Messages in this thread | | | From | Alan Cox <> | Subject | Re: As 2.0 looms | Date | Mon, 29 Apr 1996 11:40:43 +0100 (BST) |
| |
> > Yes. And the 'Do you want to be offered ALPHA test drivers' needs to be > > moved to the top of ./drivers/net/Config.in so that it can be applied to > > *all* drivers, and not just ether drivers as it is at present. > Seconded!
Done and sent to Linus. Put the drivers in alphabetical order by section, made "10 and 100Mbit ethernet" as well as "Radio Interfaces" sections.
> There's also an EtherExpressPro driver which still has problems last time I > looked -- the transmitter can get stuck if the receiver has too much work > to do (setting the card to promiscuous mode should do it), and it fails to > properly unregister itself when loaded as a module.
I'll glance at that, probably a 30 second fix in with other bits to sort its unregister. The crash is someone elses problem ;)
> The network and standard serial drivers also carry their own IRQ-probe code > around with them, I'd really like to see that cleared up before 2.0 too. > (Does _anybody_ out there use the serial driver's "wild interrupt" > feature? And why is that thing in the serial driver in the first place? > Should be a kernel command line option instead...)
I use it to save setting parameters specifically. Id not be keen to change the irq probes until after 2.0 - when it should be done. IRQ probing is a delicate art.
As to printk. Does a preprocessor guru want to hack the defines so a printk without a KERN_xxx define at the start breaks during the compile.
Alan
| |