T.R | Title | User | Personal Name | Date | Lines |
---|
61.1 | also - | ZYDECO::BODA | Realtime Products Support | Sat Jul 25 1992 00:12 | 18 |
| The VDADDRIVER is a bit confusing in that it allows VME_IRQ but doesn't
specify any irq's in the data buffer to actually set:
vdad$kav_data = 0x00000000; /* No IRQ allowed - */
vdad$kav_setup = KAV$K_ALLOW_VME_IRQ; /* (initially, at least) */
KAV$VME_SETUP( &status,
vdad$kav_setup,
&vdad$kav_data );
This seems to imply that the KAV will mask out VMEbus interrupts since
the default system settings have no irq's enabled.
At what IRQ level does the VDAD device interrupt for its VMEbus vector of 4
and VAX vector of %x810?
Thanks,
Alan
|
61.2 | re: 61.0, VMEbus config., sample KAV30.dat file | SUOSW3::GRAMMER | | Mon Jul 27 1992 10:34 | 45 |
| re: 60.0
see line "characteristic /sys..." in sample .dat file for VMEbus setup
below.
Note, I never use resp. modify VMEbus setup parameters within an
application, because it may affect another one.
vectors, adrs:
1. vdad
dev_irql= %x1; { Interrupt Level }
dev_irqv= %x4; { Interrupt Vector, cf. ebuilds .dat-file }
dev_vme_adrs= %xFE0E00; { devices physical VMEbus address }
2. smod
DEV_IRQL= %x2; { Interrupt Level }
DEV_IRQV= %x20; { Interrupt Vector, cf. ebuilds .dat-file }
DEV_VME_ADRS= %xFD0000; { device's physical VMEbus address }
3. dioz
dev_irql= %x1; { Interrupt Level }
dev_irqv= %x10; { Interrupt Vector, cf. ebuilds .dat-file }
dev_vme_adrs= %xF00000; { device's physical VMEbus address }
Alfons
-----
characteristic /remote_cli /remote_term /nofile /net_device=EZA /noserver
/objects=512 /emulator=both /jobs=32 -
/subprocesses=128 /debug=both /loader=1024
/p0_virtual_size=8192 /p1_virtual_size=1024 /io_region=8096 -
/interrupt_stack=128 /target=24
characteristic /sys1_param=%X400FEFF /sys2_param=%X400007
node /lat=active
program SETTIME /initialize
program vdad_SRV /run /warm_debug /kernel_stack=64 /mode=kernel /job_priority=10
program dioz_SRV /run /warm_debug /kernel_stack=64 /mode=kernel /job_priority=12
program smod_SRV /run /warm_debug /kernel_stack=64 /mode=kernel /job_priority=14
device TTA /vector=%X2C0
device EZA /vector=%X130 /net_def
device VDAD /vector=%X810 /noautoload
device DIOZ /vector=%X840 /noautoload
device DIOZ_W /vector=%X840 /noautoload
device SMOD /vector=%X880 /noautoload
terminal TTA1 /driver=SCN /eightbit
|
61.3 | re .0, mvmedriver .dat file | GOBANG::LEMMER | | Mon Jul 27 1992 13:42 | 22 |
|
the following is one of the .dat files I used during 'development'
of this sample driver.
MVMEDRIVER_T1 is the 'main' programm which calls the driver
MVMEDRIVER_AST is the 'ast' version of the driver (documented
in the PRI and also in sys$examples.kav after installation of
the toolkit). System parameter setting is described in the PRI
chapter 5.4.1
--------------------------------------------------------------------------------
characteristic /nonetwork /nofile /net_device=EZA /noserver /emulator=both -
/debug=none -
/p0_virtual_size=2048 /p1_virtual_size=256 /io_region=1042 /interrupt_stack=8 -
/errlog=remote /target=24 -
/sys1_param=%X370000FF /sys2_param=%X400007
program MVMEDRIVER_T1 /kernel_stack=4 /user_stack=4
program MVMEDRIVER_AST /initialize /kernel_stack=8 /user_stack=2 /mode=kernel -
/job_priority=5 /argument=("MVME")
device MVME /vector=%X808 /noautoload
terminal CONSOLE /scope /display=64914::0.0 /vt3xx
--------------------------------------------------------------------------------
|
61.4 | re .1, vdad driver | GOBANG::LEMMER | | Mon Jul 27 1992 13:46 | 7 |
|
as far as I know, the documented vdad driver did not use irq's...
there are several different drivers for the vdad board now
'hanging around'.., you may ask Alfons Grammer (see .2) for the
sources of his driver.
|
61.5 | retrieving the device dependent parameter | HERR::CROSBIE | | Mon Jul 27 1992 14:06 | 18 |
| Re: .0
Hi Alan,
The device dependent parameter is an extension to the system
configuration record structure that was primarily added for use by
rtVAX 300 based applications. I'm not sure if the method of accessing
it is documented (it may be documented in the rtVAX 300 supplement).
To retrieve the device dependent parameter it is necessary to walk
through the system device list (pointed to by KER$GA_DEVICE_LIST),
checking each system configuration record (SCR) for a match with your
device, the device dependent parameter is at an offset of
SCR$L_DEVICE_DEPENDENT within the SCR. Traversing the device list is
relatively simple as it is physically contiguous.
Graham
|
61.6 | Thanks. How about default values? | ZYDECO::BODA | Realtime Products Support | Mon Jul 27 1992 19:53 | 30 |
| Thanks to all for your replies.
Just a few loose ends:
re .4: As I read further in the VDAD device driver, it does ultimately
setup allowing for IRQ 1 in the VDAD$SETUP routine. I guess this
was just one way of masking interrupts until it was ready.
re .2, .3: In each of your System parameter #1 you have the first byte
set to %XFF indicating that default configurations should be
used at startup. However, you have different values in the
other bytes. I'm not sure whether this means you are using
the default settings or whether you are giving specific
settings. The doc'n on p5-15 is unclear in the sentence:
"When you want to use the default configurations, set this byte to FF(16)
and make sure that System Parameter 1 and System Parameter 2 contain
valid settings. When you do not want to use the default configurations,
set this byte to any value other than FF(16)."
If one wants to use the default settings and specifies %xFF in the low
byte of param 1, why should it matter what is in the rest of system param 1
and 2? It seems if you wanted to set specific values you would have
specified something other than %xFF in the low byte of param 1.
Am I missing something?
Thanks again,
Alan
|
61.7 | | HERR::CROSBIE | | Tue Jul 28 1992 10:33 | 14 |
| Hi Alan,
Re: default values, the documentation is incorrect.
The least significant byte of system parameter 1 needs to be set to FF
to override the default settings, if you set the least significant byte
to FF then you must provide valid settings for all of the parameters
that can be set up from system parameters 1 and 2.
Thanks for reporting this, I'll see that it gets changed for VAXELN
4.4.
Graham
|