Re: [uml-devel] Memoryleak in the RAM-Image in /tmp

[Johannes F. <joh...@fo...>]

At 16:17 Uhr +0100 12.03.2003, Johannes Formann wrote:
>>The only other time I've seen behaviour like this is when UML didn't close
>>file descriptors when it rebooted.  The /tmp files from the previous boot
>>were kept open and they continued to occupy /tmp.  That was fixed, and
>>AFIAK, remains fixed.
>
>I think it's no fixed completly
>
>I'd bee trying around a bit:
>
>creating a tmp for testing
>
>tmpfs                 501M     0  501M   0% /root/uml-test/tmp-test
>
>well, everything looks fine
>
>starting the UML
>nohup nice -10 su operator -s /bin/bash -c 
>/usr/bin/uml-2.4.19-50-skas uml_dir=/root/uml-test/tmp-test 
>ubd0=standarduml mem=34M eth0=tuntap,tap82 con=pty con0=fd:0,fd:1 
>umid=operator >/root/uml-test/uml.log </dev/null &
>
>after staring
>
>tmpfs                 501M   11M  490M   3% /root/uml-test/tmp-test
>
>well, lets fill the momory inside the UML
>
>tmpfs                 501M   33M  468M   7% /root/uml-test/tmp-test
>
>
>now let's try an reboot
>
>tmpfs                 501M   42M  459M   9% /root/uml-test/tmp-test
>
>fillig the RAm
>
>tmpfs                 501M   65M  436M  13% /root/uml-test/tmp-test

...

Could anyone reproduce this Bug?

Or is it fixed with -51 oder 2.4.20-1?

Greetings Johannes