Interrupt spawns Zombies and Vampires
Peter Renzland
peter at dancing.org
Wed Jan 14 08:28:40 UTC 2009
Denys, thanks for thinking about my problem report. I'm not a
developer, just a sys admin. Unfortunately I don't have a cross-
compiler and related tools. I came to this Busybox forum because I
thought that I would get competent answers (I am :-), and in the hope
that this might be a known problem (it isn't).
I will happily forward the config information, etc., to the Tomato
developer(s). ... more in line ...
On 09 Jan 13, at 21:31 , Denys Vlasenko wrote:
> On Tuesday 13 January 2009 03:37, Peter Renzland wrote:
>> OK. Attaching the config_g
>
> I see CONFIG_CROSS_COMPILER_PREFIX="mipsel-uclibc-"
>
> Can you build native i386 or x86_64 busybox with the same .config
> (just set CONFIG_CROSS_COMPILER_PREFIX="")?
> Does it still exhibit this?
I cannot do so myself.
BTW, I have a tiny Aztech DSL695EU modem/router with a very old
Busybox, which works fine. It even has job control:
BusyBox v0.61.pre (2008.10.10-08:51+0000) Built-in shell (ash)
> Which libc do you use on your embedded device?
Not sure. I see references to mipsel-uclibc in the Makefile.
I also see this, in ash.c:
#if defined(__GLIBC__) && __GLIBC__ == 2 && __GLIBC_MINOR__ < 1
typedef enum __rlimit_resource rlim_t;
#endif
BTW, for anyone interested:
http://www.polarcloud.com/f/Tomato_1_23.7z
>> Shell is ash:
>>
>> BusyBox v1.12.2 (2008-12-05 06:20:11 PST) built-in shell (ash)
>> also
>> BusyBox v1.12.3 (2008-12-14 02:54:58 PST) built-in shell (ash)
>
>
> # ### it says ash
> # ps
> PID USER VSZ STAT COMMAND
> ...
> 21313 root 1984 S -sh
> 21315 root 1960 R ps
> # ### but ps says sh
>
> So what? your /bin/sh is a symlink to ash, thus /bin/sh is ash.
> I see no problem here.
I understand, but you did ask and I didn't want to keep the
information from you. After all, commands can and do act differently,
depending on the name by which they are invoked.
>> # ping ac.ca
> PING ac.ca (69.90.47.27): 56 data bytes
> 64 bytes from 69.90.47.27: seq=0 ttl=43 time=132.217 ms
> 64 bytes from 69.90.47.27: seq=1 ttl=43 time=132.336 ms
> #
> --- ac.ca ping statistics ---
> 2 packets transmitted, 2 packets received, 0% packet loss
> round-trip min/avg/max = 132.217/132.276/132.336 ms
>
> # ### SIGINT ^C
> # ps
> # ### vampire shell eats my command ...
> # ps
> PID USER VSZ STAT COMMAND
> 1 root 1720 S init noinitrd
> 2 root 0 SW [keventd]
>
>
>
> Mine doesn't:
>
> # ping ac.ca
> PING ac.ca (69.90.47.27): 56 data bytes
> 64 bytes from 69.90.47.27: seq=0 ttl=47 time=181.496 ms
> 64 bytes from 69.90.47.27: seq=1 ttl=47 time=180.338 ms
> ^C
> --- ac.ca ping statistics ---
> 3 packets transmitted, 2 packets received, 33% packet loss
> round-trip min/avg/max = 180.338/180.917/181.496 ms
Please understand that I am reporting and providing evidence for what
clearly is a problem. You and I both know perfectly well that *this*
particular Busybox implementation is defective. So, of course, any
correct implementation won't have these problems. I'm hoping to learn
what causes them, so that I might send that information to the
developer and ask him to fix it. :-)
> # ps
> PID USER TIME COMMAND
> 1 root 0:00 /bin/sh /init
> 2 root 0:00 [kthreadd]
> 3 root 0:00 [migration/0]
>
>
>
> Having said that, these is a few differences in your and my setup:
>
> * I use native build against glibc
> * I use svn, which have a few fixes in ash.
>
> I will mail you latest terball separately. Please try it and let me
> know.
> --
> vda
Thanks!
More information about the busybox
mailing list