[lkml]   [2000]   [Aug]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: SCO: "thread creation is about a thousand times faster than on
On Mon, Aug 28, 2000 at 12:37:06AM -0400, Albert D. Cahalan wrote:
> > It's silly to expect to "fork" or to "exec" via a file system
> > in a minimal realtime environment. These are complex activities
> > however you cut it.
> The "exec" is very easy. You don't need a real filesystem.
> Your executable names can be a compiled-in table that maps
> from string to function pointer.

That's why I wrote: " "exec" via a file system". But exec without
a real fs is a pointless feature-- if you have pthread_create and the code
in memory already, the only possible use of exec is to do something
with file type resources and we just said we didn't have them.

> The "fork" is easy with your choice of position-independent
> code, real swapping, x86-style segments, or a real MMU.
> You'd use whatever works fastest on the available hardware.
> Don't have a COW.

Again, it's a possible, but pointless feature. You want fork/exec
in the presence of a file system and memory protection. But since
RTLinux has all that nice stuff available from Linux, there is little
point in duplicating it in the RT environment.

All that said, I reserve the right to add fork/exec to RTLinux sometime
if we see applications that really could make use of it. But my
original point remains: the 1003.13 spec allow us to take a coherent
subset of POSIX that makes sense in the minimal RT environment.

Victor Yodaiken
Finite State Machine Labs: The RTLinux Company.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 12:38    [W:0.088 / U:0.116 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site