433
edits
Changes
mLine 10:
Line 10: + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
no edit summary
== Basic design ==
== Basic design ==
=== Checkpoint ===
The checkpoint procedure relies heavily on '''/proc''' file system (it's a general place where crtools takes all the information it needs).
Which includes
* Files descriptors information (via '''/proc/$pid/fd''' and '''/proc/$pid/fdinfo''').
* Pipes parameters.
* Memory maps (via '''/proc/$pid/maps''').
The process dumper (lets call it simply the dumper further) does the following steps during checkpoint stage
# A '''$pid''' of a process group leader is obtained from the command line.
# By using this '''$pid''' the dumper walks though '''/proc/$pid/status''' and gathers children '''$pids''' recursively. At the end we will have a process tree.
# Then it takes every '''$pid''' from a process tree, sends ''SIGSTOP'' to every process found, and performs the following steps on each '''$pid'''.
#* Collects VMA areas by parsing '''/proc/$pid/maps'''.
#* Seizes a task via relatively new ptrace interface. Seizing a task means to put it into a special state when the task have no idea if it's being operated by ptrace.
#* Core parameters of a task (such as registers and friends) are being dumped via ptrace interface and parsing '''/proc/$pid/stat''' entry.
#* The dumper injects a parasite code into a task via ptrace interface. This allows us to dump pages of a task right from within the task's address space.
#** An injection procedure is pretty simple - the dumper scans executable VMA areas of a task (which were collected previously) and tests if there a place for <code>syscall</code> call, then (by ptrace as well) it substitutes an original code with <code>syscall</code> instructions and creates a new VMA area inside process address space.
#** Finally parasite code get copied into the new VMA and the former code which was modified during parasite bootstrap procedure get restored.
#* Then (by using a parasite code) the dumper flushes contents of a task's pages to the file. And pulls out parasite code block completely, since we don't need it anymore.
#* Once parasite removed a task get unseized via ptrace call but it remains stopped still.
#* The dumper writes out files and pipes parameter and data.
# The procedure continues for every '''$pid'''.
=== Restore ===
The restore procedure (aka restorer) proceed in the following steps
# A process tree has been read from a file.
# Every process started with saved (i.e. original) '''$pid''' via <code>clone()</code> call with new <code>CLONE_CHILD_USEPID</code> flag.
# Files and pipes are restored (by restored it's meant - they are opened and positioned).
# A new file generated. The file has an Elf format but with modified executable and program header types (telling the kernel that this particular file is not a regular Elf'oid but rather the kernel is to handle it in a slightly different way).
# Finally execve with new Elf file as an argument is executed, which initiate the kernel's stage of restore procedure.
=== Kernel requirements ===
Since checkpoint and restore processes require some help from the Linux kernel, the following kernel patches are needed
* procfs-report-eisdir-when-reading-sysctl-dirs-in-proc.patch
* proc-fix-races-against-execve-of-proc-pid-fd.patch
* proc-fix-races-against-execve-of-proc-pid-fd-fix.patch
* proc-force-dcache-drop-on-unauthorized-access.patch
* cr-statfs-callback-for-pipefs
These patches are already in -mm tree and rather a preparation patches for the next series.
* fs-proc-switch-to-dentry
* cr-proc-map-files-21
These patches introduce '''/proc/$pid/msp_files'''.
* cr-clone-with-pid-support
This one introduce ability to clone process with specified pid.
* cr-proc-add-children
This one introduce "Children" line to '''/proc/$pid/status'''.
* fs-add-do-close
* fs-proc-add-tls
* fs-proc-add-mm-task-stat
These ones provides missing pieces of process' information which is needed for checkpoint/restore.
* binfmt-elf-for-cr-5
This one provides new Elf file format.
=== Where to get '''crtools''' itself ===
The '''crtools''' utility itself is hosted at [https://github.com/cyrillos/crtools github]. Clone this repo to test new functionality out. Note the kernel patches are placed at kernel/ directory inside source code tree and includes [http://savannah.nongnu.org/projects/quilt quilt] series file.
'''crtools''' has been tested on Linux 3.1-rc3.