Messages in this thread | | | Date | Sun, 19 Sep 1999 23:51:01 +0000 (UTC) | From | Adam Fritzler <> | Subject | Re: [linux-tr] SysKonnect driver broken in new kernels? |
| |
On Sun, 19 Sep 1999, Alan Cox wrote:
> So its something it doesnt do. Ok now the right question is > "What changed in ibmtr.c over the same sort of period".
I'll try to take a look, but I have _no_ idea where to begin.
> Do you know roughly which actual kernel broke sktr ?
No idea. My first contact with this driver at all was last week. Someone who's used the driver before might know better. What changes have occured in netif_rx() (and related) lately? When did sktr first make it into the mainstream kernels? It obviously worked then or it wouldn't have made it in (right? :).
> > The major thing that breaks is sktr_wait(). Replacing it with a udelay() > > I fixed that in my tree. It uses udelay short short stops and schedule > for long ones
Did it make it into 2.3.18ac6?
> > I'm not real sure what breaks recieve. I don't know enough about the > > Well netif_rx is supposed to pass a packet up to the higher layers. The > board seems to be claling this. It also sets skb->protocol before hand > which is correct. The only thing I can't see is who sets skb->dev. It might > be adding skb->dev = dev before the netif_rx makes life happy - try it can > you ?
No good. Anyway, skb->dev does get set on line 2570ish (rpl->Skb->dev = dev);
> > Also, would a patch get accepted that renamed the SysKonnect driver to > > something more like 'Generic TMS380'? (The TMS380 chipset is sortof the > > NE2000 of token ring -- for both ISA and PCI cards.) > > Interesting. I didnt know that. You have example non syskonnect cards it works > with ? Yes. The Compaq PCI TR card is what I'm using right now. I also have an old (1990ish) Intel ISA card that should work with this driver, and I have some Thomas-Conrad PCI cards on the way that should be identical to the Compaq cards. Anything that uses these chipsets should work, according to TI: - TMS380C16+TMS38053 [1st generation COMprocessor+Ring Interface] (as on ISA Intel card) - TMS380C20+TMS38053/4 [1st generation+Ring Interface] (yet to see one of these, but its standard according to TI) - TMS380C25+TMS38054 [2nd generation COMprocessor+Ring Interface] (as on ISA SysKonnect cards) - TMS380C25+TMS38054+TMS380PCIA [2nd generation COM+RI+PCI interface] (as on PCI SysKonnect cards) - TMS380C30+TMS380PCIA [2nd+ generation COMRI+PCI (same as C25+54 in one chip) (as on Compaq/Thomas-Conrad PCI cards)
Any of the 2nd generation chips support Dedicated (switched) token ring as well. The only thing the TMS380 doesn't do is 100Mbit, which I believe only Madge does anyway.
TI has been very open with information, including releasing the microcode/firmware out with no restrictions. (I need to talk to them more about this; but from what I can see, the strict warning at the top of sktr_firmware.h can be lightened up a bit. You can get a copy of the firmware just by emailing them and asking for it.)
In any case, the SysKonnect driver is a very generic driver and it should be represented as such. When I get it working with all my cards, I'll go through and do the name change and patches (including putting myself as the maintainer).
mid
--- Adam Fritzler { mid@auk.cx, afritz@iname.com} http://www.auk.cx/~mid/ "Something in my systray is blinking wildly." -- DS
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |