Re: dac960 devices on boot-floppies
- To: Adam Di Carlo <firstname.lastname@example.org>
- Cc: email@example.com
- Subject: Re: dac960 devices on boot-floppies
- From: Mark van Walraven <firstname.lastname@example.org>
- Date: Thu, 2 Dec 1999 21:19:00 +1300
- Message-id: <19991202211900.A14650@mail.wave.co.nz>
- In-reply-to: <19991128133100.F25402@mail.wave.co.nz>; from Mark van Walraven on Sun, Nov 28, 1999 at 01:31:00PM +1300
- References: <email@example.com> <19991116081619.A20876@mail.wave.co.nz> <firstname.lastname@example.org> <19991118093015.J30121@mail.wave.co.nz> <email@example.com> <19991118175311.B3070@mail.wave.co.nz> <firstname.lastname@example.org> <19991125150734.A22950@mail.wave.co.nz> <email@example.com> <19991128133100.F25402@mail.wave.co.nz>
Well, I now have something much better for fdisk_reread(), but a couple
of issues have come up.
1. Non-disk devices: IDE CD-ROMS can be detected by ioctl(CDROMVOLREAD)
and SCSI CD-ROMs and tape drives don't have names beginning 'sd',
but what about SCSI tape, CD-ROM and other non-disk devices attached
to smart2 and dac960 controllers? How can I check that a target is a
random-access read/write device?
2. Are we making things harder for ourselves by looking for devices
by name? What if fdisk_reread() started with a list of acceptable
device numbers and walked /dev to find the names?
Any suggestions are most welcome.