On Thu, Apr 17, 2008 at 07:13:17PM -0400, Garrett Wollman wrote: >In my experience, umass over EHCI has never worked on any machine >ever, going back to 5.x and over multiple kinds of umass devices. (I >never saw panics, only triple-fault CPU resets.)
OTOH, I've had mixed results, though I haven't seen panics caused by EHCI. umass has a tendency to panic when it trips over an interface bug between bus_dmamem_alloc(9) and contigmalloc(9) which has been present since at least 4.x, though the work-arounds have improved and this is less of a problem than it was 3 years ago. On my laptop (HP nx6125, ATI SB400 chipset), I haven't had any problems with EHCI in 6.x or 7.0 (both amd64 - I can't recall if I've tried ECHI whilst running i386). My son's laptop (HP V6107au, nVIDIA MCP51 chipset, 6.x/amd64) gives varying results (sometimes it works, sometimes it doesn't) and he's found that plugging flashdisks into the USB hub on his keyboard gives better results than plugging them into the system (which doesn't make sense to me - they still show a attached to EHCI). My work desktop (Dell OptiPlex GX620, Intel ICH7 chipset 7.0/amd64) refuses to acknowledge EHCI devices - it just reports timeouts and disables that USB port. [See my recent posting to -amd64]. -- Peter Jeremy Please excuse any delays as the result of my ISP's inability to implement an MTA that is either RFC2821-compliant or matches their claimed behaviour.
pgpwvI0ALZw9N.pgp
Description: PGP signature