code review 5977045: nix: sync with plan 9

nix: sync with plan 9

[0intro, 2012-04-03 07:33:59]

Hello nixiedev@gmail.com (cc: nix-dev@googlegroups.com),

I'd like you to review this change to
https://code.google.com/p/nix-os/

[0intro, 2012-04-03 07:38:58]

I've just recreated the precedent CL (5933055) with
all the Plan 9 changes from 2012-03-28 to 2012-04-02,
and excluding the "sys/games/lib/fortunes" file,
as requested by Erik.

-- 
David du Colombier

[quanstro, 2012-04-03 13:16:48]

comment set 1 (i'm getting a chunk mismatch with the ndb man page.  :-(.

http://codereview.appspot.com/5977045/diff/3/sys/man/2/encode
File sys/man/2/encode (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/2/encode#newcode45
sys/man/2/encode:45: respectively.
the original patch had

$2n+1$, $(8n+4)/5+1$ and $(n+2)/3)*4 + 1$ bytes,

instead.  the advantage of this is that it displays
correctly in text mode, and looks much better with
man -P, too.  "over" seems out-of-place in an inline
equation.

http://codereview.appspot.com/5977045/diff/3/sys/man/2/encode#newcode88
sys/man/2/encode:88: .B /sys/src/libc/port/u[136][246].c
should list out the files at the original patch did.
this construct is misleading (u12.c?) and doesn't plumb.

.B /sys/src/libc/port/u16.c
.br
.B /sys/src/libc/port/u32.c
.br
.B /sys/src/libc/port/u64.c

[quanstro, 2012-04-03 14:07:29]

second set of changes (sorry; more rietveld errors)

unless i've missed something, i don't think we
should accept the changes that add "9boot".  this
doesn't appear to exist on sources.

http://codereview.appspot.com/5977045/diff/3/sys/man/6/vgadb
File sys/man/6/vgadb (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/6/vgadb#newcode419
sys/man/6/vgadb:419: .IR 9boot (8),
there is no 9boot(8).

http://codereview.appspot.com/5977045/diff/3/sys/man/8/booting
File sys/man/8/booting (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/8/booting#newcode28
sys/man/8/booting:28: or
some sort of copypasta here.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/booting#newcode42
sys/man/8/booting:42: .I 9load
some sort of copypasta here.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/booting#newcode291
sys/man/8/booting:291: .BR /sys/src/9/pcboot .
this doesn't exist.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/ndb
File sys/man/8/ndb (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/8/ndb#newcode756
sys/man/8/ndb:756: Print the names of all PCs that boot via PXE.
pcs aren't the only ones that pxe boot.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/ndb#newcode759
sys/man/8/ndb:759: % ndb/query -a bootf /386/9boot sys
??  /386/9boot is not on sources.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/partfs
File sys/man/8/partfs (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/8/partfs#newcode62
sys/man/8/partfs:62: .IR 9boot (8))
again, 9boot.

http://codereview.appspot.com/5977045/diff/3/sys/man/8/update
File sys/man/8/update (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/8/update#newcode123
sys/man/8/update:123: .IR mkfs (8),
again 9boot

http://codereview.appspot.com/5977045/diff/3/sys/man/8/vga
File sys/man/8/vga (right):

http://codereview.appspot.com/5977045/diff/3/sys/man/8/vga#newcode83
sys/man/8/vga:83: .IR 9boot (8).
again 9boot

http://codereview.appspot.com/5977045/diff/3/sys/man/8/vga#newcode207
sys/man/8/vga:207: .IR 9boot (8)
again 9boot

[quanstro, 2012-04-03 14:24:45]

last chunk.

it would be ideal if someone else looked at the
ohci changes, which really do look wrong to me.

http://codereview.appspot.com/5977045/diff/3/sys/src/9/pc/usbohci.c
File sys/src/9/pc/usbohci.c (right):

http://codereview.appspot.com/5977045/diff/3/sys/src/9/pc/usbohci.c#newcode1264
sys/src/9/pc/usbohci.c:1264: ctlr->ohci->intrdisable = Mie;
i don't know what this could accomplish.  ohci
needs to be on a level-triggered interrupt, and it
would be illegal for the h/w to retrigger the interrupt before the processor
executes eoi.

so disabling the interrupts in the interrupt routine
can't be right.  it might be covering for a bug by
altering the timing.

http://codereview.appspot.com/5977045/diff/3/sys/src/9/pc/usbohci.c#newcode1265
sys/src/9/pc/usbohci.c:1265: coherence();
this coherence doesn't appear to be correct.
ohci->intrdisable is a pci non-posted write,
which is always coherent---the processor waits
for the ack.

http://codereview.appspot.com/5977045/diff/3/sys/src/9/pc/usbohci.c#newcode2499
sys/src/9/pc/usbohci.c:2499: coherence();
bogus coherence after non-posted writes.

http://codereview.appspot.com/5977045/diff/3/sys/src/cmd/upas/fs/imap4.c
File sys/src/cmd/upas/fs/imap4.c (right):

http://codereview.appspot.com/5977045/diff/3/sys/src/cmd/upas/fs/imap4.c#newc...
sys/src/cmd/upas/fs/imap4.c:419: if(0 && (!imap->thumb || !okThumbprint(digest,
imap->thumb))){
i use this.  perhaps it would be better to put the
thumbprints under user control.

[0intro, 2012-04-10 07:29:04]

Hello nixiedev@gmail.com, quanstro@quanstro.net (cc: nix-dev@googlegroups.com),

Please take another look.

[0intro, 2012-04-10 21:37:57]

Hello nixiedev@gmail.com, quanstro@quanstro.net (cc: nix-dev@googlegroups.com),

Please take another look.

[0intro, 2012-04-10 21:41:39]

After some private discussions with Erik, we concluded
it was less confusing to let the 9boot changes in the
manuals out for now.

-- 
David du Colombier

[Nemo, 2012-04-12 10:14:44]

*** Submitted as bd329fe2476d ***

nix: sync with plan 9

R=nixiedev, quanstro
CC=nix-dev
http://codereview.appspot.com/5977045

Committer: Francisco J Ballesteros <nemo.mbox@gmail.com>