Syscall - Revision history

Ralf at 16:55, 11 September 2012

2012-09-11T16:55:41Z

Ralf at 08:42, 11 September 2012

2012-09-11T08:42:37Z

Ralf: /* Calling conventions */

2012-09-11T08:33:15Z

‎Calling conventions

Ralf: /* Calling conventions */

2012-09-11T08:32:07Z

‎Calling conventions

Ralf: /* Calling conventions */

2012-09-11T08:05:15Z

‎Calling conventions

Ralf at 08:02, 11 September 2012

2012-09-11T08:02:39Z

Ralf: Created page with "Syscalls on == Syscall number ranges == {| {{PrettyTable}} ! OS flavor || First || Last |- | System V Release 4 flavored syscalls || 0..."

2012-09-11T06:30:08Z

Created page with "Syscalls on == Syscall number ranges == {| {{PrettyTable}} ! OS flavor || First || Last |- | System V Release 4 flavored syscalls || 0..."

New page

Syscalls on

== Syscall number ranges ==
{| {{PrettyTable}}
! OS flavor || First || Last
|-
| System V Release 4 flavored syscalls || 0 || 999
|-
| System V syscalls. All flavors of IRIX use this number range as well. || 1000 || 1999
|-
| BSD 4.3 syscalls || 2000 || 2999
|-
| POSIX syscalls || 3000 || 3999
|-
| Linux O32 syscalls || 4000 || 4999
|-
| Linux N64 syscalls || 5000 || 5999
|-
| Linux N32 syscalls || 6000 || 6999
|}

== See also ==
* [[MIPS OS]]
* [[MIPS ABI History]]

@@ Line 7: / Line 7: @@
 {| {{PrettyTable}}
-! Register !! use on input !! use on output
+! Register                   !! use on input         !! use on output                    !! Note
 |-
-| $at                        || —                    || (caller saved)
+| $at                        || —                    || (caller saved)                   ||
 |-
-| $v0                        || syscall number       || return value
+| $v0                        || syscall number       || return value                     ||
 |-
-| $v1                        || —                    || 2nd fd only for pipe(2)
+| $v1                        || —                    || 2nd fd only for pipe(2)          ||
 |-
-| $a0 ... $a2/$a7 except $a3 || syscall arguments     || returned unmodified
+| $a0 ... $a2                || syscall arguments     || returned unmodified             || O32
 |-
-| $a3                        || 4th syscall argument || $a3 set to 0/1 for success/error
+| $a0 ... $a2, $a4 ... $a7   || syscall arguments     || returned unmodified             || N32 and 64
 |-
-| $t0 ... $t9                || —                    || (caller saved)
+| $a3                        || 4th syscall argument || $a3 set to 0/1 for success/error ||
 |-
-| $s0 ... $s8                || —                    || (callee saved)
+| $t0 ... $t9                || —                    || (caller saved)                   ||
 |-
-| $hi, $lo                   || —                    || (caller saved)
+| $s0 ... $s8                || —                    || (callee saved)                   ||
 |}

@@ Line 22: / Line 22: @@
 |-
 | $s0 ... $s8                || —                    || (callee saved)
 |}

← Older revision		Revision as of 08:32, 11 September 2012
Line 25:		Line 25:

	And of course a syscall is invoked through a <tt>SYSCALL</tt> instruction. The kernel assumes the syscall instruction ''not'' to be in a branch delay slot, that is, it will not check for branch delay slots and do branch emulation. Even iff it was, performance considerations would make it preferable to leave the <tt>SYSCALL</tt> instruction outside the branch delay slot. On return from a syscall the program counter will be set to the instruction following the <tt>SYSCALL</tt> instruction.		And of course a syscall is invoked through a <tt>SYSCALL</tt> instruction. The kernel assumes the syscall instruction ''not'' to be in a branch delay slot, that is, it will not check for branch delay slots and do branch emulation. Even iff it was, performance considerations would make it preferable to leave the <tt>SYSCALL</tt> instruction outside the branch delay slot. On return from a syscall the program counter will be set to the instruction following the <tt>SYSCALL</tt> instruction.
		+
		+	Up to 2.6.35 the kernel was assuming that the instruction immediately preceeding the <tt>SYSCALL</tt> instruction was reloading the syscall number and was using this for syscall restarting. This meant, to keep a syscall invokation restartable it was impossible to place a <tt>SYSCALL</tt> instruction in a branch delay slot anyway.

	== Syscall restarting ==		== Syscall restarting ==

@@ Line 19: / Line 19: @@
 | $a3                        || 4th syscall argument || $a3 set to 0/1 for success/error
 |-
-| $t0 ... $t9                ||                      || (caller saved)
+| $t0 ... $t9                || —                    || (caller saved)
 |-
-| $s0 ... $s8                ||                      || (callee saved)
+| $s0 ... $s8                || —                    || (callee saved)
 |}

@@ Line 56: / Line 56: @@
 For the exact syscall numbers for the three Linux ABI, please see [//git.linux-mips.org/?p=ralf/linux.git;a=blob;f=arch/mips/include/asm/unistd.h <asm/unistd.h>] of your kernel.
 == See also ==
 * [[MIPS OS]]
 * [[MIPS ABI History]]

@@ Line 1: / Line 1: @@
-Syscalls on
+This article attempts to describe the calling conventions of syscalls on Linux/MIPS.  Generally it's a less than great idea to directly access the system call interface.  Libraries such as glibc, ucLibc and others are doing a great job and will hide all the possible optimizations and version dependencies from the application programmer.  ''You have been warned.''
 == Syscall number ranges ==
@@ Line 19: / Line 50: @@
 | Linux N32 syscalls                                                          || 6000 || 6999
 |}
 == See also ==
 * [[MIPS OS]]
 * [[MIPS ABI History]]