Re: [PATCH v1 1/2] mm/page_idle: Add support for per-pid page_idle using virtual indexing

[Joel Fernandes <joel@joelfernandes.org>]

On Tue, Jul 23, 2019 at 08:05:25AM +0200, Michal Hocko wrote:
> [Cc linux-api - please always do CC this list when introducing a user
>  visible API]

Sorry, will do.

> On Mon 22-07-19 17:32:04, Joel Fernandes (Google) wrote:
> > The page_idle tracking feature currently requires looking up the pagemap
> > for a process followed by interacting with /sys/kernel/mm/page_idle.
> > This is quite cumbersome and can be error-prone too. If between
> > accessing the per-PID pagemap and the global page_idle bitmap, if
> > something changes with the page then the information is not accurate.
> > More over looking up PFN from pagemap in Android devices is not
> > supported by unprivileged process and requires SYS_ADMIN and gives 0 for
> > the PFN.
> > 
> > This patch adds support to directly interact with page_idle tracking at
> > the PID level by introducing a /proc/<pid>/page_idle file. This
> > eliminates the need for userspace to calculate the mapping of the page.
> > It follows the exact same semantics as the global
> > /sys/kernel/mm/page_idle, however it is easier to use for some usecases
> > where looking up PFN is not needed and also does not require SYS_ADMIN.
> > It ended up simplifying userspace code, solving the security issue
> > mentioned and works quite well. SELinux does not need to be turned off
> > since no pagemap look up is needed.
> > 
> > In Android, we are using this for the heap profiler (heapprofd) which
> > profiles and pin points code paths which allocates and leaves memory
> > idle for long periods of time.
> > 
> > Documentation material:
> > The idle page tracking API for virtual address indexing using virtual page
> > frame numbers (VFN) is located at /proc/<pid>/page_idle. It is a bitmap
> > that follows the same semantics as /sys/kernel/mm/page_idle/bitmap
> > except that it uses virtual instead of physical frame numbers.
> > 
> > This idle page tracking API can be simpler to use than physical address
> > indexing, since the pagemap for a process does not need to be looked up
> > to mark or read a page's idle bit. It is also more accurate than
> > physical address indexing since in physical address indexing, address
> > space changes can occur between reading the pagemap and reading the
> > bitmap. In virtual address indexing, the process's mmap_sem is held for
> > the duration of the access.
> 
> I didn't get to read the actual code but the overall idea makes sense to
> me. I can see this being useful for userspace memory management (along
> with remote MADV_PAGEOUT, MADV_COLD).

Thanks.

> Normally I would object that a cumbersome nature of the existing
> interface can be hidden in a userspace but I do agree that rowhammer has
> made this one close to unusable for anything but a privileged process.

Agreed, this is one of the primary motivations for the patch as you said.

> I do not think you can make any argument about accuracy because
> the information will never be accurate. Sure the race window is smaller
> in principle but you can hardly say anything about how much or whether
> at all.

Sure, fair enough. That is why I wasn't beating the drum too much on the
accuracy point. However, this surprisingly does work quite well.

thanks,

 - Joel