_strtol(ep, NULL, 10);
+ else
+ TftpBlkSizeOption = TFTP_MTU_BLOCKSIZE;
debug("tftp block size is %i\n", TftpBlkSizeOption);
+#endif
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
