to achieve a delay is pretty mean.
Andrew reports:
"The tulip_delay_300ns() is, well, bloody stupid on machines with a
heavily loaded PCI bus. It tries to do a delay by assuming PCI reads
will take a certain amount of time & issues a large amount of
(expensive, 5% CPU when your PCI bus is heavily loaded) pci reads.
Locally, we've replaced the calls to tulip_delay_300ns(sc) in the EMIT
macros with a simple DELAY(1) and not seen any problems. Plus we've
gained about 50Mb/sec throughput on our gigabit network cards because
of the added PCI bus bandwidth available."
Also, I do not understand why, but this change appears to stop the
Transmit Fifo underrun on one of my systems (but not the Alpha PC164SX).
This shouldn't make that much of a difference since the mii bus isn't
touched all that often, but perhaps when it does get accessed and hence
hammers the register, it was causing the chip to get upset.
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
-------------------------------------------------------------------------------
Recent versions of 3.0-current have the bktr driver built in. Older versions
of 3.0 and all versions of 2.2 need to have the driver files installed by hand:
cp ioctl_bt848.h /sys/i386/include/
cp brktree_reg.h brooktree848.c /sys/pci/
In /sys/conf/files add:
pci/brooktree848.c optional bktr device-driver
-------------------------------------------------------------------------------
In all cases you will need to add the driver to your kernel:
In your kernel configuration file:
controller pci0 #if you already have this line don't add it.
device bktr0
There is no need to specify DMA channels nor interrupts for this
driver.
-------------------------------------------------------------------------------
Finally you need to create nodes for the driver:
Create a video device:
mknod /dev/bktr0 c 92 0
Create a tuner device:
mknod /dev/tuner0 c 92 16
-------------------------------------------------------------------------------
The code attempts to auto-probe code to detect card/tuner types.
The detected card is printed in the dmesg as the driver is loaded. If
this fails to detect the proper card you can override it in brooktree848.c:
#define OVERRIDE_CARD <card type>
where <card type> is one of:
CARD_UNKNOWN
CARD_MIRO
CARD_HAUPPAUGE
CARD_STB
CARD_INTEL
-------------------------------------------------------------------------------
This model now separates the "tuner control" items into a minor device:
minor device layout: xxxxxxxx xxxT UUUU
UUUU: the card (ie UNIT) identifier, 0 thru 15
T == 0: video device
T == 1: tuner device
Access your tuner ioctl thru your tuner device handle and anything
which controls the video capture process thru the video device handle.
Certain ioctl()s such as video source are available thru both devices.
-------------------------------------------------------------------------------
If your tuner does not work properly or is not recognized properly
try setting the tuner type via or card type:
sysctl -w hw.bt848.card=<integer> current valid values are 0 to 5 inclusive
sysctl -w hw.bt848.tuner=<integer> where integer is a value from 1 to 10
systcl -w hw.bt848.reverse_mute=<1 | 0> to reverse the mute function in the
driver set variable to 1.
The exact format of the sysctl bt848 variable is:
unit << 8 | value
unit identifies the pci bt848 board to be affected 0 is the first bt848
board, 1 is the second bt848 board.
value denotes the integer value for tuners is a value from 0 to 10 for
reversing the mute function of the tuner the value is 1 or 0.
to find out all the bt848 variables:
sysctl hw.bt848
-------------------------------------------------------------------------------
The bt848 driver consists of:
src/sys/i386/include/ioctl_bt848.h
src/sys/pci/brktree_reg.h
src/sys/pci/brooktree848.c
e9fc0b372f