On Mon, Oct 03, 2016 at 03:20:13PM +0300, Jarkko Sakkinen wrote:
> On Sun, Oct 02, 2016 at 03:25:51PM -0600, Jason Gunthorpe wrote:
> > On Sat, Oct 01, 2016 at 10:32:39PM +0300, Jarkko Sakkinen wrote:
> > > > chip held for the duration of any fops. Can open() start after the
> > > > kref is dropped, etc.
> > > 
> > > Why not make tpm_securityfs_data refcounted in order to remove
> > > binding to the chip?
> > 
> > The chip is already kref'd. How does swapping one kref for another
> > solve anything?
> > 
> > The possible issue is that the krefs are not covering
> > the right code.
> > 
> > The scheme you suggested is also way off the mark for how fops works,
> > fops->close has no relation to the needed duration for 'data', the
> > duration is related to securityfs_remove.
> 
> Right, the above would not work because it's not linked to
> securityfs_remove by any means.
> 
> Are you trying to say that after securityfs_remove() there might be a
> "grace period" when user space could still see the files visible and
> open them?

You have to provide something factors more concrete. Otherwise,
I'm inclined to accept the approach in Naynas patch. It's an
improvement.

/Jarkko

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most 
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
tpmdd-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tpmdd-devel

Reply via email to