Changes

1,702 bytes added ,  00:27, 1 September 2016
created (not finished)
Line 1: Line 1:  +
This articles describes some intricacies of handling shared memory mappings, i.e. mappings that are shared between a few processes.
 +
 +
== Checkpoint ==
 +
 +
Every process has one or more mmaped files. Some mappings (for example, ones of shared libraries)
 +
are shared between a few processes. During the checkpointing, CRIU need to figure out
 +
all the mappings that are shared in order to dump them as such.
 +
 +
It does so by performing <code>fstatat()</code> for each entry in <code>/proc/$PID/map_files/</code>,
 +
noting the ''device'' and ''inode'' fields of the structure returned by fstatat(). This information
 +
is collected and sorted. Now, if any few processes have a mapping with same ''device'' and ''inode'',
 +
this mapping is a shared one and should be dumped as such.
 +
 +
== Restore ==
 +
 +
Upon restore, CRIU already knows which mappings are shared, and the trick is to restore them as such.
 +
For that, two different approaches are used, depending on the availability.
 +
 +
The common part is, between the processes sharing a mapping, the one with the lowest PID
 +
among the group performs the actual <code>mmap()</code>, while all the others wait
 +
for the mapping to appear and, once it's available, use it.
 +
 +
=== memfd ===
 +
 +
Linux kernel v3.17 adds a [http://man7.org/linux/man-pages/man2/memfd_create.2.html memfd_create()]
 +
syscall. CRIU restore checks if it is available from the running kernel; it yes, it is used.
 +
 +
FIXME how
 +
 +
=== /proc/$PID/map_files/ ===
 +
 +
This method is used if memfd is not available. The limitation is, /proc/$PID/map_files/ is not available
 +
for users inside user namespaces (due to security concerns), so it's not possible to use it if there
 +
are any user namespaces in the dump.
 +
 +
FIXME how
 +
 +
===
 +
 
[[Category:Memory]]
 
[[Category:Memory]]
 
[[Category:Under the hood]]
 
[[Category:Under the hood]]
 
[[Category:Empty articles]]
 
[[Category:Empty articles]]