[lkml]   [1996]   [Jun]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Linux-2.0

On Thu, 6 Jun 1996, Linus Torvalds wrote:

> Just kidding, folks..
> Anyway, I wasn't kidding about the "final pre-release", so you really _should_
> take a good look at pre2.0.13 and holler about any problems you encounter. Note
> that Linux-2.0 will not be the "Linux to end all Linuxes", and if your
> favourite feature hasn't made it into the latest kernels, don't feel too bad.
> We have at least 5 major numbers to go before we can even catch up with some of
> the competition.
> Now, I haven't been very good at writing the release notes, because I stopped
> at some point when vger was so slow that it didn't make much sense. vger has
> been back, but I've been slow to start writing release notes again.. Anyway, as
> pre2.0.13 is the final pre-release, you obviously automatically know that all
> the bugs are gone, and that this is just a release to check that you're all
> awake and still following the list. The real 2.0 will be the same, except we'll
> change the name of some things.
> HOWEVER, in the unlikely case that there are still problems, I'd like to hear
> about them. We may not be able to fix all the problems in the world in the last
> day or two of 2.0 development, but at least we can try. So consider it your
> civic duty to bother enough that you try this out, ok?
> Yes, I know you're all laughing your heads off and feeling a pity for poor
> optimistic Linus, but hey, deciding when to release the d*mn things is my
> prerogative. So there..
> Linus
> ----
> In case you care, pre-2.0.12 -> pre-2.0.13 does:
> - various documentation stuff.
> - random drievr update
> - mcdx and sbpcd updates
> - pty packet mode disable fixes
> - baycom driver bugfix
> - lance driver support for PCnet PCI II (where do people find these silly
> names?)
> - SMC ultra packet corruption fix
> - BusLogic driver update
> - standard UIO_MAXIOV name (and size)
> - SMB conditional disable of long filenames
> - various net fixes - the TCP send_queue used to get corrupted by local
> firewalling rules etc.
> - mmap() checks sizes.

I'm not sure if this is fixed, but in pre2.0.12, I could not compile the
kernel without defining the loopback character device in the make
config/make xconfig process. I hope this gets fixed, because I don't
need the loopback device, and it is a waste of memory. Also, just for
info, I have checked the sizes of the kernels that I use/have used, and I
think you are doing a good job of keeping bloat down. 1.2.13 was 375708
bytes, and pre2.0.6 was 390374 bytes. pre2.0.12 is 400006 bytes. Not
bad, since the extra 10 or so K probably came from that dumb loopback
device I had to put in there... BTW.. the error came in the final link,
and claimed that something having _loopback in it was not defined. Hope
that is a helpful report.

*ALSO* I really wouldn't mind having a real CPU quota system! I brought
this up before, and it would be incredibly usefull for me and lots of
other people. The tuner daemon seems to make the system act like an XT,
but it is much more stable under heavy loads. The only problem is that
it is extremely unconfigurable, and makes everything slower, including
daemons, servers, NFS, etc. CPU quotas that could be determined by total
CPU time, % of CPU usage, etc would be awesome. Could someone start on this?

- --
Jeff Johnson GCS d- s: !a C+++ UA++(+++) P+ L+ E---- W+++ N+++(+++++) K- w(+) O(-)
KE4QWX M- V-(--) PS+ PE Y++ PGP+++(+++++) t- 5 X+++(+++++) R tv+ b++ DI-- D G++ e* !h r y?
Nerdity Test = 66% Hacker Test = 45%
1024/3397E001 1995/06/10 5B 92 8B 34 84 E9 42 26 DC FB F7 C4 1E 0E 80 29

Version: 2.6.3e-586
Charset: noconv
Comment: Processed by mkpgp1.6, a Pine/PGP interface.


 \ /
  Last update: 2005-03-22 13:37    [W:0.176 / U:1.344 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site