Re: [PATCH for-7.2] rtl8139: honor large send MSS value

[Jason Wang <jasowang@redhat.com>]

在 2022/11/17 19:26, Tobias Fiebig 写道:
> Heho,
> Ok, that explains a lot. I was also thinking that the vlan bit seem to overlap with the MTU field, and wanted to look at that later today.
>
> Re the 12b: IIRC, the standard 1500 MTU for ethernet is already without the ethernet header; That can have up to 26b (18b basis, 4b 802.1q, 4b 802.1ad), but leads to a total frame length of 1526 (with other additions (MPLS) also just making the frame bigger, without touching the MTU/MSS). MSS than as usual -40 for v4 and ~ -60 for v6.
>
> So I doubt that those 12b are subtracted for the ethernet header.
>
> Does somebody still have an RTL8139 around, to test how the real hardware behaved?


This would be very hard, I can think that the Qemu rtl8139 emulation is 
probably the only user for kernel 8139cp drivers for years.

Thanks


>
> With best regards,
> Tobias
>
> -----Original Message-----
> From: Stefan Hajnoczi <stefanha@gmail.com>
> Sent: Thursday, 17 November 2022 12:16
> To: Tobias Fiebig <tobias@fiebig.nl>
> Cc: Jason Wang <jasowang@redhat.com>; Stefan Hajnoczi <stefanha@redhat.com>; qemu-devel@nongnu.org; qemu-stable@nongnu.org; Russell King - ARM Linux <linux@armlinux.org.uk>
> Subject: Re: [PATCH for-7.2] rtl8139: honor large send MSS value
>
> After looking more closely at txdw0 it seems that the code mixes "Tx command mode 0", "Tx command mode 1", and "Tx status mode". The bits have different meanings in each mode, so this leads to confusion :).
>
> Stefan
>