RE: spinlocks() are severely broken in 2.2.X and 2.4.X for modules

From: Chris Lattner (sabre@skylab.org)
Date: Fri Jun 30 2000 - 20:47:32 EDT

Next message: Andi Kleen: "Re: Arp Timeout Problem, help needed."

Previous message: Jeff V. Merkey: "Re: Patented algorithms in kernel"
Maybe in reply to: Jeff V. Merkey: "spinlocks() are severely broken in 2.2.X and 2.4.X for modules"
Next in thread: Gérard Roudier: "Re: spinlocks() are severely broken in 2.2.X and 2.4.X for modules"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Okay, one question, one comment.

I too am having "issues" with linux spinlocks... but in a massively
parrallel situation. It appears that something isn't working right and I
don't think it's _my_ code. ;) I was wondering what the incentive is for
moving code to a different section (besides that it totally confuses my
debugger! :)... Why not just leave it inline? Trying to make fall
through happen?

Okay, my comment is that there is a lot of errata on the pentium line (and
others I'm sure) with reguards to caching and the lock prefix. Have you
tried just adding a lock prefix to the spin_unlock "movb" instruction?

-Chris

-
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/

Next message: Andi Kleen: "Re: Arp Timeout Problem, help needed."
Previous message: Jeff V. Merkey: "Re: Patented algorithms in kernel"
Maybe in reply to: Jeff V. Merkey: "spinlocks() are severely broken in 2.2.X and 2.4.X for modules"
Next in thread: Gérard Roudier: "Re: spinlocks() are severely broken in 2.2.X and 2.4.X for modules"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

This archive was generated by hypermail 2b29 : Fri Jun 30 2000 - 20:51:21 EDT