On Tue, Aug 05, 2008 at 10:10:47AM -0400, Ian Shields wrote:
> Fair point. However, if you think it's important to know something of the 
> difference between SCSI and ATA disks, then I think it's crucial to 
> understand that libATA gives ATA disks much of the appearance of SCSI. 
> However, there are differences. One difference is that ATA disks still 
> have a limit of 4 primary partitions. Another is that SCSI limits the 
> total number of partitions much more than ATA does. As a topic on its own, 
> the limited set of questions might become gimmes. As an important part of 
> understanding SCSI/ATA differences, there are a few questions that could 
> be a reasonable part of a larger topic.

The PC platform limits partition tables to 4 primary entries.  Doesn't
matter if it is PATA, SATA, SCSI, ESDI or MFM.  Extended doesn't
technically have a limit although I suspect most operating systems put a
limit somewhere.

The linux device entries also have the limit, which is one major per
pair of PATA devices using /dev/hdX with the full device using 0 and 64
as the minor number respectively, and 1-63 and 65-127 for the
partitions, and hence 4 primary and 59 secondary partitions per drive
(which actually works out to only 62 useful partitions since you loose a
primary to run the secondaries).  SCSI style /dev/sdX only allocate 16
minors per drive and hence have up to 4 primary and 11 secondaries per
drive.  In either case the use of LVM and other logical disk management
can easily work around any limitations that may cause.

Systems using GPT instead of the older MBR partition system, can have up
to 128 partitions per drive (as far as I recall) although the linux
major/minor limitations still apply.

LibATA mainly is nice new code that seems cleaner and to have less
duplication, and it elliminates the need for ide-scsi for cd burners
(for those applications that insisted on using SCSI interfaces), and
makes all disks show up consistently, which seems nice.  After all SCSI,
USB and firewire already shared one method.

-- 
Len Sorensen
_______________________________________________
lpi-examdev mailing list
lpi-examdev@lpi.org
http://list.lpi.org/cgi-bin/mailman/listinfo/lpi-examdev

Reply via email to