-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2008-10-13, Stuart Henderson <[EMAIL PROTECTED]> wrote:
>>> There are not "issues with re(4) which are being worked on" which >>> pertain to this issue. The problem mentioned by the original poster >>> is an issue with the (emulated) RTL8139 driver provided by QEMU >>> which KVM is derived from. The driver does not fully emulate the hw >>> by default unless compiled with a particular define, in this case it >>> is the timer interrupt. This needs to be fixed upstream with QEMU >>> and then also bug the KVM developers to copy the change over to >>> their source tree. >> I was looking for the bug report in Qemu but I could not find it. You >> have some link to be able to have a reference and track its state? > No, but there's a nice post about it here. > http://forums.citrix.com/thread.jspa?messageID=1343896&tstart=0 > OpenBSD and FreeBSD both have this defined in their respective ports > trees. > http://www.openbsd.org/cgi-bin/cvsweb/ports/emulators/qemu/patches/patch-hw_rtl8139_c?rev=1.4 > http://www.freebsd.org/cgi/cvsweb.cgi/ports/emulators/qemu/files/rtl8139-re-patch?rev=1.1 >> It's peculiar that testing both Linux distributions and Windows XP, I >> did not have this problem using the same network card. > Not really. Not all drivers for the same hardware work the same way. Thanks for the reply and the links, Stuart. I found temporarily a workaround using model=ne2k_pci with kvm. Regards, Daniel iEYEARECAAYFAkj3ChIACgkQZpa/GxTmHTd84ACeLnASR8uKIGzlDoxq/0OBx7W9 kb4AnA2SB4wtuaHcqDv9uUscL+aiJ9A2 =+n1O -----END PGP SIGNATURE-----
