Messages in this thread | | | Date | Wed, 03 Feb 1999 19:52:07 -0500 | From | Edwin Foo <> | Subject | Re: Kernel interface changes (was Re: cdrecord problems on recent Linux versions) |
| |
At 01:02 AM 2/4/99 +0000, Alan Cox wrote: >> Currently, this model of changing kernel interfaces within supposedly >> stable releases DOESN'T WORK. It must change, or Linux will >> inevitably lose. > >No the model of binary kernel modules doesn't work. Thats not my problem. >Binary only compatibility is the cause of many of the problems in windows
Unfortunately, I think it is a problem you have to take up and deal with. Recompiling sources for entire server setups in a live production environment because of a short->int conversion is nothing short of a nightmare for those of us who have to do it. For huge security holes I can see the necessity, but I think the larger problem of things changing just for change's sake is still not answered. The examples which Monty and Derek bring up are good ones which have nothing to do with security holes, AFAIK.
Besides, binaries are still the best way to get up and running as fast as possible. Waiting to bring up a replacement server because it's still compiling itself is a good way to get yourself fired. And as Derek points out, binary-only releases will not just go away anytime soon. Having been involved in one major use of an open-source OS within a major PC company (Compaq), I can say from firsthand experience that it was difficult to deploy systems and keep morale up with kernel interfaces changing on a moment's notice. Trying to tell non-hacker-type people over the phone how to recompile userland is not fun either.
I think either (a) slower and better documented releases of the kernel/C library or (b) a good binary compatibility layer that allows older kernel modules to still work would help alleviate this. Examples could be either NetBSD or MacOS/PPC's emulation of 680x0 code (heck, I was sincerely impressed that I could run old MacOS extensions on the PowerPC). Of course, this is just my own thought (and especially not MIT's nor Compaq's), and I'm not going to start another religious OS war here. But I will admit that I've had much less headaches and had to recompile far less stuff ever since I switched to NetBSD for my servers a year ago. I still use Linux for my personal machine since it is cooler in terms of features and supported devices, but I find myself recompiling/rebooting a lot more often even when just keeping up with 2.0.xx kernels.
-Edwin
------------------------------------------------------------------------ Edwin Foo | MIT Computer Science - BS '98, MEng. '99 efoo@mit.edu | Compaq Computer Corp. - Cambridge Research Lab (617) 225-9715 | MIT Residential Computing Consultant (RCC) Developer | MIT 6.270 - http://web.mit.edu/6.270 "Love must be sincere; Hate what is evil; cling to what is good." http://hesed.mit.edu/~efoo - Romans 12:9 <>< ------------------------------------------------------------------------
- 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/
| |