On Tue, Nov 05, 2013 at 09:02:22AM +0200, Konstantin Belousov wrote: K> First, all reported instances have ata attachment for the ada0, except K> of milu (possibly). So this means that kernel does transient remapping, K> and very different code path is executed comparing with what I thought K> initially. The path is simpler than the pure unmapped i/o. K> K> Second, I use QEMU with the ata0 attachment for the disks regularly, and K> I do not have an issue. I also did not see a report from the real h/w. K> K> Third point is that all reports are i386. Is there anybody with amd64, K> vbox, ata and the same corruption hiddent by disabling unmapped i/o ? K> K> In what way the non-working images were installed ? Is it possible to K> produce the problematic installs by doing it one way, and non-problematic K> by another ? The only known way to get FreeBSD working is to turn off unmapped io in loader. Any tweaking of "hardware" in VB or changes in the install process do not affect. -- Totus tuus, Glebius.Received on Tue Nov 05 2013 - 06:06:15 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:44 UTC