Messages in this thread | | | Date | Tue, 4 May 1999 11:54:27 -0400 (EDT) | From | Alexander Viro <> | Subject | Re: pre-patch-2.2.8-series candidate, TESTERS NEEDED... |
| |
On Tue, 4 May 1999, Andre M. Hedrick wrote:
> Andries B. discovered that there is a jumper that allows one to select > between 15/16 hardwired head reporting. They are shipped with the jumper > selected to 15 and you loose about 5% capacity.
Worse yet, some laptops are shipped with 15-heads setting and backup/re{format|partition}/restore is not too pleasant option.
> Since the geometry rules for disk translations larger than 8.4GB do not > allow for 15 heads, he and I are discussing various options to account for > the error that IBM did.
IIRC one the laptop where I've seen that lossage BIOS geometry was munged to ???*240*63.
> The request is reasonable to add 15/16 combination, but the rule breaking > leaves us wondering some. Since this is a geometry issue and Andries is > primarily responsible for that part of the kernel, I will defer to his > judgement.
> If you have a list of drive models and names they could be useful for > determining the action needed to address the problem.
Hmm... Laptop in question wasn't mine, so I'll just Cc the owner. Oleg, could you give the model name of that sucker/give additional details?
Linux managed to boot with the unpatched ide-disk.c and mounted everything OK, but fdisk was very unhappy until I did s/15/16/ in the function in question. After that geometry was recoginzed and fdisk returned into sane state.
So there... I don't have additional datapoints. I think that reasonable course of actions would be to recognize 240 heads as valid munge and check for this 0x3fff*15*63 bogosity, maybe informing happy owner about the situation. But then I'm VFS/filesystems hacker, and underlying stuff is *not* an area I'm comfortable with ;-) Cheers, Al
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |