[Bug 576590] Re: Command not found consumes lots of memory and CPU for longer input

[Ricardo Cabral]

I temporarily fixed my own installation of cnf by changing line 89 of /usr/lib/command-not-found from:
    if len(args) == 1:
to
    if len(args) == 1 and len(args[0])<100:

That means cnf will not do anything for commands longer than 100
characters. But seriously, can anyone tell me a command that has 100
characters?

This bug can be used by an attacker to bring down a system to its knees.

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to command-not-found in Ubuntu.
https://bugs.launchpad.net/bugs/576590

Title:
  Command not found consumes lots of memory and CPU for longer input

Status in command-not-found handler:
  Confirmed
Status in “command-not-found” package in Ubuntu:
  Confirmed

Bug description:
  Binary package hint: command-not-found

  This concerns command-not-found 0.2.40ubuntu5 in Ubuntu Lucid.

  Once upon a time, I was working on my shiny new Lucid system, when
  things suddenly became very slow. I ran top(1) to check what was going
  on, and saw this:

  top - 13:34:17 up 15 min,  5 users,  load average: 3.79, 1.41, 0.54
  Tasks: 158 total,   2 running, 156 sleeping,   0 stopped,   0 zombie
  Cpu(s):  0.6%us,  1.5%sy,  0.0%ni, 15.8%id, 82.1%wa,  0.0%hi,  0.0%si,  0.0%st
  Mem:   3917500k total,  3890004k used,    27496k free,      480k buffers
  Swap:  4200956k total,  1919184k used,  2281772k free,    50180k cached

    PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND            
   2348 userme    20   0 4618m 3.2g  440 D    3 86.7   0:06.95 python             
   2370 userme    20   0 11000 1308  956 S    1  0.0   0:00.03 top                
   1587 root      20   0  101m 5788 1524 S    0  0.1   0:04.48 Xorg               

  Python was going out of control! What was it doing?

  # ps uww 2348
  USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
  userme    2348  6.5 84.6 4729240 3316468 tty1  D    13:32   0:07 /usr/bin/python /usr/lib/command-not-found -- mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm...

  (The m's went on for like two or three screens.)

  Now, I have no idea what caused this bizarre invocation of c-n-f. (I
  don't typically sleep with my forehead on the keyboard or anything.)
  However, I copied-and-pasted the complete invocation as reported by
  ps(1) into a script, and tried running it:

  $ time -p sh c-n-f-bug.sh 
  mmmm[...]mmmm: command not found
  real 42.60
  user 39.17
  sys 1.10

  During that time, c-n-f used over a gigabyte of memory. That's not
  right!

  I am attaching the c-n-f-bug.sh script. Interestingly, it is exactly
  4096 bytes long, which suggests to me that the original command
  invocation may have been even longer and not reported in full simply
  due to ps(1) limiting the output.

To manage notifications about this bug go to:
https://bugs.launchpad.net/command-not-found/+bug/576590/+subscriptions