[Stgt-devel] Fw: Implementation of vtl

aftab azmi max_aftab
Wed May 7 17:27:25 CEST 2008



----- Forwarded Message ----
From: aftab azmi <max_aftab at yahoo.com>
To: Mark Harvey <markh794 at gmail.com>
Sent: Wednesday, 7 May, 2008 8:56:36 PM
Subject: Re: [Stgt-devel] Implementation of vtl

Actually i couldnt find any value for file mark so i choose ascii code 28(file separator) as filemark.i guess it will not be present in any user file so will not produce false positive...

I have assummed fixed block write..i forgot to include check for the FIXED bit in read and write command.. 

For variable block write i think we have to use your structure...for underrun we can compare the blk_size in blk_header against the MINIMUM BLOCK LENGTH LIMIT field in READ BLOCK LIMITS data and for overrun we can compare blk_size and MAXIMUM BLOCK LENGTH LIMIT.

----- Original Message ----
From: Mark Harvey <markh794 at gmail.com>
To: aftab azmi <max_aftab at yahoo.com>
Cc: stgt-devel at lists.berlios.de; fujita.tomonori at lab.ntt.co.jp
Sent: Wednesday, 7 May, 2008 1:52:45 PM
Subject: Re: [Stgt-devel] Implementation of vtl

Hello aftab azmi,

First off, thanks... You beat me to the punch :) It might save me some work....

While I can't speak authoritatively, I'm sure this will be rejected
first off due the the code base not following the Linux kernel coding
style... You can find the coding style document in the Linux kernel
source under the Documentation dir.

I've passed it thru the scripts/checkpatch.pl script and attached are
the coding style errors: (Lots of tab/space issues)

2nd. The 'detection' of a filemark appears a little week. i.e. first
byte at a 512 byte boundary to contain '28'. This could produce many
false positives.

3rd. I see no way to detect variable block underrun/overrun detection
(the main sticking point with my code base)..

4th. You assume variable block writes. This code would fail if the
user set the tape block size to 'fixed block'.

Point 1 is easily resolved with a little effort.
Do you have any ideas on the solution to points 2 & 3 (just not coded yet) ?
Point 4 is an easy code fix.

Keep up the good work.

p.s. To make it much easier for people on this list to review the
code, please refer to the Linux kernel 'submitting patches' document.
You will receive much faster code reviews...

Regards
Mark Harvey

On Tue, May 6, 2008 at 5:24 PM, aftab azmi <max_aftab at yahoo.com> wrote:
> Hiii everyone...
>     Attached herewith  is our implementation  of  Virtual  Tape Library....I have implemented all the mandatory commands mentioned in ssc-r42  except  erase command.....
>  For this to run u have  to change the default-bst in iscsid.c  to "ssc"..
>  We can use tar command to read/write  backup..
>  Please do send in your comments and suggestions for further devlopment.
>
>
>
>
>
>
>       Bollywood, fun, friendship, sports and more. You name it, we have it on http://in.promos.yahoo.com/groups/bestofyahoo/
> _______________________________________________
>  Stgt-devel mailing list
>  Stgt-devel at lists.berlios.de
>  https://lists.berlios.de/mailman/listinfo/stgt-devel
>
>



      Check out the all-new face of Yahoo! India. Go to http://in.yahoo.com/



      Bring your gang together. Do your thing. Find your favourite Yahoo! group at http://in.promos.yahoo.com/groups/




More information about the stgt mailing list