bpo-41586: Add pipesize parameter to subprocess & F_GETPIPE_SZ and F_SETPIPE_SZ to fcntl. (GH-21921)

* Add F_SETPIPE_SZ and F_GETPIPE_SZ to fcntl module
* Add pipesize parameter for subprocess.Popen class

This will allow the user to control the size of the pipes.
On linux the default is 64K. When a pipe is full it blocks for writing.
When a pipe is empty it blocks for reading. On processes that are
very fast this can lead to a lot of wasted CPU cycles. On a typical
Linux system the max pipe size is 1024K which is much better.
For high performance-oriented libraries such as xopen it is nice to
be able to set the pipe size.

The workaround without this feature is to use my_popen_process.stdout.fileno() in
conjuction with fcntl and 1031 (value of F_SETPIPE_SZ) to acquire this behavior.

1 files changed, 18 insertions, 1 deletions

diff --git a/Lib/subprocess.py b/Lib/subprocess.py
index 86fdf27..6a6c2fc 100644
--- a/Lib/subprocess.py
+++ b/Lib/subprocess.py
@@ -62,6 +62,11 @@ try:
     import grp
 except ImportError:
     grp = None
+try:
+    import fcntl
+except ImportError:
+    fcntl = None
+
 
 __all__ = ["Popen", "PIPE", "STDOUT", "call", "check_call", "getstatusoutput",
            "getoutput", "check_output", "run", "CalledProcessError", "DEVNULL",
@@ -756,7 +761,7 @@ class Popen(object):
                  startupinfo=None, creationflags=0,
                  restore_signals=True, start_new_session=False,
                  pass_fds=(), *, user=None, group=None, extra_groups=None,
-                 encoding=None, errors=None, text=None, umask=-1):
+                 encoding=None, errors=None, text=None, umask=-1, pipesize=-1):
         """Create new Popen instance."""
         _cleanup()
         # Held while anything is calling waitpid before returncode has been
@@ -773,6 +778,11 @@ class Popen(object):
         if not isinstance(bufsize, int):
             raise TypeError("bufsize must be an integer")
 
+        if pipesize is None:
+            pipesize = -1  # Restore default
+        if not isinstance(pipesize, int):
+            raise TypeError("pipesize must be an integer")
+
         if _mswindows:
             if preexec_fn is not None:
                 raise ValueError("preexec_fn is not supported on Windows "
@@ -797,6 +807,7 @@ class Popen(object):
         self.returncode = None
         self.encoding = encoding
         self.errors = errors
+        self.pipesize = pipesize
 
         # Validate the combinations of text and universal_newlines
         if (text is not None and universal_newlines is not None
@@ -1575,6 +1586,8 @@ class Popen(object):
                 pass
             elif stdin == PIPE:
                 p2cread, p2cwrite = os.pipe()
+                if self.pipesize > 0 and hasattr(fcntl, "F_SETPIPE_SZ"):
+                    fcntl.fcntl(p2cwrite, fcntl.F_SETPIPE_SZ, self.pipesize)
             elif stdin == DEVNULL:
                 p2cread = self._get_devnull()
             elif isinstance(stdin, int):
@@ -1587,6 +1600,8 @@ class Popen(object):
                 pass
             elif stdout == PIPE:
                 c2pread, c2pwrite = os.pipe()
+                if self.pipesize > 0 and hasattr(fcntl, "F_SETPIPE_SZ"):
+                    fcntl.fcntl(c2pwrite, fcntl.F_SETPIPE_SZ, self.pipesize)
             elif stdout == DEVNULL:
                 c2pwrite = self._get_devnull()
             elif isinstance(stdout, int):
@@ -1599,6 +1614,8 @@ class Popen(object):
                 pass
             elif stderr == PIPE:
                 errread, errwrite = os.pipe()
+                if self.pipesize > 0 and hasattr(fcntl, "F_SETPIPE_SZ"):
+                    fcntl.fcntl(errwrite, fcntl.F_SETPIPE_SZ, self.pipesize)
             elif stderr == STDOUT:
                 if c2pwrite != -1:
                     errwrite = c2pwrite