The driver now works on -stable again. Post 4.3-RELEASE MFCs of the
pccard code caused a minor breakage.
Submitted by: Christian Plazas <[email protected]>
Workaround size mismatch in struct drive on alpha platform only.
This struct is shared between kernel and userland and contains a
dev_t, which now refers to a completely different structure in
userland and kernel: int the kernel, userland dev_t is called
udev_t, and dev_t is struct specinfo *. In practical terms, kernel
dev_t is a pointer and userland dev_t is an int, which on the alpha
are different sizes.
As stated, this is a workaround. The correct fix would be in two
parts:
1. Don't define structures in the kernel with the same name and a
different content as a userland structure. This is a
kernel-wide issue: kernel dev_t needs another name.
2. Don't export unnecessary kernel structures to userland. This
can be done within Vinum. It is currently on the tuit queue.
Sleuth work by: Normand Leclerc <[email protected]>
Tested by: Normand Leclerc <[email protected]>
Blanket approved by: jkh
waiting for a lock, and between the time it is woken and
it retries the lock, another process gets it and places it
in the first entry in the table.
This problem has not been observed, but it's possible, and
it's easy enough to fix.
Submitted by: tegge
vinumunlock: Catch a real bug capable of hanging a system. When
releasing a lock, vinumunlock() called wakeup_one. This
caused wakeups to sometimes get lost. After due
consideration, we think that this is due to the fact that
you can't guarantee that some other process is also
waiting on the same address. This makes wakeup_one a
very dangerous function to use.
These changes have not yet been committed to -CURRENT. In view of the
seriousness of the second one and the imminent 4.3-RELEASE, it seems
excusable to commit to -STABLE first.
36 hour test by: grog
Suggested by: alfred
Suffered by: Andrew Gordon <[email protected]>
Approved by: jkh
This was forgotten in the last slew of MFCs, and caused an immediate
panic when starting a striped plex.
Suffered by: "John J. Rushford Jr" <[email protected]>
Approved by: jkh
later. We now call cdesw_remove() to unregister the device, NULL
out any pointers to the vn_softc that exist in dev_t `si_drv1'
fields, and perform a destroy_dev() corresponding to the initial
make_dev().
This is not an MFC; vn(4) is no longer in -current.
PR: kern/18270, i386/26049
Submitted by: [email protected], [email protected]
Reviewed by: phk
Approved by: jkh
destroy child devices on detach to prevent ever-increasing
numbers of pcm/midi devices trying to attach if the module
is repeatedly loaded and unloaded.
Approved by: jkh
devices. opening /dev/{dsp,dspW,audio}0 and then opening a different device
from that list and closing it resulted in a panic when any operation is
performed on the first fd.
we prevent this happening by denying the second open unless it uses the same
minor device as the first.
PR: kern/25519
Approved by: jkh
problems which are fixed in -CURRENT. Now's still not a good time to
do it, so I'd be grateful if people would hammer on this stuff.
Approved by: jkh
Pointy hat to: grog
termination level setting can be found in the seeprom. The old code
could get some false positives causing the termination to be set incorrectly.
This was already in -current, but was missed in the last merge.
Nick Hibma