> Put it into the main kernel and make sys_sendfile() use it - it's *MUCH*
> better than expanding the export list and duplicating code.
Ehm. sys_sendfile is for fd->fd, mine is for file -> socket. I don't know
if I'm able to merge some of this... (And I don't think Linus would like
this)
> >
> > "sendfile" holds the global kernel-lock. Is it possible to remove this
> > somehow?
>
> No. It uses a lot of code that is not SMP-safe. BTW, the same applies to
> your sendfile-from-struct-file thing.
My version also grabs the kernel-lock, because I didn't know if it was
allowed NOT to use it for this special case.
> Which is one of the reasons *not* to
> drag it out of the main kernel - it would make analysis of SMP issues in
> VFS and filesystems much harder.
I don't think Linus will accept a http-server in the main kernel. But who
knows... It is a nice tool to pinpoint bottlenecks in the kernel.
Greetings,
Arjan van de Ven
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/