| Age | Commit message (Collapse) | Author |
|---|
|
*hci cannot receive interrupts before it has been initialized.
this was changed to accomodate the pcibios strategy change back
in pre-2.9 times since *hci_init enables interrupts at it's end
and thus unmapped pci interrupt hangs the machine in the
endless loop trying to deliver it (and no hci handler to ack it).
this new way we disable interrupts by hands before we map 'em
on pci and thus ensure that they will not crash an uninitialized
*hci which, in turn, will run right after the interrupt map/establish
and by the time it enables *hci interrupts all gonna be just fine.
please allow us to end this stretched remark for cvs
is getting a power outage in less than 20 minutes.
|
|
umass.c and ukbd.c as the diffs are a bit hairier.
|
|
|
|
|
|
form of a diff to augustss@netbsd.org so that future syncs will be very easy.
This commit also adds support for ADMtek AN986 "Pegasus" based USB Ethernet,
CATC USB-EL1210A based USB Ethernet, and USB Printers (all untested).
|
|
|
|
|
|
|
|
Make sure not to call tsleep() from suspend/resume routine.
Move more of the transfer completion processing to HC independent code.
Fix some problems with transfer abort & timeout.
Merge the soft_{td,qh} with the real {td,qh}. This saves time and space.
|
|
Diffs sent to augustsson.
|
|
|
