Messages in this thread | | | Date | Tue, 28 Mar 2000 14:43:45 -0800 (PST) | From | Blu3Viper <> | Subject | Re: Location of shmfs; devfs automagics |
| |
On Tue, 28 Mar 2000, Jeff Garzik wrote: > not an option. cannot force all users of 'hackkernel' package to change > their tried-and-true LILO settings. Vendor kernels are different from > user-built kernels. There are considerations of scale which don't come > into play when you're talking about a single person building their own > custom kernel.
Sure it's an option. Look at the boot commands of various distributions. Look at how wildly they range. Look at everyone who distributes FB setups.
If a user adds a new feature to their kernel, then they need to accomplish the necessary setup to utilize it. If Alan installs a scanner, I the vendor have the reasonable expectation that he should install or enable the appropriate driver support as well. Distributions come with predefined setups. Redhat starts a variety of services and has their scripts layed out in a particular fashion. They have a reasonable expectation that the end user use their scripts in the fashion they designed them. Their scripts are completely different from those found on SUSE or slack.
If I the end user choose a distribution, I choose to use the options and format they made. If I choose devfs, then I choose to use it in one of the available fashions it was built for.
We'll never have a perfect system that pleases everyone.
In rolling out devfs on systems across the corp, I find it is significantly easier to follow the basic setup. Compile, put on devfsd and default conf, add it to boot script, reboot.
It's very simple and works much better than expecting the end user to grok an arcane mknod for his new digital toy or figure out how to run MAKEDEV to create the particular device. Frankly, MAKEDEV is quite lacking.
Not only are new things added automatically (save the odd new driver), but many of the tty problems are gone. No worrying about modes and helper apps that run suid.
-d
- 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/
| |