> Replace the remaining KAUTH_GENERIC_ISSUSER authorization calls with
> something meaningful. All relevant documentation has been updated or
> written.
mutex and reduce the overhead.
Also, even if we're polling, we need to use the lock to control access to
the bus, otherwise we get anomolous results.
(These same changes were made several weeks ago to other drivers; I
missed this one.)
a relatively slow CPU can overload a relatively fast Rage 128 in some
circumstances.
This cures weird artifacts seen when scrolling lots and lots of text very
fast. Why this only showed up with glyph cacheing which should reduce the
frequency of host blits isn't quite clear though.
workqueue MPSAFE. This could lead to bnx_start() being running at
the same time on 2 CPUs, with packets being sent twice.
Problem found and reported by Beverly Schwartz and Greg Troxel.
misc configuration devices
- print proper error message if no misc configuration device is found
Fixes kernel crash right after starting Xserver with radeondrm
on ASRock AM2NF3-VSTA (AM2 + nForce3 250 AGP) with Athlon II X2 CPU.
only RW_WRITERs for these, and no RW_READERs, so no need to incur the
extra overhead of allowing for both. As discussed on tech-kern.
For piixpm and ichsmb, the acquire/release protocol needs to be used,
even if the request is I2C_F_POLL'd (or if the device supports only
polled mode). Otherwise multiple requests can be running at the same
time, and they stomp on each other and create anomolous results.
Part 2 addresses my PR kern/45889
3 ACKs from releng
1) Move core entropy-pool code and source/sink/sample management code
to sys/kern from sys/dev.
2) Remove use of NRND as test for presence of entropy-pool code throughout
source tree.
3) Remove use of RND_ENABLED in device drivers as microoptimization to
avoid expensive operations on disabled entropy sources; make the
rnd_add calls do this directly so all callers benefit.
4) Fix bug in recent rnd_add_data()/rnd_add_uint32() changes that might
have lead to slight entropy overestimation for some sources.
5) Add new source types for environmental sensors, power sensors, VM
system events, and skew between clocks, with a sample implementation
for each.
ok releng to go in before the branch due to the difficulty of later
pullup (widespread #ifdef removal and moved files). Tested with release
builds on amd64 and evbarm and live testing on amd64.
where it looks straightforward, and pci_aprint_devinfo_fancy in a few
others where drivers want to supply their own device names instead
of the pcidevs generated one. More complicated cases, where names
are composed at runtime, are left alone for now. It certainly makes
sense to simplify the drivers here rather than inventing a catch-all API.
This should serve as as example for new drivers, and also ensure
consistent output in the AB_QUIET ("boot -q") case. Also, it avoids
excessive stack usage where drivers attach child devices because the
buffer for the device name is not kept on the local stack anymore.
used by drivers: a short name for the quiet/naive case and a string
to override the "pcidevs" based name by one provided by the driver,
ride on yesterday's kernel minor version bump
simplifying uvm_map handling (no special kernel entries anymore no relocking)
make malloc(9) a thin wrapper around kmem(9)
(with private interface for interrupt safety reasons)
releng@ acknowledged
by purpose) - this is a stack hog, and with this change my uTCA amd64
system boots again
a lot of similar code can be eliminated from pci device drivers this way,
but before doing so (and making the new function part of the module API)
I'd like to consider a modification to make it work with drivers which
prefer to print names from other sources (like pciide)
Glyphs which use the default attribute will be saved into offscreen video
memory the first time they're used and subsequent uses will just copy them
from there instead of rendering them every time.
This should give a nice speedup especially on slow CPUs.
nonaka