adrian [Fri, 15 Jul 2016 06:39:35 +0000 (06:39 +0000)]
[ath] [ath_hal] break out the duration calculation to optionally include SIFS.
The pre-11n calculations include SIFS, but the 11n ones don't.
The reason is that (mostly) the 11n hardware is doing the SIFS calculation
for us but the pre-11n hardware isn't. This means that we're over-shooting
the times in the duration field for non-11n frames on 11n hardware, which
is OK, if not a little inefficient.
Now, this is all fine for what the hardware needs for doing duration math
for ACK, RTS/CTS, frame length, etc, but it isn't useful for doing PHY
duration calculations. Ie, given a frame to TX and its timestamp, what
would the end of the actual transmission time be; and similar for an
RX timestamp and figuring out its original length.
So, this adds a new field to the duration routines which requests
SIFS or no SIFS to be included. All the callers currently will call
it requesting SIFS, so this /should/ be a glorious no-op. I'm however
planning some future work around airtime fairness and positioning which
requires these routines to have SIFS be optional.
Notably though, the 11n version doesn't do any SIFS addition at the moment.
I'll go and tweak and verify all of the packet durations before I go and
flip that part on.
Tested:
* AR9330, STA mode
* AR9330, AP mode
* AR9380, STA mode
When building multiple kernels using KERNCONF, non-existent KERNCONF
files will produce an error and buildkernel will fail. Previously missing
KERNCONF files silently failed giving no indication as to why, only to
subsequently discover during installkernel that the desired kernel was
never built in the first place.
- Update aio_return/waitcomplete wrappers for the ssize_t return type.
- Fix the aio_return() wrapper to fail with EINVAL on a pending job.
This matches the semantics of the in-kernel system call. Also,
aio_return() returns errors via errno, not via the return value.
Let DDB's buf printer handle NULL pointers in the buf page array.
A buf's b_pages and b_npages fields may be inconsistent after a panic.
For instance, vfs_vmio_invalidate() sets b_npages to zero only after all
pages are unwired and their page array entries are cleared.
Make PCI interupts allocation static when using bootrom (UEFI).
This makes factual interrupt routing match one shipped with UEFI firmware.
With old firmware this make legacy interrupts work reliable for functions 0
of PCI slots 3-6. Updated UEFI image fixes problem completely.
andrew [Thu, 14 Jul 2016 17:10:54 +0000 (17:10 +0000)]
Move structures only used by the GICv3 ITS driver from a shared header to
the ITS driver file. There is no need for other drivers to need to know
about these structures.
Obtained from: ABT Systems Ltd
MFC after: 1 month
Sponsored by: The FreeBSD Foundation
Always panic if an invalid capability is passed to `capable(..)` instead of
just with INVARIANTS
rwatson's point was valid in the sense that if the data passed at runtime is
invalid, it should always trip the invariant, not just in the debug case.
This is a deterrent against malicious input, or input caused by hardware
errors.
https://www.illumos.org/issues/6878
Summary of changes:
* Replace generic "scan done" message with "scan aborted, restarting",
"scan cancelled", or "scan done"
* Log number of errors using spa_get_errlog_size
* Refactor scan restarting check into static function
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Nav Ravindranath <nav@delphix.com>
MFC after: 2 weeks
https://www.illumos.org/issues/6940
Similar to #6334, but this time with empty directories:
$ zfs create tank/quota
$ zfs set quota=10M tank/quota
$ zfs snapshot tank/quota@snap1
$ zfs set mountpoint=/mnt/tank/quota tank/quota
$ mkdir /mnt/tank/quota/dir # create an empty directory
$ mkfile 11M /mnt/tank/quota/11M
/mnt/tank/quota/11M: initialized 9830400 of 11534336 bytes: Disc quota exceeded
$ rmdir /mnt/tank/quota/dir # now unlink the empty directory
rmdir: directory "/mnt/tank/quota/dir": Disc quota exceeded
From user perspective, I would expect that ZFS is always able to remove files
and directories even when the quota is exceeded.
Reviewed by: Dan McDonald <danmcd@omniti.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Simon Klinkert <simon.klinkert@gmail.com>
MFC after: 2 weeks
https://www.illumos.org/issues/6513
If a ZFS object contains a hole at level one, and then a data block is created
at level 0 underneath that l1 block, l0 holes will be created. However, these
l0 holes do not have the birth time property set; as a result, incremental
sends will not send those holes.
Fix is to modify the dbuf_read code to fill in birth time data.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Boris Protopopov <bprotopopov@hotmail.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Paul Dagnelie <pcd@delphix.com>
MFC after: 3 weeks
https://www.illumos.org/issues/6844
dnode_next_offset is used in a variety of places to iterate over the holes or
allocated blocks in a dnode. It operates under the premise that it can iterate
over the blockpointers of a dnode in open context while holding only the
dn_struct_rwlock as reader. Unfortunately, this premise does not hold.
When we create the zio for a dbuf, we pass in the actual block pointer in the
indirect block above that dbuf. When we later zero the bp in
zio_write_compress, we are directly modifying the bp. The state of the bp is
now inconsistent from the perspective of dnode_next_offset: the bp will appear
to be a hole until zio_dva_allocate finally finishes filling it in. In the
meantime, dnode_next_offset can detect a hole in the dnode when none exists.
I was able to experimentally demonstrate this behavior with the following
setup:
1. Create a file with 1 million dbufs.
2. Create a thread that randomly dirties L2 blocks by writing to the first L0
block under them.
3. Observe dnode_next_offset, waiting for it to skip over a hole in the middle
of a file.
4. Do dnode_next_offset in a loop until we skip over such a non-existent hole.
The fix is to ensure that it is valid to iterate over the indirect blocks in a
dnode while holding the dn_struct_rwlock by passing the zio a copy of the BP
and updating the actual BP in dbuf_write_ready while holding the lock.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Boris Protopopov <bprotopopov@hotmail.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Alex Reece <alex@delphix.com>
MFC after: 3 weeks
https://www.illumos.org/issues/6874
When we do a clone swap (caused by "zfs rollback" or "zfs receive"), the ZPL
doesn't completely reload the state from the DMU; some values remain cached in
the zfsvfs_t.
steps to reproduce:
```
#!/bin/bash -x
zfs destroy -R test/fs
zfs destroy -R test/recvd
zfs create test/fs
zfs snapshot test/fs@a
zfs set userquota@$USER=1m test/fs
zfs snapshot test/fs@b
zfs send test/fs@a | zfs recv test/recvd
zfs send -i @a test/fs@b | zfs recv test/recvd
zfs userspace test/recvd
1. should show 1m quota
dd if=/dev/urandom of=/test/recvd/file bs=1k count=1024
sync
dd if=/dev/urandom of=/test/recvd/file2 bs=1k count=1024
2. should fail with ENOSPC
sync
zfs unmount test/recvd
zfs mount test/recvd
zfs userspace test/recvd
3. if bug above, now shows 1m quota
dd if=/dev/urandom of=/test/recvd/file3 bs=1k count=1024
4. if bug above, now fails with ENOSPC
```
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Approved by: Garrett D'Amore <garrett@damore.org>
Author: Matthew Ahrens <mahrens@delphix.com>
MFC after: 3 weeks
fix-up for configuration of AMD Family 10h processors borrowed from Linux
http://lxr.free-electrons.com/source/arch/x86/kernel/cpu/amd.c#L643
BIOS may configure Family 10h processors to convert WC+ cache type
to CD. That can hurt performance of guest VMs using nested paging.
Back out non-collating [a-z] ranges.
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
Back out non-collating [a-z] ranges.
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
To mimic system glob, we definitely don't need manual upper/lower hack.
The author clearly disagree in the comment, so this patch will be not
submitted upstream.
Back out non-collating [a-z] ranges.
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
Back out non-collating [a-z] ranges.
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
Back out non-collating [a-z] ranges.
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
Back out non-collating [a-z] ranges (r302594).
Instead of changing the whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and focus on fixing bugs in what we have now,
too many small obstacles we have choicing other way, counting ports.
Corresponding libc changes are backed out in r302824.
1) Eliminate possibility to call __*collate_range_cmp() with inclomplete
locale (which cause core dump) by removing whole 'table' argument
by which it passed.
2) Restore __collate_range_cmp() in __sccl().
3) Collating [a-z] range in regcomp() only for single bytes locales
(we can't do it now for other ones). In previous state only first 256
wchars are considered and all others are just silently dropped from the
range.
Back out non-collating [a-z] ranges.
Instead of changing whole course to another POSIX-permitted way
for consistency and uniformity I decide to completely ignore missing
regex fucntionality and concentrace on fixing bugs in what we have now,
too many small obstacles instead, counting ports.
It is not safe to iterate the sub-channel list w/o lock on the
close path, while it's even more difficult to hold the lock
and iterate the sub-channel list. We leverage the
vmbua_{get,rel}_subchan() functions to solve this dilemma.
- Support DEVICE_POLLING
- Increase TX descriptors to 1024
- Add support for passing a chain of mbufs to if_input, reducing the
number of calls to mtx_unlock/mtx_lock under load.
- Remove duplicate byteswap when setting TX_INT_CTL in TX descriptor.
- Set undocumented "TX_NEXT_FRAME" bit in TX control 1 register.
According to the A83T BSP, setting this bit allows the DMA engine to
operate on a packet while receiving another.
Tested on A83T (1000Mbps PHY) and H3 (100Mbps PHY).
Reviewed by: manu
Differential Revision: https://reviews.freebsd.org/D7031