[9fans] naive acid debugging

[9fans] naive acid debugging

[gdiaz]

Hello

While understanding how refs work inside nupas, i decided to try to implement the same function nupas has to find mailbox references in acid langauge, so i did:

defn findmboxrefs(mb,fids) {
	local f, refs;
	f=fids;
	while f!=0 do {
		print("f=", f\X,"\n");
		if f->mb==mb then
			refs++;
		print("f->mb=", f->mb\X,"	=?	mb=",mb\X,"\n");
		f=f->next;
	}
	print("refs ==", refs,"\n");
}

using a Broken nupas (broked by an assert call), i got this:

acid: findmboxrefs(*checkmboxrefs:mb,fids);
f=0x0003a090
f->mb=0x00000000	=?	mb=0x0004a468
f=0x0004bf68
f->mb=0x8d1c2444	=?	mb=0x0004a468
f=0x89ffffe4
<stdin>:75: (error) indir: can't read address 0x89ffffe4: bad arg in system call

is that related to something wrongly done in the acid function?
that means that there is an f which is pointing to an incorrect place?
or means that as the process is broken, acid cannot read some parts of it?

i forgot another possible issue, the broken process is owned by another user, i just chmod'ed /proc/pid/*. . . .

thanks

gabi

Re: [9fans] naive acid debugging

[erik quanstrom]

> While understanding how refs work inside nupas, i decided to try to implement the same function nupas has to find mailbox references in acid langauge, so i did:
>
> defn findmboxrefs(mb,fids) {
> 	local f, refs;
> 	f=fids;
> 	while f!=0 do {
> 		print("f=", f\X,"\n");
> 		if f->mb==mb then
> 			refs++;
> 		print("f->mb=", f->mb\X,"	=?	mb=",mb\X,"\n");
> 		f=f->next;
> 	}
> 	print("refs ==", refs,"\n");
> }

first, i want to complement you on a fine use of acid.
very nice.

>
> using a Broken nupas (broked by an assert call), i got this:
>
> acid: findmboxrefs(*checkmboxrefs:mb,fids);
> f=0x0003a090
> f->mb=0x00000000	=?	mb=0x0004a468
> f=0x0004bf68
> f->mb=0x8d1c2444	=?	mb=0x0004a468
> f=0x89ffffe4
> <stdin>:75: (error) indir: can't read address 0x89ffffe4: bad arg in system call

your assmption that each fid has a mailbox is wrong.
see rattach() and sanefid().  they encode the rules.

- erik

Re: [9fans] naive acid debugging

[gdiaz]

Hello

Thanks Russ, seems that even though i tried to use your iwp9 2007 talk as a reference, i missed the point, i should read more carefully.

gabi

>>>               print("f->mb=", f->mb\X,"       =?      mb=",mb\X,"\n");
>>>               f=f->next;
>
> Always use . (never ->) in acid.
>
> If f is a pointer, then the acid expression f.mb
> is like the C expression (*f).mb aka f->mb.
> The acid expression f->mb is like the C expression (**f).mb.
> By using -> here you are indirecting one level too far.
>
> -> is one of those acid features that seems
> to exist only to cause confusion.
>
> Russ

Re: [9fans] naive acid debugging

[Russ Cox]

>>               print("f->mb=", f->mb\X,"       =?      mb=",mb\X,"\n");
>>               f=f->next;

Always use . (never ->) in acid.

If f is a pointer, then the acid expression f.mb
is like the C expression (*f).mb aka f->mb.
The acid expression f->mb is like the C expression (**f).mb.
By using -> here you are indirecting one level too far.

-> is one of those acid features that seems
to exist only to cause confusion.

Russ

Re: [9fans] naive acid debugging

[gdiaz]

hello

thank you very much  Erik, I'll look at that code today night, i hope to see the light soon ☺

gabi

> While understanding how refs work inside nupas, i decided to try to implement the same function nupas has to find mailbox references in acid langauge, so i did:
>
> defn findmboxrefs(mb,fids) {
> 	local f, refs;
> 	f=fids;
> 	while f!=0 do {
> 		print("f=", f\X,"\n");
> 		if f->mb==mb then
> 			refs++;
> 		print("f->mb=", f->mb\X,"	=?	mb=",mb\X,"\n");
> 		f=f->next;
> 	}
> 	print("refs ==", refs,"\n");
> }

first, i want to complement you on a fine use of acid.
very nice.

>
> using a Broken nupas (broked by an assert call), i got this:
>
> acid: findmboxrefs(*checkmboxrefs:mb,fids);
> f=0x0003a090
> f->mb=0x00000000	=?	mb=0x0004a468
> f=0x0004bf68
> f->mb=0x8d1c2444	=?	mb=0x0004a468
> f=0x89ffffe4
> <stdin>:75: (error) indir: can't read address 0x89ffffe4: bad arg in system call

your assmption that each fid has a mailbox is wrong.
see rattach() and sanefid().  they encode the rules.

- erik