deadlock in VOP_FSYNC() if the unreferenced vnode picked for
reclamation happened to be stacked on top of a vnode the process
already had locked. This could happen if the same filesystem was
accessed both through a union mount and directly; it seemed to happen
most frequently when the direct access was through NFS.
Avoid this deadlock by changing vinvalbuf to pass a new FSYNC_RECLAIM
flag bit to VOP_FSYNC() to indicate that a reclaim is in progress and
only a `shallow' fsync is necessary.
Do nothing in *_fsync() in umapfs, nullfs, and unionfs when
FSYNC_RECLAIM is set; the underlying vnodes will shortly be released
in *_reclaim and may be reclaimed (and fsync'ed) later.
was changes to comments only, but..)
Build vfs_getcwd.c as standard part of kernel.
Add implementation of fchroot(), since two emulations already had it.
Call vn_isunder() in fchdir(), chroot(), and fchroot() to make it harder
to escape chroot().
Add kernel implementation of getcwd() which uses this cache, falling
back to reading the filesystem on a cache miss.
Along for the ride: add new VOP_FSYNC flag FSYNC_RECLAIM indicating
that a reclaim is being done, so only a "shallow" fsync is needed.
* Rearrange the speed mapping table and adjust the code so that the highest
rate can actually be used. Previously we ended up rounding up slightly
lower speeds and then losing because set_params couldn't set the mode
back to the current one.
* Allow 260 as a valid I/O address, since the SB1 can be jumpered to this.
* Change the MPU-401 code so it can be attached as a separate device.
(XXX Really, the SB code ought to just attach a subdevice itself.)
* Do not attach an OPL on the SB1. Writing to the OPL registers at
SB_base+0 on this card wedges my machine.
(XXX Should we access it at 388 instead? The Creative web site claims
that this board *does* have an OPL2, but I haven't played with this
extensively.)
* Allocate the SB DMA channels at open time, rather than attach time, so
that a single DRQ can be used for multiple cards (if only one is in use
at a given time).
(XXX Let me tell you why this is a horrible hack. If the ISA DMA code
tries to allocate a bounce buffer after boot time, it will generally fail,
because there is no contiguous memory below 16MB and the code to allocate
contiguous pages doesn't know how to move things around. Now, we
shouldn't ever be using bounce buffers here, because we use
isa_dmamem_alloc(). So we just turn off BUS_DMA_ALLOCNOW and we don't
actually try to. That's cool, and it even works, but isa_dmamem_alloc()
has the same problem. It just happens that we allocate the ring buffers
at boot time, and whenever we reallocate them (due to the buffer size
changing), we just deallocated the previous (contiguous) buffer, so we get
lucky. This is absolutely disgusting and needs to be fixed.)
Fix group mapping so members of group 0 get other group-ids mapped as well.
Avoid rename panic by checking (*this_vp_p) against NULLVP before
dereferencing it (same change as to NULLFS some time ago).
sommerfe