GNU bug report logs - #36455
Question on reading from socket(pair) instead from pipe

Previous Next

Package: grep;

Reported by: "Dr. Werner Fink" <werner <at> suse.de>

Date: Mon, 1 Jul 2019 06:23:02 UTC

Severity: normal

Done: Paul Eggert <eggert <at> cs.ucla.edu>

Bug is archived. No further changes may be made.

Full log


Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: "Dr. Werner Fink" <werner <at> suse.de>
To: bug-grep <at> gnu.org
Subject: Question on reading from socket(pair) instead from pipe
Date: Mon, 1 Jul 2019 08:22:36 +0200
[Message part 1 (text/plain, inline)]
Hi,

with grep 3.3 I see for a FIFO this

  echo XXX | strace grep -E XXX
  [...]
  fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(0x88, 0xc), ...}) = 0
  stat("/dev/null", {st_mode=S_IFCHR|0666, st_rdev=makedev(0x1, 0x3), ...}) = 0
  fstat(0, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
  lseek(0, 0, SEEK_CUR)                   = -1 ESPIPE (Illegal seek)
  read(0, "XXX\n", 98304)                 = 4
  fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(0x88, 0xc), ...}) = 0
  write(1, "XXX\n", 4XXX)                    = 4
  read(0, "", 98304)                      = 0
  close(1)                                = 0
  close(2)                                = 0
  exit_group(0)                           = ?
  +++ exited with 0 +++

prefect but with libpipeline configured to use much faster socketpairs I see
this:

  strace -f ./mgrep -e XXX
  clone(strace: Process 2020 attached
  child_stack=NULL, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x7fb632fb4a10) = 2020
  [...]
  [pid  2020] execve("/usr/bin/grep", ["grep", "-e XXX"], 0x7ffd31f42ea8 /* 128 vars */ <unfinished ...>
  [...]
  [pid  2020] fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(0x88, 0xc), ...}) = 0
  [pid  2020] stat("/dev/null", {st_mode=S_IFCHR|0666, st_rdev=makedev(0x1, 0x3), ...}) = 0
  [pid  2020] fstat(0, {st_mode=S_IFSOCK|0400, st_size=0, ...}) = 0
  [pid  2020] lseek(0, 0, SEEK_CUR)       = -1 ESPIPE (Illegal seek)
  [pid  2020] read(0, "XXX\n", 98304)     = 4
  [pid  2020] read(0, "", 98304)          = 0
  [pid  2020] close(1)                    = 0
  [pid  2020] close(2)                    = 0
  [pid  2020] exit_group(1)               = ?
  [pid  2020] +++ exited with 1 +++

IMHO reading data from a FIFO/pipe or from a FSOCK/socket(pair) should result
in the same result.  The mgrep is a simple test C code which uses libpipeline
to write "XXX\n" to a FILE stream used on the input fd of the sub process for
the grep command:

  FILE *xy;
  int fd = dup(1);
  pipeline *p = pipeline_new();
  pipeline_want_in(p, -1);
  pipeline_want_out(p, fd);
  pipecmd *grep = pipecmd_new("/usr/bin/grep");
  [... handle options and arguments ...]
  xy = pipeline_get_infile (p);
  fprintf(xy, "XXX\n");
  fflush(xy);
  fclose(xy);

Nevertheless in both cases stdout for grep is 136/12 aka /dev/pts/12

Werner

-- 
  "Having a smoking section in a restaurant is like having
          a peeing section in a swimming pool." -- Edward Burr
[signature.asc (application/pgp-signature, inline)]

This bug report was last modified 5 years and 328 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.