sys_times() return value

sys_times() return value

[Russell King]

Guys,

ARM folk have recently pointed out a problem with sys_times().
When the kernel boots, we set jiffies to -5 minutes.  This causes
sys_times() to return a negative number, which increments through
zero.

However, some negative numbers are used to return error codes.
Hence, there's a period of time when sys_times() returns values
which are indistinguishable from error codes shortly after boot.

This probably only affects 32-bit architectures.  However, one
wonders whether sys_times() needs force_successful_syscall_return().

Also, it appears that glibc does indeed interpret the return value
from sys_times in the way I describe above on at least ARM and x86.
Other architectures may be similarly affected.  Hopefully the ARM
glibc folk will raise a cross-architecture bug in glibc for this.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core

Re: sys_times() return value

[Andrew Morton]

Russell King <rmk+lkml@arm.linux.org.uk> wrote:
>
> Guys,
> 
> ARM folk have recently pointed out a problem with sys_times().
> When the kernel boots, we set jiffies to -5 minutes.  This causes
> sys_times() to return a negative number, which increments through
> zero.
> 
> However, some negative numbers are used to return error codes.
> Hence, there's a period of time when sys_times() returns values
> which are indistinguishable from error codes shortly after boot.

What a strange system call.

> This probably only affects 32-bit architectures.  However, one
> wonders whether sys_times() needs force_successful_syscall_return().

I'd say so, yes.  But lots of architectures seem to have a no-op there.

> Also, it appears that glibc does indeed interpret the return value
> from sys_times in the way I describe above on at least ARM and x86.
> Other architectures may be similarly affected.  Hopefully the ARM
> glibc folk will raise a cross-architecture bug in glibc for this.
> 

Re: sys_times() return value

[Russell King]

On Sat, Jul 23, 2005 at 03:35:31PM +1000, Andrew Morton wrote:
> Russell King <rmk+lkml@arm.linux.org.uk> wrote:
> > ARM folk have recently pointed out a problem with sys_times().
> > When the kernel boots, we set jiffies to -5 minutes.  This causes
> > sys_times() to return a negative number, which increments through
> > zero.
> > 
> > However, some negative numbers are used to return error codes.
> > Hence, there's a period of time when sys_times() returns values
> > which are indistinguishable from error codes shortly after boot.
> 
> What a strange system call.
> 
> > This probably only affects 32-bit architectures.  However, one
> > wonders whether sys_times() needs force_successful_syscall_return().
> 
> I'd say so, yes.  But lots of architectures seem to have a no-op there.

As I mentioned below, these other architectures need glibc to be fixed.

> > Also, it appears that glibc does indeed interpret the return value
> > from sys_times in the way I describe above on at least ARM and x86.
> > Other architectures may be similarly affected.  Hopefully the ARM
> > glibc folk will raise a cross-architecture bug in glibc for this.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core