Page 3 of 4
Posted: Tue May 06, 2008 4:17 am
by JamesM
jal wrote:pcmattman wrote:This looks suspicious, but I'm not 100% sure:
Shouldn't that & be an &&? Or do I just not understand the syntax?
It looks suspicious indeed, but it works, since & has a low priority. So id > 7 and id < 15 are evaluated (yielding 0 or 1), and the result is 0 or 1.
JAL
Given that he's using ~ to mean NOT I see two possibilities:
(a) ~ = logical NOT, & = logical AND, the problem is elsewhere.
(b) ~ = one's complement, & = binary AND, the OP doesn't know the difference and the problem is here.
Being that I don't know the macro language being used I'll leave it up to someone more qualified to make the choice of those two!
Cheers,
James
Posted: Tue May 06, 2008 6:28 am
by White-spirit
I've debugged my isr14 and it seems to be correct, it does just one push :
Code: Select all
(3219620396) Breakpoint 3219620264, in 0008:001031b1 (0x001031b1)
Next at t=298296331
(0) [0x001031b1] 0008:00000000001031b1 (unk. ctxt): cli ; fa
<bochs:5> s
Next at t=298296332
(0) [0x001031b2] 0008:00000000001031b2 (unk. ctxt): push 0x0000000e ; 6a0e
<bochs:6>
Next at t=298296333
(0) [0x001031b4] 0008:00000000001031b4 (unk. ctxt): jmp .+0x00000077 (0x0010322d) ; eb77
<bochs:7>
So I think the macro is correct .
Posted: Thu May 08, 2008 6:29 am
by White-spirit
I corrected the macro and the "if" condition must be :
Code: Select all
if ~ ( id = 8 | id > 9 & id < 15 )
But it still doesn't work, I've also discovered that all the interrupts 8, 10-14 don't work, but the rest does his work perfectly ....
Can I post a floppy image here ?
Thanks
Posted: Fri May 09, 2008 1:01 am
by jal
White-spirit wrote:But it still doesn't work, I've also discovered that all the interrupts 8, 10-14 don't work, but the rest does his work perfectly ....
I'd rather you disassemble the assembled code and look what the output is, so you can check whether the macro is assembled correctly (i.e. as you intended). A floppy image doesn't help at all, we really believe you when you say it doesn't work.
JAL
Posted: Fri May 09, 2008 1:12 am
by White-spirit
Thanks, I've disassembled my idt.s.o and it seems that the interrupts 8, 10-14 are doing only one push ( so it's correct ) .
It has nothing to do with this ? AlfaOmega08 has a similar problem, and he uses also JamesM's macro :
http://www.osdev.org/phpBB2/viewtopic.php?t=16922
Code: Select all
idt.s.o: file format elf32-i386
Disassembly of section .flat:
00000000 <load_idt>:
0: 55 push %ebp
1: 89 e5 mov %esp,%ebp
3: 8b 5c 24 08 mov 0x8(%esp),%ebx
7: 0f 01 1b lidtl (%ebx)
a: c9 leave
b: c3 ret
0000000c <isr0>:
c: fa cli
d: 6a 00 push $0x0
f: 6a 00 push $0x0
11: e9 f1 00 00 00 jmp 107 <isr31+0x7>
00000016 <isr1>:
16: fa cli
17: 6a 00 push $0x0
19: 6a 01 push $0x1
1b: e9 e7 00 00 00 jmp 107 <isr31+0x7>
00000020 <isr2>:
20: fa cli
21: 6a 00 push $0x0
23: 6a 02 push $0x2
25: e9 dd 00 00 00 jmp 107 <isr31+0x7>
0000002a <isr3>:
2a: fa cli
2b: 6a 00 push $0x0
2d: 6a 03 push $0x3
2f: e9 d3 00 00 00 jmp 107 <isr31+0x7>
00000034 <isr4>:
34: fa cli
35: 6a 00 push $0x0
37: 6a 04 push $0x4
39: e9 c9 00 00 00 jmp 107 <isr31+0x7>
0000003e <isr5>:
3e: fa cli
3f: 6a 00 push $0x0
41: 6a 05 push $0x5
43: e9 bf 00 00 00 jmp 107 <isr31+0x7>
00000048 <isr6>:
48: fa cli
49: 6a 00 push $0x0
4b: 6a 06 push $0x6
4d: e9 b5 00 00 00 jmp 107 <isr31+0x7>
00000052 <isr7>:
52: fa cli
53: 6a 00 push $0x0
55: 6a 07 push $0x7
57: e9 ab 00 00 00 jmp 107 <isr31+0x7>
0000005c <isr8>:
5c: fa cli
5d: 6a 08 push $0x8
5f: e9 a3 00 00 00 jmp 107 <isr31+0x7>
00000064 <isr9>:
64: fa cli
65: 6a 00 push $0x0
67: 6a 09 push $0x9
69: e9 99 00 00 00 jmp 107 <isr31+0x7>
0000006e <isr10>:
6e: fa cli
6f: 6a 0a push $0xa
71: e9 91 00 00 00 jmp 107 <isr31+0x7>
00000076 <isr11>:
76: fa cli
77: 6a 0b push $0xb
79: e9 89 00 00 00 jmp 107 <isr31+0x7>
0000007e <isr12>:
7e: fa cli
7f: 6a 0c push $0xc
81: e9 81 00 00 00 jmp 107 <isr31+0x7>
00000086 <isr13>:
86: fa cli
87: 6a 0d push $0xd
89: eb 7c jmp 107 <isr31+0x7>
0000008b <isr14>:
8b: fa cli
8c: 6a 0e push $0xe
8e: eb 77 jmp 107 <isr31+0x7>
00000090 <isr15>:
90: fa cli
91: 6a 00 push $0x0
93: 6a 0f push $0xf
95: eb 70 jmp 107 <isr31+0x7>
00000097 <isr16>:
97: fa cli
98: 6a 00 push $0x0
9a: 6a 10 push $0x10
9c: eb 69 jmp 107 <isr31+0x7>
0000009e <isr17>:
9e: fa cli
9f: 6a 00 push $0x0
a1: 6a 11 push $0x11
a3: eb 62 jmp 107 <isr31+0x7>
000000a5 <isr18>:
a5: fa cli
a6: 6a 00 push $0x0
a8: 6a 12 push $0x12
aa: eb 5b jmp 107 <isr31+0x7>
000000ac <isr19>:
ac: fa cli
ad: 6a 00 push $0x0
af: 6a 13 push $0x13
b1: eb 54 jmp 107 <isr31+0x7>
000000b3 <isr20>:
b3: fa cli
b4: 6a 00 push $0x0
b6: 6a 14 push $0x14
b8: eb 4d jmp 107 <isr31+0x7>
000000ba <isr21>:
ba: fa cli
bb: 6a 00 push $0x0
bd: 6a 15 push $0x15
bf: eb 46 jmp 107 <isr31+0x7>
000000c1 <isr22>:
c1: fa cli
c2: 6a 00 push $0x0
c4: 6a 16 push $0x16
c6: eb 3f jmp 107 <isr31+0x7>
000000c8 <isr23>:
c8: fa cli
c9: 6a 00 push $0x0
cb: 6a 17 push $0x17
cd: eb 38 jmp 107 <isr31+0x7>
000000cf <isr24>:
cf: fa cli
d0: 6a 00 push $0x0
d2: 6a 18 push $0x18
d4: eb 31 jmp 107 <isr31+0x7>
000000d6 <isr25>:
d6: fa cli
d7: 6a 00 push $0x0
d9: 6a 19 push $0x19
db: eb 2a jmp 107 <isr31+0x7>
000000dd <isr26>:
dd: fa cli
de: 6a 00 push $0x0
e0: 6a 1a push $0x1a
e2: eb 23 jmp 107 <isr31+0x7>
000000e4 <isr27>:
e4: fa cli
e5: 6a 00 push $0x0
e7: 6a 1b push $0x1b
e9: eb 1c jmp 107 <isr31+0x7>
000000eb <isr28>:
eb: fa cli
ec: 6a 00 push $0x0
ee: 6a 1c push $0x1c
f0: eb 15 jmp 107 <isr31+0x7>
000000f2 <isr29>:
f2: fa cli
f3: 6a 00 push $0x0
f5: 6a 1d push $0x1d
f7: eb 0e jmp 107 <isr31+0x7>
000000f9 <isr30>:
f9: fa cli
fa: 6a 00 push $0x0
fc: 6a 1e push $0x1e
fe: eb 07 jmp 107 <isr31+0x7>
00000100 <isr31>:
100: fa cli
101: 6a 00 push $0x0
103: 6a 1f push $0x1f
105: eb 00 jmp 107 <isr31+0x7>
107: 60 pusha
108: 66 8c d8 mov %ds,%ax
10b: 50 push %eax
10c: 66 b8 10 00 mov $0x10,%ax
110: 8e d8 mov %eax,%ds
112: 8e c0 mov %eax,%es
114: 8e e0 mov %eax,%fs
116: 8e e8 mov %eax,%gs
118: e8 fc ff ff ff call 119 <isr31+0x19>
11d: 5b pop %ebx
11e: 8e db mov %ebx,%ds
120: 8e c3 mov %ebx,%es
122: 8e e3 mov %ebx,%fs
124: 8e eb mov %ebx,%gs
126: 61 popa
127: 83 c4 08 add $0x8,%esp
12a: fb sti
12b: cf iret
Posted: Fri May 09, 2008 1:26 am
by JamesM
Hi,
My macro works perfectly.
Problem
Exists
Between
Keyboard
And
Chair
, IMHO. Why you are trying to trigger the exceptions which push an error code manually, I have no idea.
Cheers,
James
Posted: Fri May 09, 2008 1:40 am
by White-spirit
JamesM wrote:Hi,
My macro works perfectly.
Problem
Exists
Between
Keyboard
And
Chair
, IMHO. Why you are trying to trigger the exceptions which push an error code manually, I have no idea.
Cheers,
James
Well, I just want to test a page fault exception by accessing to a non-present page .
Posted: Fri May 09, 2008 2:08 am
by AJ
If I was debugging this, one of the first things I'd do would be to take out the call to isr_handler. When the first PFE occurs, your OS will then go in to an infinite loop, but at least you will find out if it's the stub causing the problem or not.
Cheers,
Adam
Posted: Fri May 09, 2008 3:13 am
by White-spirit
I've removed the isr_handler call, and it still doesn't work, so the problem comes from the stub .
Posted: Fri May 09, 2008 3:49 am
by AJ
Wait a minute - the paging interupt is int 14. From your disassembly, the PFE ISR does not push a dummy error code and it should do.
Cheers,
Adam
Posted: Fri May 09, 2008 4:16 am
by White-spirit
JamesM said that the interrupts 8, 10-14 push error codes, so I need to do only one push, or not ?
Posted: Fri May 09, 2008 4:51 am
by AJ
Bugg*r - sorry. My Mistake
Posted: Fri May 09, 2008 5:48 am
by White-spirit
Here's a screenshot if it can help you .
Thanks
Posted: Fri May 09, 2008 9:01 am
by White-spirit
With QEmu I get interrupts 14 instead of 13... Is that an emulator related issue ?
Posted: Fri May 09, 2008 10:18 am
by Combuster
It means you borked something else.
Are you perchance using an
unlinked version of your code? The address of the call instruction is just pointing nowhere
