coryxie / nix-os Goto Github PK

we don't use GiB PTEs in the KSEG2 map

This code was lost in the shuffle. There's no reason not to, since we require 
CPUs that support them and a 512-fold reduction in PTEs is a good thing. Also, 
at least until we have more than 512GiB of memory, we can fit all KSEG2 PTEs on 
one page. As it is each GiB of memory takes one page and 512 PTEs!

The fix is in physinit. I'm raising this issue because we're going to get 
pushback but I'm not fond of supporting obsolete CPUs. We've had 10 years of 
that already!

[deleted issue]

On my AMD Athlon II X2 250 (dual core),
only one core is recognized. The other
gets stuck in squidboy() as archhz()
returns 0.

What steps will reproduce the problem?
1. hg clone
2. cd nix-os
3. 9vx -r .

Need a post-clone script that will get run. It should
mkdir usr/username
mkdir /mail/box/username
get a font that works -- euro.9.font is not there. 

cd /sys/src
mk all
cd cmd
mk all
pcc -c -D_POSIX_SOURCE basename.c
mk -f /sys/src/cmd/mkfile cc.6
mk: don't know how to make 'cc.6' in directory /sys/src/ape/cmd
mk: mk -f /sys/src/cmd/mkfile cc.6  : exit status=rc 5766: mk 5768: error
mk: cd cmd mk all  : exit status=rc 5757: mk 5760: error
mk: mk lib.all mk ...  : exit status=rc 3173: mk 5756: error
mk: date for (i ...  : exit status=rc 145: rc 3170: mk 3172: error

What steps will reproduce the problem?
1. run /sys/lib/newuser for a new user
2. type acme
3. note it fails with usage

Just fix it someone, I don't understand this damned shell.

What steps will reproduce the problem?
1. do a clean pull of nix-os
2. mkdir /usr/your-name
3. /sys/lib/newuser
4. run acme

What is the expected output? What do you see instead?
acme should run. 

 acme /usr/rminnich/lib/profile
imageinit: can't open font /lib/font/bit/lucidasans/euro.8.font: 
'/lib/font/bit/lucidasans/euro.8.font' does not exist

Can somebody smart please fix this? We keep getting it wrong. I know I'll screw 
it up.

ron

What steps will reproduce the problem?
1. plumb url
2. ...
3. crash

What is the expected output? What do you see instead?


What version of the product are you using? On what operating system?


Please provide any additional information below.

here's the fix
; diff -c fsys.c /mnt/term/usr/quanstro/nix-os/sys/src/cmd/plumb/
fsys.c:10,16 - /mnt/term/usr/quanstro/nix-os/sys/src/cmd/plumb//fsys.c:10,16

  enum
  {
-   Stack = 16*1024
+   Stack = 8*1024
  };

  typedef struct Dirtab Dirtab;

Man pages should be updated, in general.
There are new ones for (some of) the new nix
features, but other ones ignore that this system
differs from Plan 9 in some respects. Most notably,
introduction pages should perhaps change.

It would be good to review all of them to check out that
they are correct for nix. But I know that's a herculean task.

run /sys/lib/newuser and watch it fail in many ways

What steps will reproduce the problem?
1. cd /sys/src/cmd
2. mk all


term% mk all
mk: don't know how to make '6.aan' in directory /sys/src/cmd
term% 

mk install fails
mk 5l
6c -FTVw asm.c
<eof> unknown op in type complex: OXXX
mk: 6c -FTVw asm.c  : exit status=rc 15851: 6c 15853: error
mk: for(i in 1a ...  : exit status=rc 15515: rc 15848: mk 15850: error
mk: test -e 8._cp ...  : exit status=rc 14315: mk 15499: error
mk: date for (i ...  : exit status=rc 7717: rc 14297: mk 14299: error

Don't know whether you are aware of http://nixos.org/nixos/ 
Name clash might be confusing...