Hi,


----- Original Message ----
> From: Serge E. Hallyn <[email protected]>
> To: Kamalesh Babulal <[email protected]>
> Cc: CAI Qian <[email protected]>; [email protected]; [email protected]; 
> [email protected]; [email protected]
> Sent: Saturday, January 31, 2009 12:54:42 AM
> Subject: Re: [LTP] [PATCH] proc01: SELinux with attr/* Interface - version 2
> 
> Quoting Kamalesh Babulal ([email protected]):
> > selinux enabled
> > ----------------
> > proc01      0  INFO  :  /proc/acpi/event: open: known issue: Device or 
> resource busy
> > proc01      0  INFO  :  /proc/sys/net/ipv6/route/flush: is write-only.
> > proc01      0  INFO  :  /proc/sys/net/ipv4/route/flush: is write-only.
> > proc01      0  INFO  :  /proc/sys/fs/binfmt_misc/register: is write-only.
> > proc01      0  INFO  :  /proc/sysrq-trigger: is write-only.
> > proc01      0  INFO  :  /proc/kmsg: read: known issue: Resource temporarily 
> unavailable
> > proc01      0  INFO  :  /proc/self/task/2875/mem: read: known issue: 
> Input/output error
> > proc01      0  INFO  :  /proc/self/mem: read: known issue: Input/output 
> > error
> > proc01      1  PASS  :  readproc() completed successfully, total read: 
> > 1096865 
> bytes, 885 objs
> > 
> > the EINVAL is returned only when the LSM is does not support the
> > interface, and found_errno() checks for the know return value or else
> > it handled the way the unknow error is hanlded.
> 
> Right, but I think CAI is concerned that if there is a regression with
> selinux enabled and it mistakenly returns -EINVAL this won't catch it.
> 

That is correct. In addition, if it happens, there will be some false positive
and bad messages like,

proc01 0 INFO: /proc/self/attr/exec: read: known issue: Invalid argument

> As Stephen pointed out, if that happens then you likely won't get a
> successful boot to begin with...
> 

Perhaps,  but by testing those entries the same way as other procfs entries
here,  we can catch EINVAL in same scenarios and facilities the test
provided. For example, using different sizes of read buffers .

CAI Qian

> thanks,
> -serge


------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to