On 15/08/17 12:06, Boaz Harrosh wrote:

On 14/08/17 19:03, Dan Williams wrote:

quoted

On Mon, Aug 14, 2017 at 7:04 AM, Boaz Harrosh [off-list ref] wrote:

quoted

Thank you Jan, I'm patiently waiting for this MAP_SYNC flag since I asked for
it in 2014. I'm so glad its time is finally do.

<>

quoted

quoted

4] Once we have this flag. And properly implemented at least in one FS
   and optionally in /dev/pmemX we no longer have any justification for
   /dev/daxX and it can die a slow and happy death.

I'm all for replacing /dev/dax with filesystem equivalent
functionality, but I don't think MAP_SYNC gets us fully there. That's
what the MAP_DIRECT proposal [1] is meant to address.

OK This is true. Could you please summarise for us the exact semantics of both
proposed flags?

That said I think the big difference is the movability of physical blocks
underneath the mmap mapping. Now for swap files that is a problem. because
of the deadlocks that can happen with memory needed if blocks start moving.
But for an application like nvml? why does it care. Why can't an nvml image file
not be cloned and COWed underneath the NVM application transparently.

OK Sorry didn't do my homework. Struck this out, it is all about RDMA and friends
from an "immutable" file.

I'll go dig into this now. 

Thanks
Boaz

Sorry for being slow but I don't see why you need MAP_DIRECT from user-mode
If you have MAP_SYNC. Please advise

(not that the immutable patchset is not a very needed fixing)

quoted

[1]: https://lkml.org/lkml/2017/8/13/160

Thanks
Boaz