Skip to contentSkip to conversation

OOM aware applications

RFC BALBIR SINGH 3 participants 3 messages Sep 25, 2001

lore
Share & tools
lore
#0 lore raw

I was looking at the code in oom_kill.c, I had the following suggestions

Warning: I am not aware of what was discussed earlier about these issues or if they were discusses at all.

1. I was wondering that instead of killing the application using oom_kill_task() directly, should the OOM issue some kind of a warning by sending a signal (some signal with
si_code set to a value indicating that
the application is causing memory to run out). Then, wait for a
while and then see if the application is still
misbehaving and if so kill it.

2. Minor changes suggested

In the code, I see

points *= 2 and points /= 4 in a few places, recommend changing them to

points <<= 1 and points >>= 2 to help the compiler generate better code

Dialog between the kernel and an OOM aware application

on OOM condition

kernel to the application: Hey u are running the system out of memory
application: Sorry my lord, I commited a blunder and will rectify myself

application rectifies itself and everybody lives happily everafter application is stuborn

kernel: You must go now, u have been unfair in your demands and have caused a lot of problem
application: Does not get to say anything, it is zapped

Comments, Suggestions, Criticism
Balbir

----------------------------------------------------------------------------------------------------------------------
Information transmitted by this E-MAIL is proprietary to Wipro and/or its Customers and is intended for use only by the individual or entity to which it is addressed, and may contain information that is privileged, confidential or exempt from disclosure under applicable law. If you are not the intended recipient or it appears that this mail has been forwarded to you without proper authority, you are notified that any use or dissemination of this information in any manner is strictly prohibited. In such cases, please notify us immediately at mailto:[email protected] and delete this mail from your records.
----------------------------------------------------------------------------------------------------------------------

#1 lore raw
Quoted 5 lines · BALBIR SINGH

In the code, I see

points *= 2 and points /= 4 in a few places, recommend changing them to

points <<= 1 and points >>= 2 to help the compiler generate better code

Does your compiler really write different code for *= 2 and <<= 1? If so, try -O :-)

My gcc writes addl %eax,%eax for either case.

Signature · 4 lines
Matthias Andree

"Those who give up essential liberties for temporary safety deserve
neither liberty nor safety." - Benjamin Franklin
#2 lore raw
In reply to BALBIR SINGH
Quoted 2 lines · BALBIR SINGH

Warning: I am not aware of what was discussed earlier about these issues or if they were discusses at all.

They have been discussed many times. Please search for SIGDANGER in the linux-kernel mailing list archive.

Quoted 6 lines

1. I was wondering that instead of killing the application using oom_kill_task() directly, should the OOM issue some kind of a warning by sending a signal (some signal with si_code set to a value indicating that the application is causing memory to run out). Then, wait for a while and then see if the application is still misbehaving and if so kill it.

This is exactly what SIGDANGER is.

Cheers, Andreas
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert