HIL status (fwd)
|On Wed, Nov 10, 1999 at 04:03:39AM -0500, Alex deVries wrote: |> |> Having ASP interrupts working will enable a lot of different boxes to be |> supported, mostly 715s. I have a /50, /75 and probably /100 to test on. | |The /50 and /75 are also Scorpio machines. The /100 is either a Mirage or |a Turnip. As such, I don't know if it has ASP.
/100 is a LASI machine, like 80 and 64. They are the original 3 mirages. Turnip is a mirage with bigger caches, shouldn't affect driver stuff.
|> The closest things we have to ASP docs are: |> - take the portions of the Lasi docs that describe how Lasi is not like |> ASP, and read backwards | |umm.. 4MB download this evening then... | |> Hrm. That's uncomfortable. So did it actually work at 0xf0800000? | |Yep. | |> The problem comes when we hit an ASP that's not at 0xf0800000, which I'd |> think might come up with ASP add on cards. | |Do ASP add-in cards exist? We could just refuse to support them until |we know more :-)
I've never heard of one, and I've also never heard of a slot that might support one. SO, with 99.9% confidence, no they don't exist.
|
|> I don't think that firmware is lying, though. Perhaps we just need to
|> subtract 0x2f000 for some reason.
|
|Found out what this is in the mach sources. It's called CORE_ASP:
|
|/*
| * Structures defining memory-mapped registers in the Core IOSS (ASP).
| */
| vu_char reset; /* (WO) IOSS reset */
| vu_char version; /* (RO) version byte */
| vu_char scsi_dsync; /* (WO) SCSI DSYNC enable bit (1:enable) */
| vu_char error; /* bus errors while ASP masters SGC/VSC */
| vu_char lan_oen; /* (WO) output enable Ethernet/FDDI */
|
|Shortly below this is:
|
|/*
| * Structures defining memory-mapped registers in the Core IOSS (LASI).
| */
|struct core_lasi {
| vu_int power;
| vu_int error;
| vu_int version;
| vu_int reset;
| vu_int arbmask;
|};
|
|which is defined to be at 0xF010C000.
|
|> > Here's the driver list printed at boot on the 715 (is someone collecting
|> > these somewhere?)
|>
|> Within a week, I'll send some mail about The Puffin PA RISC Hardware
|> Guide. Details to come.
|
|Oh good.If you have specific Scorpio/ASP questions, mail them to me, and I'll look them up in the docs.
-Bill
[parisc-linux] ASP
Quoted 5 lines · Bill Katz William
|Do ASP add-in cards exist? We could just refuse to support them until |we know more :-)
I've never heard of one, and I've also never heard of a slot that might support one. SO, with 99.9% confidence, no they don't exist.
Excellent.
Quoted 2 lines
If you have specific Scorpio/ASP questions, mail them to me, and I'll look them up in the docs.
OK, here's one...
When I don't attempt to program anything, the ASP defaults to irq 3. Is this changeable? What other devices can cause interrupts on Scorpio- class machines, that don't go through ASP?
The following devices _do_ go through ASP:
EISA, 8042, SCSI, FDDI, RS232, Centronics, Ethernet, Graphics (1 + 2), Audio.
which seems pretty comprehensive. I'm wondering about slimming down our interrupt architecture for older machines; it's a trifle over-engineered for a machine which can't be expanded.
Signature · 4 lines
Matthew Wilcox <[email protected]> "Windows and MacOS are products, contrived by engineers in the service of specific companies. Unix, by contrast, is not so much a product as it is a painstakingly compiled oral history of the hacker subculture." - N Stephenson
Re: [parisc-linux] ASP
Matthew Wilcox wrote:
...
Quoted 7 lines
The following devices _do_ go through ASP:
EISA, 8042, SCSI, FDDI, RS232, Centronics, Ethernet, Graphics (1 + 2), Audio.
which seems pretty comprehensive. I'm wondering about slimming down our interrupt architecture for older machines; it's a trifle over-engineered for a machine which can't be expanded.
We need to be able to boot a CD on *any* HP box which parisc-linux wants to support in order to install parisc-linux. If the kernel modules will get rebuilt as part of a second pass to optimize code path for older machines, go for it.
cheers,
grant
Grant Grundler
Unix Developement Lab
+1.408.447.7253
Re: [parisc-linux] ASP
In reply to Matthew WilcoxQuoted 3 lines
which seems pretty comprehensive. I'm wondering about slimming down our interrupt architecture for older machines; it's a trifle over-engineered for a machine which can't be expanded.
it is ? Of course, you could hardcode the interrupt handlers in the ASP main interrupt handler, so you would do
switch(ffs(mask)) {
case 0:
return; case 1: asp_something_interrupt(); break; case 2: asp_something_else_interrupt(); break; case 3: asp_keyboard_interrupt(); break; }
but my guess is this would just result in a jump table of function calls rather than an indirect call - not going to improve performance measurably.
Philipp Rumpf
Re: [parisc-linux] ASP
Quoted 5 lines · Philipp Rumpf
but my guess is this would just result in a jump table of function calls rather than an indirect call - not going to improve performance measurably.
I think it depends how compiled deal with pointer to function, in the following example you got the asm code gened (with HP compiler), in compile #1 regular pointer to function, then the jump table is better
as the call to dyncall to perform the function call cost about dozen of instruction, space register switching, etc...
On compile2 instructing the compiler that there is no shared lib (what does happen on HP-UX kenel despite the DLKM, then the +ESfic canbe used greatly reducing the pointer to function deref.
So I would say, if your compiler have a +ESfic kind of feature, you'd better use the pointer to function array.
ic.c::
int f0(){}
int f1(){}
int f2(){}
int f3(){}
int (*fp[4])()={f0,f1,f2,f3};
f(i)
{ switch(i)
{ case 0:f0();break;
case 1:f1();break;
case 2:f2();break;
case 3:f3();break;}
(*fp[i])();
}
compile 1================================================
$ cc -S ic.c -O ; more ic.s
.LEVEL 1.1 .SPACE $TEXT$,SORT=8
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f0
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x0
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f1
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x8
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f2
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x10
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f3
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x18
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f
.PROC
.CALLINFO CALLER,FRAME=16,ENTRY_GR=%r3,SAVE_RP,ARGS_SAVED
.ENTRY
STW %r2,-20(%r30) ;offset 0x20
STW,MA %r3,64(%r30) ;offset 0x24
CMPIB,<< 3,%r26,$00000001 ;offset 0x28
COPY %r26,%r3 ;offset 0x2c
CMPIB,=,N 3,%r3,$00000005 ;offset 0x30
CMPIB,<<,N 1,%r3,$00000004 ;offset 0x34
CMPIB,=,N 1,%r3,$00000003 ;offset 0x38
CMPB,<>,N %r3,%r0,$00000001 ;offset 0x3c
$00000002
.CALL RTNVAL=GR ;out=28;
B,L f0,%r2 ;offset 0x40
NOP ;offset 0x44
B $00000001+4 ;offset 0x48
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x4c
$00000003
.CALL RTNVAL=GR ;out=28;
B,L f1,%r2 ;offset 0x50
NOP ;offset 0x54
B $00000001+4 ;offset 0x58
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x5c
$00000004
.CALL RTNVAL=GR ;out=28;
B,L f2,%r2 ;offset 0x60
NOP ;offset 0x64
B $00000001+4 ;offset 0x68
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x6c
$00000005
.CALL RTNVAL=GR ;out=28;
B,L f3,%r2 ;offset 0x70
NOP ;offset 0x74
$00000001
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x78
LDO RR'fp-$global$(%r1),%r31 ;offset 0x7c
LDW,S %r3(%r31),%r22 ;offset 0x80
.CALL RTNVAL=GR ;in=22;out=28;
B,L $$dyncall,%r31 ;offset 0x84
COPY %r31,%r2 ;offset 0x88
LDW -84(%r30),%r2 ;offset 0x8c
BV %r0(%r2) ;offset 0x90
.EXIT
LDW,MB -64(%r30),%r3 ;offset 0x94
.PROCEND ;in=26;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SPACE $PRIVATE$,SORT=16
.SUBSPA $DATA$,QUAD=1,ALIGN=8,ACCESS=0x1f,SORT=16
fp
.ALIGN 8
.WORD P'f0
.WORD P'f1
.WORD P'f2
.WORD P'f3
.IMPORT $global$,DATA
.SPACE $TEXT$
.SUBSPA $CODE$
.EXPORT f0,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f1,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f2,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f3,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SPACE $PRIVATE$
.SUBSPA $DATA$
.EXPORT fp
.SPACE $TEXT$
.SUBSPA $CODE$
.EXPORT f,ENTRY,PRIV_LEV=3,ARGW0=GR,RTNVAL=GR
.IMPORT $$dyncall,MILLICODE
.ENDcompile 2=======================================================
$ cc -S ic.c +ESfic -O; more ic.s
.LEVEL 1.1 .SPACE $TEXT$,SORT=8
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f0
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x0
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f1
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x8
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f2
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x10
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f3
.PROC
.CALLINFO FRAME=0
.ENTRY
.EXIT
BV,N %r0(%r2) ;offset 0x18
.PROCEND ;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY,SORT=24
f
.PROC
.CALLINFO CALLER,FRAME=80,ENTRY_GR=%r3,SAVE_RP,ARGS_SAVED
.ENTRY
STW %r2,-20(%r30) ;offset 0x20
STW,MA %r3,128(%r30) ;offset 0x24
CMPIB,<< 3,%r26,$00000001 ;offset 0x28
COPY %r26,%r3 ;offset 0x2c
CMPIB,=,N 3,%r3,$00000005 ;offset 0x30
CMPIB,<<,N 1,%r3,$00000004 ;offset 0x34
CMPIB,=,N 1,%r3,$00000003 ;offset 0x38
CMPB,<>,N %r3,%r0,$00000001 ;offset 0x3c
$00000002
.CALL RTNVAL=GR ;out=28;
B,L f0,%r2 ;offset 0x40
NOP ;offset 0x44
B $00000001+4 ;offset 0x48
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x4c
$00000003
.CALL RTNVAL=GR ;out=28;
B,L f1,%r2 ;offset 0x50
NOP ;offset 0x54
B $00000001+4 ;offset 0x58
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x5c
$00000004
.CALL RTNVAL=GR ;out=28;
B,L f2,%r2 ;offset 0x60
NOP ;offset 0x64
B $00000001+4 ;offset 0x68
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x6c
$00000005
.CALL RTNVAL=GR ;out=28;
B,L f3,%r2 ;offset 0x70
NOP ;offset 0x74
$00000001
ADDIL LR'fp-$global$,%r27,%r1 ;offset 0x78
LDO RR'fp-$global$(%r1),%r31 ;offset 0x7c
LDW,S %r3(%r31),%r24 ;offset 0x80
STW %r24,-112(%r30) ;offset 0x84
LDW -112(%r30),%r23 ;offset 0x88
STW %r23,-120(%r30) ;offset 0x8c
.CALL RTNVAL=GR ;out=28;
BE,L 0(%sr4,%r23),%r31 ;offset 0x90
COPY %r31,%r2 ;offset 0x94
LDW -148(%r30),%r2 ;offset 0x98
LDW -120(%r30),%r23 ;offset 0x9c
BV %r0(%r2) ;offset 0xa0
.EXIT
LDW,MB -128(%r30),%r3 ;offset 0xa4
.PROCEND ;in=26;out=28;
.SPACE $TEXT$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SUBSPA $CODE$
.SPACE $PRIVATE$,SORT=16
.SUBSPA $DATA$,QUAD=1,ALIGN=8,ACCESS=0x1f,SORT=16
fp
.ALIGN 8
.WORD P'f0
.WORD P'f1
.WORD P'f2
.WORD P'f3
.IMPORT $global$,DATA
.SPACE $TEXT$
.SUBSPA $CODE$
.EXPORT f0,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f1,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f2,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SUBSPA $CODE$
.EXPORT f3,ENTRY,PRIV_LEV=3,RTNVAL=GR
.SPACE $PRIVATE$
.SUBSPA $DATA$
.EXPORT fp
.SPACE $TEXT$
.SUBSPA $CODE$
.EXPORT f,ENTRY,PRIV_LEV=3,ARGW0=GR,RTNVAL=GR
.END======================================================================
--
mailto:[email protected]
WTEC Project. Kernel debugging tools