RISC OS Open: Forum: BASIC and ZeroPain

Oct 26, 2015 7:45pm

Tank (53) 375 posts

Repeatable error from Function VARINT from castle.RiscOS.Sources.Programmer.BASIC.s.Factor.

Time: Mon Oct 26 19:11:41 2015 Location: Offset 00004768 in module BASIC Current Wimp task: Card Last app to start: TaskWindow 08003974 00000000


R0  = 00000006 R1  = 00000002 R2  = 00000004 R3  = 0000006e

R4  = 00008f39 R5  = 00000000 R6  = 00008f3d R7  = 0000003d

R8  = 00008700 R9  = 00000004 R10 = 0000000d R11 = 00008f3d

R12 = 00008f32 R13 = 000a7fcc R14 = fc1a79fc R15 = fc1a8a84

DFAR = 00000004  Mode USR32  Flags nzCv if   PSR = 20000110
fc1a8a3c : 11a00230 : MOVNE   R0,R0,LSR R2

fc1a8a40 : 12623020 : RSBNE   R3,R2,#&20         ; =" "

fc1a8a44 : 11800311 : ORRNE   R0,R0,R1,LSL R3

fc1a8a48 : 11a01231 : MOVNE   R1,R1,LSR R2

fc1a8a4c : e21110ff : ANDS    R1,R1,#&FF         ; =“ÿ”

fc1a8a50 : 03300000 : TEQEQ   R0,#0

fc1a8a54 : e2003102 : AND     R3,R0,#&80000000

fc1a8a58 : 13800102 : ORRNE   R0,R0,#&80000000

fc1a8a5c : e3b09102 : MOVS    R9,#&80000000

fc1a8a60 : e1a0f00e : MOV     PC,R14

fc1a8a64 : e5d00000 : LDRB    R0,[R0,#0]

fc1a8a68 : e3b09101 : MOVS    R9,#&40000000

fc1a8a6c : e1a0f00e : MOV     PC,R14

fc1a8a70 : e2101003 : ANDS    R1,R0,#3

fc1a8a74 : e3c02003 : BIC     R2,R0,#3

fc1a8a78 : 05900000 : LDREQ   R0,[R0,#0]

fc1a8a7c * 18920005 * LDMNEIA R2,{R0,R2}

fc1a8a80 : 11a01181 : MOVNE   R1,R1,LSL #3

fc1a8a84 : 11a00130 : MOVNE   R0,R0,LSR R1

fc1a8a88 : 12611020 : RSBNE   R1,R1,#&20         ; =" "

fc1a8a8c : 11800112 : ORRNE   R0,R0,R2,LSL R1

fc1a8a90 : e3b09101 : MOVS    R9,#&40000000

fc1a8a94 : e1a0f00e : MOV     PC,R14

fc1a8a98 : 0a00000e : BEQ     &FC1A8AD8

fc1a8a9c : e3590c01 : CMP     R9,#&0100          ; =256

fc1a8aa0 : 2a00316c : BCS     &FC1B5058

fc1a8aa4 : e2482c06 : SUB     R2,R8,#&0600       ; =1536

fc1a8aa8 : e2823c01 : ADD     R3,R2,#&0100       ; =256

fc1a8aac : e2833001 : ADD     R3,R3,#1

fc1a8ab0 : e4d01001 : LDRB    R1,[R0],#1

fc1a8ab4 : e4c21001 : STRB    R1,[R2],#1

fc1a8ab8 : e1320003 : TEQ     R2,R3

Using the following BASIC lines

10 SYS "OS_Hardware" ,&403,0,,,,,,,4 TO ,info%

20 proc=!info%

Line 20 causes the problem

Oct 26, 2015 8:29pm

Jeff Doggett (257) 234 posts

According to https://www.riscosopen.org/wiki/documentation/show/HAL_IRQSource
the call 0,4 to OS_Hardware is the HAL_IRQSource which returns a number in R0 & junk in R1!

Oct 27, 2015 5:34pm

Tank (53) 375 posts

Wrong call Jeff….
See here
It returns the device list, and R2 is a pointer to the descriptor.
To call the routine you mentioned, R8 is 0 and R9 is 4

Oct 27, 2015 7:24pm

Jeff Doggett (257) 234 posts

Dammit, I knew that I must me misreading it somehow! In fact I said as much in the first attempt at posting which the Textile system lost!
I got confused by the leading space before the first comma!

Ok, so it now looks to me that info% is read from R1 – or am I reading that wrong as well?

Oct 27, 2015 8:37pm

Jeffrey Lee (213) 6048 posts

Jeff’s right – OS_Hardware 4 returns the device pointer in R2, but you’re dereferencing R1, which is meant to be an opaque iterator handle.

To save yourself from future shame Tank, feel free to send in a patch to BASIC which implements the ! and ? operator sanity checks which were discussed a while ago :-) (not sure of the thread – sorry!)

Oct 27, 2015 8:45pm

Tank (53) 375 posts

You’ve got me there…
Yes it should be r2..
Mea culpa.
Please ignore this thread.

BASIC and ZeroPain

Reply

Search forums

Social

ROOL Store

Donate! Why?

RISC OS IPR

Description

Voices

Options

Oct 26, 2015 7:45pm Tank (53) 375 posts	Repeatable error from Function VARINT from castle.RiscOS.Sources.Programmer.BASIC.s.Factor. `Time: Mon Oct 26 19:11:41 2015 Location: Offset 00004768 in module BASIC Current Wimp task: Card Last app to start: TaskWindow 08003974 00000000` R0 = 00000006 R1 = 00000002 R2 = 00000004 R3 = 0000006e R4 = 00008f39 R5 = 00000000 R6 = 00008f3d R7 = 0000003d R8 = 00008700 R9 = 00000004 R10 = 0000000d R11 = 00008f3d R12 = 00008f32 R13 = 000a7fcc R14 = fc1a79fc R15 = fc1a8a84 DFAR = 00000004 Mode USR32 Flags nzCv if PSR = 20000110 fc1a8a3c : 11a00230 : MOVNE R0,R0,LSR R2 fc1a8a40 : 12623020 : RSBNE R3,R2,#&20 ; =" " fc1a8a44 : 11800311 : ORRNE R0,R0,R1,LSL R3 fc1a8a48 : 11a01231 : MOVNE R1,R1,LSR R2 fc1a8a4c : e21110ff : ANDS R1,R1,#&FF ; =“ÿ” fc1a8a50 : 03300000 : TEQEQ R0,#0 fc1a8a54 : e2003102 : AND R3,R0,#&80000000 fc1a8a58 : 13800102 : ORRNE R0,R0,#&80000000 fc1a8a5c : e3b09102 : MOVS R9,#&80000000 fc1a8a60 : e1a0f00e : MOV PC,R14 fc1a8a64 : e5d00000 : LDRB R0,[R0,#0] fc1a8a68 : e3b09101 : MOVS R9,#&40000000 fc1a8a6c : e1a0f00e : MOV PC,R14 fc1a8a70 : e2101003 : ANDS R1,R0,#3 fc1a8a74 : e3c02003 : BIC R2,R0,#3 fc1a8a78 : 05900000 : LDREQ R0,[R0,#0] fc1a8a7c * 18920005 * LDMNEIA R2,{R0,R2} fc1a8a80 : 11a01181 : MOVNE R1,R1,LSL #3 fc1a8a84 : 11a00130 : MOVNE R0,R0,LSR R1 fc1a8a88 : 12611020 : RSBNE R1,R1,#&20 ; =" " fc1a8a8c : 11800112 : ORRNE R0,R0,R2,LSL R1 fc1a8a90 : e3b09101 : MOVS R9,#&40000000 fc1a8a94 : e1a0f00e : MOV PC,R14 fc1a8a98 : 0a00000e : BEQ &FC1A8AD8 fc1a8a9c : e3590c01 : CMP R9,#&0100 ; =256 fc1a8aa0 : 2a00316c : BCS &FC1B5058 fc1a8aa4 : e2482c06 : SUB R2,R8,#&0600 ; =1536 fc1a8aa8 : e2823c01 : ADD R3,R2,#&0100 ; =256 fc1a8aac : e2833001 : ADD R3,R3,#1 fc1a8ab0 : e4d01001 : LDRB R1,[R0],#1 fc1a8ab4 : e4c21001 : STRB R1,[R2],#1 fc1a8ab8 : e1320003 : TEQ R2,R3 Using the following BASIC lines `10 SYS "OS_Hardware" ,&403,0,,,,,,,4 TO ,info%` `20 proc=!info%` Line 20 causes the problem

Oct 26, 2015 8:29pm Jeff Doggett (257) 234 posts	According to https://www.riscosopen.org/wiki/documentation/show/HAL_IRQSource the call 0,4 to OS_Hardware is the HAL_IRQSource which returns a number in R0 & junk in R1!

Oct 27, 2015 5:34pm Tank (53) 375 posts	Wrong call Jeff…. See here It returns the device list, and R2 is a pointer to the descriptor. To call the routine you mentioned, R8 is 0 and R9 is 4

Oct 27, 2015 7:24pm Jeff Doggett (257) 234 posts	Dammit, I knew that I must me misreading it somehow! In fact I said as much in the first attempt at posting which the Textile system lost! I got confused by the leading space before the first comma! Ok, so it now looks to me that info% is read from R1 – or am I reading that wrong as well?

Oct 27, 2015 8:37pm Jeffrey Lee (213) 6048 posts	Jeff’s right – OS_Hardware 4 returns the device pointer in R2, but you’re dereferencing R1, which is meant to be an opaque iterator handle. To save yourself from future shame Tank, feel free to send in a patch to BASIC which implements the ! and ? operator sanity checks which were discussed a while ago :-) (not sure of the thread – sorry!)

Oct 27, 2015 8:45pm Tank (53) 375 posts	You’ve got me there… Yes it should be r2.. Mea culpa. Please ignore this thread.