Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-02 Thread Thomas Gleixner
On Mon, 2 Oct 2017, Don Zickus wrote: > On Mon, Oct 02, 2017 at 07:32:57PM +, Thomas Gleixner wrote: > > static void softlockup_reconfigure_threads(void) > > There is a second copy of , you will need to add identical locking there > too. Good catch, but then we might move it further out

Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-02 Thread Don Zickus
On Mon, Oct 02, 2017 at 07:32:57PM +, Thomas Gleixner wrote: > On Mon, 2 Oct 2017, Linus Torvalds wrote: > > Side note: would it perhaps make sense to have that > > cpus_read_lock/unlock() sequence around the whole reconfiguration > > section? > > > > Because while looking at that sequence, it

Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-02 Thread Thomas Gleixner
On Mon, 2 Oct 2017, Linus Torvalds wrote: > Side note: would it perhaps make sense to have that > cpus_read_lock/unlock() sequence around the whole reconfiguration > section? > > Because while looking at that sequence, it looks a bit odd to me that > cpu's can come and go in the middle of the nmi

Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-02 Thread Linus Torvalds
On Mon, Oct 2, 2017 at 11:46 AM, Thomas Gleixner wrote: > > I agree that adding that 'run' argument was certainly not a piece of > art. Though I disagree with the sentiment that non-functional garbage is > preferrable over functionally correct code which merily contains a bad > implementation choi

Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-02 Thread Thomas Gleixner
On Sun, 1 Oct 2017, Linus Torvalds wrote: > So get rid of that kind of shit, and I may reconsider. But as is, I > look at that patch and say "no, this is worse than the garbage it used > to be". I agree that adding that 'run' argument was certainly not a piece of art. Though I disagree with the se

Re: [RFC GIT Pull] core watchdog sanitizing

2017-10-01 Thread Linus Torvalds
I refuse to pull this. Look, I understand what you want to do, but the code is disgusting. Maybe most of it is fine, but I just couldn't stomach looking at it after just a few lines. Look at that abortion called "watchdog_nmi_reconfigure()". It's one single function that does two completely dif

[RFC GIT Pull] core watchdog sanitizing

2017-10-01 Thread Thomas Gleixner
Linus, please consider to pull the latest core-watchdog-for-linus git tree from: git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git core-watchdog-for-linus The watchdog (hard/softlockup detector) code is pretty much broken in its current state. The patch series addresses this by remov