Re: [rohc] COPY-LITERAL and COPY-OFFSET torture test

[Cristian CONSTANTIN <[e-mail]>]

On Wed, Jul 06, 2005 at 07:11:47PM +0200, Cristian CONSTANTIN wrote:
> On Tue, Jul 05, 2005 at 03:45:34PM +0100, Surtees, Abigail wrote:
> > Hi Cristian,
> > 
> > You are quite right.  The offset value of 1 is not used and this test does require a large amount of memory.  There isn't really anyway for the interpreter to guess the amount of memory needed.  If a large amount of memory is needed then it should be written in the draft as part of the test.
> > 
> > However, I think this test could be done without needing an extra large amount of memory.  I've uploaded an updated version of the draft with a new test at <http://sigcomp.srmr.co.uk/~ahs/draft-ietf-rohc-sigcomp-torture-tests-01.txt> (see my previous email).  I'd appreciate any comments, particularly on the COPY and COPY-LITERAL and COPY-OFFSET tests.
> 
> cristian: hi! I have run the COPY and COPY-LITERAL and COPY-OFFSET tests;
> got the same output as you did. no memory problems this time.
> 
> I find the comments really helpful.

cristian: I mean the comments inside the udvm code examples.
> 
> bye now!
> cristian
> > 
> > Best regards,
> > 
> > Abbie
> > 
> > Cristian wrote:
> > 
> > hi!
> > 
> > about test "2.7  COPY-LITERAL and COPY-OFFSET" in the draft "SigComp
> > Torture Tests".
> > in different memory configurations I get different results.
> > when interpreting the sigcomp payload spec'ed in the annex 
> > "A.1.7  COPY-LITERAL and COPY-OFFSET" of the same draft, the
> > instruction:
> > 
> > LOAD ($offset, 1)
> > 
> > unless the udvm memory is over 0x4042 bytes, causes a memory fault, 
> > trying to access memory outside the total memory available for the udvm.
> > am I wrong or is it suppossed to be so?
> > was the "$" intended (i.e writing through indirection)? the next
> > instruction:
> > 
> > COPY-OFFSET ($offset, 5, $destination)
> > 
> > does not seem to make use of what the previous LOAD has written.
> > that of course if my udvm has behaved itself up to the 
> > LOAD ($offset, 1).
> > using:
> > 
> > LOAD (offset, 1)
> > 
> > doesn't lead to this kind of problem; the output would be differrent,
> > though.
> > and since I'm at it: how can the interpreter "guess" the minimum udvm
> > memory size needed to run such a bytecode?
> > 
> > bye now!
> > cristian
> 
> -- 
>  _          | There are many people who use UNIX or Linux  
> (_'_        | who IMHO do not understand UNIX.
>   (_'rist   |              --David Korn  
> GPG public key at wwwkeys.de.pgp.net



> _______________________________________________
> Rohc mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/rohc


-- 
 _          | There are many people who use UNIX or Linux  
(_'_        | who IMHO do not understand UNIX.
  (_'rist   |              --David Korn  
GPG public key at wwwkeys.de.pgp.net

_______________________________________________
Rohc mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/rohc