Skip to contentSkip to conversation

syscall numbers larger than 255?

DISCUSSION Chris Friesen 3 participants 3 messages Nov 10, 2003

lore
Share & tools
lore
#0 lore raw

Just a quick and simple question for someone that knows the answer.

Stock 2.4.20 for i386 uses syscalls up to 252. I want to add about a half-dozen new syscalls (forward porting stuff that we've got on 2.4.18).

Does x86 support syscall numbers > 255? If yes, do I have to do anything special to use them? If not, what are my options?

Thanks,

Chris

Signature · 4 lines
Chris Friesen                    | MailStop: 043/33/F10
Nortel Networks                  | work: (613) 765-0557
3500 Carling Avenue              | fax:  (613) 765-2986
Nepean, ON K2H 8E9 Canada        | email: [email protected]
#1 lore raw
Quoted 6 lines · Chris Friesen

Just a quick and simple question for someone that knows the answer.

Stock 2.4.20 for i386 uses syscalls up to 252. I want to add about a half-dozen new syscalls (forward porting stuff that we've got on 2.4.18).

Quoted 6 lines

Does x86 support syscall numbers > 255? If yes, do I have to do anything special to use them? If not, what are my options?

Thanks,

Chris

Sure. With the old int 0x80 calling convention, the syscall number is just put into the EAX register which can contain 32 bits. There may be some mask on the kernel side to limit 'damage', but you can change this if you are adding system-calls. The newer calling convention also has just the 32 bit limitation.

Cheers,
Dick Johnson
Penguin : Linux version 2.4.22 on an i686 machine (797.90 BogoMips).
Note 96.31% of all statistics are fiction.

#2 lore raw
In reply to Chris Friesen
Quoted 8 lines · Chris Friesen

Just a quick and simple question for someone that knows the answer.

Stock 2.4.20 for i386 uses syscalls up to 252. I want to add about a half-dozen new syscalls (forward porting stuff that we've got on 2.4.18).

Does x86 support syscall numbers > 255? If yes, do I have to do anything special to use them? If not, what are my options?

Currently there's a discrepancy between include/asm-i386/unistd.h and arch/i386/kernel/entry.S. While the first define function numbers up to 252, the entry.S file fill up the table up to 258 (epoll crosses the 255 boundary :).

- Davide