bug#22496: why does the parted math look wrong - Q1 ?

[Narcis Garcia]

I see inconsistent data when printing table. It shoud be:
 Number  Start       End         Size       File system  Name  Flags
 1      0cyl        121600cyl** 121601cyl*               zfs
 9      121600cyl** 121601cyl   2cyl*

* Starting cylinder should be the first cylinder and count in the amount
** Strange ending at same cylinder that other partition begins.

The problem may be partitions aren't created in cylinders (but sectors),
and you try to calculate numbers based on a worng unit now.


On 30/01/16 21:48, Chris Johnson wrote:
> Hi,
> 
> I need some education on how parted is doing math.  ;-)
> 
> The drive is 1 TB.
> 
> # parted /dev/rdsk/c7t1d0p0
> GNU Parted 2.3.0
> Using /dev/rdsk/c7t1d0p0
> Welcome to GNU Parted! Type 'help' to view a list of commands.
> (parted) unit cyl                                                        
> unit cyl
> (parted) p                                                               
> p
> Model: Generic Ide (ide)
> Disk /dev/rdsk/c7t1d0p0: 121601cyl
> Sector size (logical/physical): 512B/512B
> BIOS cylinder,head,sector geometry: 121601,255,63.  Each cylinder is 8225kB.
> Partition Table: gpt
> 
> Number  Start      End        Size       File system  Name  Flags
>  1      0cyl       121600cyl  121600cyl               zfs
>  9      121600cyl  121601cyl  1cyl
> 
> *
> Question 1)*
> 
> Based on the parted value  8225KB per cylinder,  I would think that
> sectors per cylinder could be calculated thus
> 
>     8225 KB/cyl   * 1024 Bytes/KB  = 8422400 Bytes/cyl 
> 
> and  8422400 Bytes/cyl   / 512 Bytes/sector  = 16450   sectors per cylinder
> 
> However, I would also think that sectors per cylinder could be
> calculated thus:
>                              
>                                                                               
>                
> 255  * 63  = 16065   sectors per cylinder
> 
> Obviously these numbers do not match.
> 
> Just for the heck of it, in case when parted says 1 KB it means 1,000
> bytes not 1,024 bytes, I did this math.
> 
>     8225 * 1000 = 8225000    / 512 = 16064.4531 sectors per cylinder
> 
> which is certainly closer to the value you get from 255 * 63 but it
> really doesn't make any sense to me to be thinking that we are getting a
> fractional number of sectors in a given cylinder.
> 
>       
> Thanks for your insight.