Messages in this thread |  | | Subject | General thoughts for 2.1+ | Date | Wed, 23 Oct 1996 15:13:38 -0400 (EDT) | From | "Aaron Tiensivu" <> |
| |
I'm a little worried this might start something ugly.. but bear with me. :)
The 'Documentation' directory isn't being used by some areas of the kernel source. I realize some are probably there for a good reason, but I figured we should probably full endorse the idea of a 'Documentation' area.
The offenders: ./fs/nfs/README ./net/README ./net/decnet/README ./net/ipv6/README ./drivers/net/README1.PLIP ./drivers/net/README2.PLIP ./drivers/net/README.DLINK ./drivers/net/README.wavelan ./drivers/net/README.ewrk3 ./drivers/net/README.de4x5 ./drivers/net/README.eql ./drivers/net/README.tunnel ./drivers/net/README.multicast ./drivers/net/README.pt ./drivers/net/README.smc9 ./drivers/net/README.dgrs ./drivers/block/README.ide ./drivers/block/README.hd ./drivers/block/README.fd ./drivers/block/README.md ./drivers/char/README.scc ./drivers/char/README.stallion ./drivers/char/ftape/README.PCI ./drivers/char/README.baycom ./drivers/char/README.cyclomY ./drivers/scsi/README.st ./drivers/scsi/README.aic7xxx ./drivers/scsi/README.BusLogic ./drivers/scsi/README.g_NCR5380 ./drivers/scsi/README.AM53C974 ./drivers/scsi/README.ppa ./drivers/scsi/README.FlashPoint ./drivers/scsi/README.dtc3x80 ./drivers/scsi/README.qlogicisp ./drivers/scsi/README.qlogicfas ./drivers/scsi/README.ncr53c8xx ./drivers/scsi/README.in2000 ./drivers/sound/lowlevel/README ./arch/i386/math-emu/README ./arch/sparc/boot/README ./arch/m68k/ifpsp060/README ./arch/m68k/fpsp040/README ./scripts/README.Menuconfig
Other areas of concern:
1. The GGI people need help with getting that project going full-tilt. I'll be helping out somehow when time permits. Video card programming is not my forte, but I think it is for a very good and valid cause. Their main WWW site is at: http://synergy.caltech.edu/~ggi/
Hopefully, maybe I can help out with the ATI Mach32 support and VGAWonder XL support. [The two video cards I have]
A small sub-set of GGI might be nice for kernel. Not necessarily a gung-ho graphics library, but 'simple' things like being aware of certain extended registers of various chipsets. That work fix a majority of the video corruption problems we see today.
2. This is a really minor 'grip' but Paul Gortmaker already posted the patch and I think it should probably be stuck in 2.1. It changes the '/proc/interrupts ' display to look like:
----------------------------------------- olga:~> cat /proc/interrupts 0: 19970147 timer 1: 180474 keyboard 2: 0 cascade 3: 138931 + serial 4: 600716 (currently free) 5: 861 (currently free) 6: 111 (currently free) 7: 52 (currently free) 8: 10495 + rtc 9: 2933492 WD8013 10: 189 PAS16 11: 865252 + BusLogic BT-747A 12: 320325 + serial 13: 1 math error 14: 114459 + ide0 15: 197 (currently free) olga:~> --------------------------------------------
This patch is even more valid with the increased use of modules, in general.
3. Documentation of the magic cookies found under '/proc/sys' :)
4. Allow nfsiod processes to die when last mount is unmounted
5. Clean-up the 'printk' messages in the kernel - giving them at least a logging level. This might be a real pain.
6. If we're going to fix something that's 'broken' in nature and will break binaries, let's try to cover as many as possible, early in the 2.1 cycle. This way, if we fix things that should be fixed, but might break binary compatibility, it'll just be one large upgrade instead of doing something half-assed. I don't think I've explained this well, but my main motivation is for, in the future, fairly painless updates of kernels for 'non-experimental' people. The main goal is so that people won't go nuts and complain about procps breaking again, or something like that.
7. Friendlier kernel panics. I know this goes against the grain of standard kernels, but for instance, if you muck up the root drive setting in your kernel image, instead of just saying "Kernel panic: cannot find major 03 01" (or whatever), it might give you an option of trying /dev/hda2 instead, or something.
|  |