Re: write_lock_irq(&tasklist_lock)

[Will Deacon]

On Wed, May 23, 2018 at 09:26:35AM -0700, Linus Torvalds wrote:
> On Wed, May 23, 2018 at 8:35 AM Will Deacon <will.deacon@xxxxxxx> wrote:
> 
> > In other words, qrwlock requires consistent locking order wrt spinlocks.
> 
> I *thought* lockdep already tracked and detected this. Or is that only with
> with the sleeping versions?

There are patches in-flight to detect this:

https://marc.info/?l=linux-kernel&m=152483640529740&w=2k

as part of Boqun's work into recursive read locking.

Will