On Tue, 2006-01-03 at 14:45 -0500, Joey Hess wrote: > J.F. Gratton wrote: > > - I know my hardware, it's unlikely to change in a near-future; a new > > kernel is more likely to come out thant my hardware to change; why using > > an initrd then if I know exactly what needs to be put in modules and > > must not ? > > This assumes a best case scenario that you will never need to get some > new peice of hardware working at a time when taking the time out to set > up a new kernel will be painful. No matter what percentage of time this > best case scenario is true, it will never be true 100% of the time, and > as time goes on the chances that it will fail to be true at some point > approaches one. Some of the failure scenarios are very painful. After it > has failed to be true a couple of times, people tend to switch over to > modular kernels. >
I agree, although not using initrd does not preclude having a modular kernel. In my eyes, where initrd is strong is when it comes to not having to worry about which IDE/SCSI driver using for the initial boot. There, having all drivers compiled as modules and crammed into your initrd image is a winner. I do not think though that you change IDE controllers that many times in a kernel lifetime, so that's why I'm saying that the hardware is not that "dynamic" onto a machine. Other hardware can be used with modules. IDE/SCSI drivers ? It's a bit more dicey. Well at least for the root device, it is. -- Jeff -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
