Am 27.09.2013 12:02, schrieb Paolo Bonzini: > Il 24/07/2012 09:49, christian schilling ha scritto: >> On Mon, Jul 23, 2012 at 6:28 PM, Stefan Weil <s...@weilnetz.de> wrote: >>> Am 23.07.2012 11:25, schrieb initcr...@gmail.com: >>>> A driver using simplified mode that works on real hardware >>>> did not work in qemu. >>>> >>>> Signed-off-by: Christian Schilling <initcr...@gmail.com> >>>> --- >>>> hw/eepro100.c | 7 +++++++ >>>> 1 files changed, 7 insertions(+), 0 deletions(-) >>>> >>> >>> Do you really think that's a trivial patch? >> It's only three lines plus comments, but ok small != trivial. >> >>> I have a different fix for simplified mode in my QEMU tree: >>> >>> http://repo.or.cz/w/qemu/ar7.git/blob/HEAD:/hw/eepro100.c >>> >>> That version is implemented according to the Intel specifications >>> and avoids hacks for specific guest drivers. >> My fix isn't a hack for a specific guest driver, but is also in >> accordance with the >> intel specs. >> >>> Maybe you can give it a try. >> I have, it does work. >> Overall the code look better to me, but one thing irritates me: >> The comment on line 821 contradicts the trace on 830 and 833. >> in fact i don't understand the code from 820 to 340. It seems to me it should >> handle extended flexible TCBs, but if it does what does the code >> following line 855 do? > Anything new about this one-year-old patch? > > Paolo
No, there isn't, sorry. I still have to send a patch which merges my modifications into official QEMU. And I also still wish a good set of test cases which can be used for all nic emulations. My version of eepro100.c is in http://repo.or.cz/w/qemu/ar7.git. Regards, Stefan
