Submit Hint Search The Forums LinksStatsPollsHeadlinesRSS
14,000 hints and counting!


Click here to return to the 'clarification.' hint
The following comments are owned by whoever posted them. This site is not responsible for what they say.
clarification.
Authored by: bhines on Jun 25, '02 06:32:29AM
This was discussed in march on darwin-development, search the archives for pdisk. In summary I don't think we have to worry about it on OS X.
Date: Sun, 24 Mar 2002 11:00:15 +0000 Subject: Re: pdisk man page comment clarification... From: Ken FitzGerald-Smith To: darwin-development@lists.apple.com > From the pdisk man page: > > "Creating more than fifteen partitions is not advised. There is > currently a > bug in the some (all?) of the kernels which causes access to the > whole disk > fail if more than fifteen partitions are in the map." > > Is this still true? What is the problem? > > -Kevin Elliott > Software Engineer There is 1 good reason why 15 partitions could have been a limitation and that is because if you were creating a ISO/HFS hybrid CD your partition map would necessarily be limited to 63 sectors, ie sector 1 to 63. Since the map needed to be readable at boot time, it had to be laid out in 2K blocks thus yielding 15 as a max. I do not include the ddm in sector 0 as part of the map in the above maths. This condition is still true today if you want to preserve the ISO blocks that must start at block 16/Sector 64 on a CD. > I just create a disk image with 61 data partitions (the max for a > 63-sector pmap partition created by pdisk, with 1 for the pmap > itself and > one for the Apple_Free at the end). There is no requirement for an apple free at the end of the partition map. > All device nodes were created correctly. Interestingly, I but a UFS > partition on each one, but autodiskmount totally croaked and > claimed there > were no filesystems. Hmmmm... I will see if I can reproduce... > Ken From: "David A. Gatwood" To: Ken FitzGerald-Smith cc: darwin-development@lists.apple.com Subject: Re: pdisk man page comment clarification... On Sun, 24 Mar 2002, Ken FitzGerald-Smith wrote: > There is 1 good reason why 15 partitions could have been a > limitation and that is because if you were creating a ISO/HFS > hybrid CD your partition map would necessarily be limited to 63 > sectors, ie sector 1 to 63. Since the map needed to be readable at > boot time, it had to be laid out in 2K blocks thus yielding 15 as a > max. I do not include the ddm in sector 0 as part of the map in the > above maths. The cause of this issue has to do with the major/minor number allocation in Linux. Each device is assigned exactly one major number, and there are only 16 minor numbers per major number. One minor number is reserved for the "whole disk partition", leaving room for fifteen partitions. If the mapping mechanism is broken, you'd get partition 16 instead of the "whole disk" partition, or worse.


[ Reply to This | # ]
clarification.
Authored by: gvitale on Jun 25, '02 06:55:50AM

I'm sorry, but after reading this I do not see why the 15 partitions limit is not an issue in Darwin



[ Reply to This | # ]
Is is an issue?
Authored by: porkchop_d_clown on Jun 25, '02 08:15:28AM

Well, the important point is that it "isn't a bug it's a feature!" in other words, the partitioning code is broken as designed, not broken because of an error in the kernel programming.

As the previous poster noted there are valid technical reasons that limit the size of the partition map and the absolute number of partitions on a drive.

I guess for me the question is why does Apple reserve so many hidden partitions?



[ Reply to This | # ]
clarification.
Authored by: bhines on Jun 26, '02 05:54:54AM

"This condition is still true today if you want to preserve the ISO
blocks that must start at block 16/Sector 64 on a CD."

Is your main hard drive CD?



[ Reply to This | # ]