]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
8 years agoether82563: initial i217 support from http://www.9legacy.org/9legacy/patch/pc-ether82...
cinap_lenrek [Thu, 7 Apr 2016 19:24:13 +0000 (21:24 +0200)]
ether82563: initial i217 support from http://www.9legacy.org/9legacy/patch/pc-ether82563-i210.diff (thanks k0ga)

This patch is only an adaptation for 9front of the patch located in
http://www.9legacy.org/9legacy/patch/pc-ether82563-i210.diff.  The
major difference is that this patch ignores errors in checksum of
eeprom, because in my system the checksum was wrong.  After 3 months,
I didn't have problems, and I think the patch can be used.  although
it has some things that need to be fixed.  If the link is inactive
when the system boots then it will remain inactive forever.

8 years agotcp567: run authserver with p9sk1 tickets disabled preventing offline password brute...
cinap_lenrek [Thu, 7 Apr 2016 19:19:48 +0000 (21:19 +0200)]
tcp567: run authserver with p9sk1 tickets disabled preventing offline password brute-force

8 years agojpg: fd == 0 is valid, so don't ignore it
BurnZeZ [Thu, 7 Apr 2016 00:22:42 +0000 (20:22 -0400)]
jpg: fd == 0 is valid, so don't ignore it

8 years agolibdraw: fix out of bounds memory access after subfont array reallocation (thanks...
cinap_lenrek [Tue, 5 Apr 2016 09:24:07 +0000 (11:24 +0200)]
libdraw: fix out of bounds memory access after subfont array reallocation (thanks ray)

/n/bugs/open/libdrawfont.c_buffer_overflow
http://bugs.9front.org/open/libdrawfont.c_buffer_overflow/readme

ray@raylai.com

Hi all,

In plan9port this bug keeps crashing mc when I run lc in a directory with Chinese characters. This is a diff from OpenBSD but it should apply cleanly to the various plan9 sources.

The code is basically trying to do a realloc (I guess realloc wasn't available back then?) but it copies too much from the original buffer.

Since realloc is available, just use it. If realloc isn't available outside plan9port (I haven't checked) the memmove line should be changed from:
memmove(f->subf, of, (f->nsubf+DSUBF)*sizeof *subf);
to:
memmove(f->subf, of, f->nsubf*sizeof *subf);

I hope this is helpful.

Ray

8 years agorio: /dev/kbd cleanup
cinap_lenrek [Mon, 4 Apr 2016 04:47:26 +0000 (06:47 +0200)]
rio: /dev/kbd cleanup

- use free running unsigned indeces for window kbdq
- emalloc() vs malloc() in xfid

8 years agokbdfs(8): document new /dev/kbd behaviour
cinap_lenrek [Sun, 3 Apr 2016 21:16:08 +0000 (23:16 +0200)]
kbdfs(8): document new /dev/kbd behaviour

8 years agochange /dev/kbd to return multiple messages per read
cinap_lenrek [Sun, 3 Apr 2016 20:54:22 +0000 (22:54 +0200)]
change /dev/kbd to return multiple messages per read

8 years agomercurial: CVE-2016-3630
cinap_lenrek [Sun, 3 Apr 2016 02:04:56 +0000 (04:04 +0200)]
mercurial: CVE-2016-3630

backported the following patches from mercurial 3.7.3:

https://selenic.com/repo/hg/rev/b6ed2505d6cf
https://selenic.com/repo/hg/rev/b9714d958e89

8 years agokernel: always clunk closed fids asynchronously, regardless of caching
cinap_lenrek [Fri, 1 Apr 2016 12:12:50 +0000 (14:12 +0200)]
kernel: always clunk closed fids asynchronously, regardless of caching

8 years ago6in4: ingress filter multicast and link-local, but allow relay traffic
cinap_lenrek [Thu, 31 Mar 2016 21:53:10 +0000 (23:53 +0200)]
6in4: ingress filter multicast and link-local, but allow relay traffic

8 years ago6in4: allow setting the local IPv4 address with -i flag (thanks k0ga)
cinap_lenrek [Thu, 31 Mar 2016 18:35:02 +0000 (20:35 +0200)]
6in4: allow setting the local IPv4 address with -i flag (thanks k0ga)

8 years ago9fs: bind -b /n/9front.org!7734 /n/lists
stanley lieber [Thu, 31 Mar 2016 18:15:45 +0000 (14:15 -0400)]
9fs: bind -b /n/9front.org!7734 /n/lists

8 years agokernel: remove unused NSMAX, NSLOG, NSCACHE constants from portdat.h
cinap_lenrek [Thu, 31 Mar 2016 02:23:27 +0000 (04:23 +0200)]
kernel: remove unused NSMAX, NSLOG, NSCACHE constants from portdat.h

8 years agofortunes: Ah, nice. the woodpeckers are back.
stanley lieber [Thu, 31 Mar 2016 01:14:39 +0000 (21:14 -0400)]
fortunes: Ah, nice. the woodpeckers are back.

8 years agodevsegment: cleanups
cinap_lenrek [Wed, 30 Mar 2016 20:49:13 +0000 (22:49 +0200)]
devsegment: cleanups

- return distinct error message when attempting to create Globalseg with physseg name
- copy directory name to up->genbuf so it stays valid after we unlock(&glogalseglock)
- cleanup wstat() handling, allow changing uid
- make sure global segment size is below SEGMAXSIZE
- move isoverlap() check from globalsegattach() into segattach()
- remove Proc* argument from globalsegattach(), segattach() and isoverlap()
- make Physseg.attr and segattach attr parameter an int for consistency

8 years agobullshit: +progressive +enhancement
cinap_lenrek [Wed, 30 Mar 2016 00:36:33 +0000 (02:36 +0200)]
bullshit: +progressive +enhancement

8 years agozynq: cleanup devarch, flushmmu() after procflushpsecg()
cinap_lenrek [Tue, 29 Mar 2016 03:06:32 +0000 (05:06 +0200)]
zynq: cleanup devarch, flushmmu() after procflushpsecg()

8 years agokernel: fix procflushmmu()
cinap_lenrek [Tue, 29 Mar 2016 00:09:49 +0000 (02:09 +0200)]
kernel: fix procflushmmu()

fix bug introduced in previous change for zynq, broke
procflushseg() function only flushing the first proc
matching the segment.

8 years agokernel: print pid as %lud instead %lux (in tsleep() debug print)
cinap_lenrek [Mon, 28 Mar 2016 21:01:54 +0000 (23:01 +0200)]
kernel: print pid as %lud instead %lux (in tsleep() debug print)

8 years agolibsec: fix tlsid for TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES...
cinap_lenrek [Mon, 28 Mar 2016 19:58:37 +0000 (21:58 +0200)]
libsec: fix tlsid for TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256

the previous cipher id's where fore TLS_ECDH_* not TLS_ECDHE_*... sorry :(

8 years agodevtls: print the path of the underlying chan in status file
cinap_lenrek [Mon, 28 Mar 2016 18:12:54 +0000 (20:12 +0200)]
devtls: print the path of the underlying chan in status file

to figure out what network connection a particular tls
conversation refers to, we add the path of the underlying
we send the encrypted tls traffic over in the status file,
example:

term% grep -n '^Chan:' '#a'/tls/*/status
#a/tls/0/status:7: Chan: /net/tcp/6/data
#a/tls/1/status:7: Chan: /net/tcp/0/data

8 years agodevip: applying changes for bug: multicasts_and_udp_buffers
cinap_lenrek [Mon, 28 Mar 2016 14:58:09 +0000 (16:58 +0200)]
devip: applying changes for bug: multicasts_and_udp_buffers

/n/bugs/open/multicasts_and_udp_buffers
http://bugs.9front.org/open/multicasts_and_udp_buffers/readme

michal@Lnet.pl

I have ported my small MPEG-TS analisis tool to Plan9.

To allow this application working I had to fix a bug in the kernel IPv4 code and increase UDP input buffer.

Bug is related to listening for IPv4 multicast traffic. There is no problem if you listen for only one group or multiple groups with different UDP ports. This works:

Write to UDP ctl:

anounce PORT
addmulti INTERFACE_ADDR MULTICAST_ADDR
headers

and you can read packets from data file.

You need to set headers option because otherwise every UDP packet for MULTICAST_ADDR!PORT is treat as separate connection. This is a bug and should be fixed too, but I didn't tried it.

There is a problem when you need to receive packets for multiple multicast groups. Usually the same destination port is used by multiple streams and above sequence of commands fails for second group because the port is the same.

Simple and probably non-intrusive fix is adding "|| ipismulticast(addr)" to if statement at /sys/src/9/ip/devip.c:861 line:

if(ipforme(c->p->f, addr) || ipismulticast(addr))

This fixes the problem and now you can use the following sequence to listen for multiple multicast groups even if they all have the same destination port:

announce MULTICAST_ADDR!PORT
addmulti INTERFACE_ADDR MULTICAST_ADDR
headers

After that my application started working but signals packet drops at >2 Mb/s input rate. The same is reported by kernel netlog. Increase capacity of UDP connection input queue fixes this problem /sys/src/9/ip/udp.c:153

c->rq = qopen(512*1024, Qmsg, 0, 0);

--
MichaƂ Derkacz

8 years agozynq: introduce SG_FAULT to prevent access to AXI segment while PL is not ready
cinap_lenrek [Sun, 27 Mar 2016 18:57:01 +0000 (20:57 +0200)]
zynq: introduce SG_FAULT to prevent access to AXI segment while PL is not ready

access to the axi segment hangs the machine when the fpga
is not programmed yet. to prevent access, we introduce a
new SG_FAULT flag, that when set on the Segment.type or
Physseg.attr, causes the fault handler to immidiately
return with an error (as if the segment would not be mapped).

during programming, we temporarily set the SG_FAULT flag
on the axi physseg, flush all processes tlb's that have
the segment mapped and when programming is done, we clear
the flag again.

8 years agofile: deal with negative coordinates in plan9 bitmaps, print image size
cinap_lenrek [Sun, 27 Mar 2016 14:38:39 +0000 (16:38 +0200)]
file: deal with negative coordinates in plan9 bitmaps, print image size

8 years agokernel: fix tsleep()/twakeup()/tsemacquire() race
cinap_lenrek [Sat, 26 Mar 2016 01:37:42 +0000 (02:37 +0100)]
kernel: fix tsleep()/twakeup()/tsemacquire() race

tsleep() used to cancel the timer with:

if(up->tt != nil)
timerdel(up);

which still can result in twakeup() to fire after tsleep()
returns (because we set Timer.tt to nil *before* we call the tfn).
in most cases, this is not an issue as the Rendez*
usually is just &up->sleep, but when it is dynamically allocated
or on the stack like in tsemacquire(), twakeup() will call
wakeup() on a potentially garbage Rendez structure!

to fix the race, we execute the wakup() with the Timer lock
held, and set p->trend to nil only after we called wakeup().

that way, the timerdel(); which unconditionally locks the Timer;
can act as a proper barrier and use up->trend == nil as the
condition if the timer has already fired.

8 years agorio: flushimage() before wscrsleep()
cinap_lenrek [Wed, 23 Mar 2016 19:17:58 +0000 (20:17 +0100)]
rio: flushimage() before wscrsleep()

8 years agodevtls: zero secret information before freeing, cleanup
cinap_lenrek [Wed, 23 Mar 2016 12:50:58 +0000 (13:50 +0100)]
devtls: zero secret information before freeing, cleanup

8 years agostats: fix display->locking race
cinap_lenrek [Wed, 23 Mar 2016 10:19:36 +0000 (11:19 +0100)]
stats: fix display->locking race

8 years agoadd portable AES-GCM (Galois/Counter Mode) implementation to libsec and devtls
cinap_lenrek [Wed, 23 Mar 2016 01:45:35 +0000 (02:45 +0100)]
add portable AES-GCM (Galois/Counter Mode) implementation to libsec and devtls

8 years agorio: add flushimage() after mb2 menu action
cinap_lenrek [Tue, 22 Mar 2016 16:48:08 +0000 (17:48 +0100)]
rio: add flushimage() after mb2 menu action

8 years agomothra(1): fix url for https://code.9front.org/hg/plan9front/
stanley lieber [Tue, 22 Mar 2016 00:57:18 +0000 (20:57 -0400)]
mothra(1): fix url for https://code.9front.org/hg/plan9front/

8 years agorio botch
BurnZeZ [Sun, 20 Mar 2016 00:27:09 +0000 (20:27 -0400)]
rio botch

8 years agofile(1): print warnings on /fd/2, remove some duplicate strings, whitespace fixes
BurnZeZ [Sun, 20 Mar 2016 00:09:44 +0000 (20:09 -0400)]
file(1): print warnings on /fd/2, remove some duplicate strings, whitespace fixes

8 years agogames/life: add -d delay parameter, -b color reversal parameter, and update manpage
BurnZeZ [Sun, 20 Mar 2016 00:03:10 +0000 (20:03 -0400)]
games/life: add -d delay parameter, -b color reversal parameter, and update manpage

8 years agogames(1) cleanup
BurnZeZ [Sun, 20 Mar 2016 00:00:03 +0000 (20:00 -0400)]
games(1) cleanup

8 years agolib9p: remove duplicate variable assignment in createfile()
BurnZeZ [Sat, 19 Mar 2016 21:58:35 +0000 (17:58 -0400)]
lib9p: remove duplicate variable assignment in createfile()

8 years agorio: fix some spacing, slightly cleaner arg handling
BurnZeZ [Sat, 19 Mar 2016 21:50:33 +0000 (17:50 -0400)]
rio: fix some spacing, slightly cleaner arg handling

8 years agolibdraw: have openfont() set error string
BurnZeZ [Sat, 19 Mar 2016 21:47:25 +0000 (17:47 -0400)]
libdraw: have openfont() set error string

8 years agolibc: trailing whitespace cleanup
BurnZeZ [Sat, 19 Mar 2016 21:35:36 +0000 (17:35 -0400)]
libc: trailing whitespace cleanup

8 years agoip/gping: add main pid to pid list for killall()
cinap_lenrek [Fri, 18 Mar 2016 07:59:01 +0000 (08:59 +0100)]
ip/gping: add main pid to pid list for killall()

8 years agoabandon streaming experiment
cinap_lenrek [Thu, 17 Mar 2016 16:48:19 +0000 (17:48 +0100)]
abandon streaming experiment

for queue like non-seekable files, it is impossible to implement an
exportfs because one has to run the kernels devtab read() and write()
in separate processes, and that makes it impossible to maintain 9p message
order as the scheduler can come in and randomly schedule one process before
another.

so as soon as we have a transition from 9p -> syscalls, we'r screwed.

i currently see just two possibilities:

- introduce special file type like QTSEQ with strictly ordered i/o semantics
- fix all fileservers and exportfs to only do one outstanding i/o to QTSEQ files
which means maintaining a queue per fid

this doesnt propagate. so exporting slow 9p mount again will be limited
again by latency of the inner mount.

other option:

- return offset in Rread, so client can bring responses back into order. this
requires changing all fileservers and drivers to maintain such an per fid offset
and change the protocol to include it in the response, and also pass it to userspace
(new syscalls or pass it in TOS)

this only works for read pipelining, write is still screwed.

both options suck.

--
cinap

8 years agolibc: fix runestrecpy() return value (thanks spew)
cinap_lenrek [Wed, 16 Mar 2016 16:27:00 +0000 (17:27 +0100)]
libc: fix runestrecpy() return value (thanks spew)

8 years agoppp: fix buffer overflow, set correct state after chap negotiation (thanks k0ga)
cinap_lenrek [Tue, 15 Mar 2016 21:31:03 +0000 (22:31 +0100)]
ppp: fix buffer overflow, set correct state after chap negotiation (thanks k0ga)

(ppp->secret comes from factotum and it can have any size)
This patch also sets the correct state after success and
failure cases in chap negotiation (without them the code was
working because it expected the other point to pass to net
phase or due to the timer).

8 years agoppp: Small format and log fixes (thanks k0ga)
cinap_lenrek [Tue, 15 Mar 2016 21:19:19 +0000 (22:19 +0100)]
ppp: Small format and log fixes (thanks k0ga)

8 years agoppp: md5 and mschap inside chap, do not request encryption with -c or -C (thanks...
cinap_lenrek [Tue, 15 Mar 2016 21:18:05 +0000 (22:18 +0100)]
ppp: md5 and mschap inside chap, do not request encryption with -c or -C (thanks k0ga)

p2.patch: Do not request encriptation with -c or -C in ppp
(it was a bit annoying request compression, and when the ACK
from the server was received then send a NAK).

p3.patch: Add support for md5 and mschap in in chap
(without this patch ppp was passing to the net stage
without worring about chap).

8 years agopppoe: Add support for -c and -C to enable/disable header compression (thanks k0ga)
cinap_lenrek [Tue, 15 Mar 2016 21:10:54 +0000 (22:10 +0100)]
pppoe: Add support for -c and -C to enable/disable header compression (thanks k0ga)

8 years agorio: flushimage() improves your ability to see, i repeat twice.
cinap_lenrek [Tue, 15 Mar 2016 21:06:11 +0000 (22:06 +0100)]
rio: flushimage() improves your ability to see, i repeat twice.

8 years agorio: remove more unneccesary flushimage() calls
cinap_lenrek [Tue, 15 Mar 2016 17:32:05 +0000 (18:32 +0100)]
rio: remove more unneccesary flushimage() calls

8 years agotruss: add fake __NSEC syscall name for ape
cinap_lenrek [Mon, 14 Mar 2016 02:15:59 +0000 (03:15 +0100)]
truss: add fake __NSEC syscall name for ape

8 years agotruss: fix "bad string" error due to missing ape _nsec() syscall
cinap_lenrek [Mon, 14 Mar 2016 02:11:43 +0000 (03:11 +0100)]
truss: fix "bad string" error due to missing ape _nsec() syscall

8 years agouhtml: dont trust charset=utf-8 attribute, verify.
cinap_lenrek [Sun, 13 Mar 2016 22:47:24 +0000 (23:47 +0100)]
uhtml: dont trust charset=utf-8 attribute, verify.

when the charset is explicitely specified as utf-8, ignore it
for now. we'll assume utf-8 when all bytes have been properly
utf-8 encoded.

8 years agog: dont pollute the environment (thanks BurnZeZ)
cinap_lenrek [Sun, 13 Mar 2016 21:34:08 +0000 (22:34 +0100)]
g: dont pollute the environment (thanks BurnZeZ)

8 years agokbdfs: map Kup/Kdown the same in shift tab (fixes shift-[up]/[down] one-line scroll...
cinap_lenrek [Sun, 13 Mar 2016 20:38:28 +0000 (21:38 +0100)]
kbdfs: map Kup/Kdown the same in shift tab (fixes shift-[up]/[down] one-line scroll in rio over drawterm/vncs)

8 years agolibdraw: don't flush in readmouse() when theres nothing to flush
cinap_lenrek [Sat, 12 Mar 2016 23:03:42 +0000 (00:03 +0100)]
libdraw: don't flush in readmouse() when theres nothing to flush

8 years agoip/gping: ignore advice, fix packet corruption check, remove dead code
cinap_lenrek [Sat, 12 Mar 2016 22:27:49 +0000 (23:27 +0100)]
ip/gping: ignore advice, fix packet corruption check, remove dead code

8 years agodevip: handle ignoreadvice flag for all protocols
cinap_lenrek [Sat, 12 Mar 2016 22:07:58 +0000 (23:07 +0100)]
devip: handle ignoreadvice flag for all protocols

8 years agovncs: parse screenid as long, not short: BGSHORT() -> BGLONG()
cinap_lenrek [Sat, 12 Mar 2016 21:36:00 +0000 (22:36 +0100)]
vncs: parse screenid as long, not short: BGSHORT() -> BGLONG()

8 years agolibdraw: remove flushimage calls from fontresize() and loadchar()
cinap_lenrek [Sat, 12 Mar 2016 21:01:43 +0000 (22:01 +0100)]
libdraw: remove flushimage calls from fontresize() and loadchar()

8 years agoip/gping: set display->locking before starting mouseproc (race)
cinap_lenrek [Sat, 12 Mar 2016 20:44:58 +0000 (21:44 +0100)]
ip/gping: set display->locking before starting mouseproc (race)

8 years agolibsec: remove weakCipher[] array check as we do not support any of these weak ciphers
cinap_lenrek [Sat, 12 Mar 2016 12:24:46 +0000 (13:24 +0100)]
libsec: remove weakCipher[] array check as we do not support any of these weak ciphers

8 years agomothra: dynamically allocate buffer for refresh url (was 20 bytes stack buffer before...
cinap_lenrek [Fri, 11 Mar 2016 20:46:49 +0000 (21:46 +0100)]
mothra: dynamically allocate buffer for refresh url (was 20 bytes stack buffer before) (thanks BurnZeZ for reporting)

8 years agolibvorbis: update to 1.3.5
ftrvxmtrx [Thu, 10 Mar 2016 21:43:00 +0000 (22:43 +0100)]
libvorbis: update to 1.3.5

8 years agomake kernel UTFmax and Runemax consistent with libc (21-bit runes) (thanks maurice)
cinap_lenrek [Thu, 10 Mar 2016 19:02:36 +0000 (20:02 +0100)]
make kernel UTFmax and Runemax consistent with libc (21-bit runes) (thanks maurice)

8 years agomerge
cinap_lenrek [Thu, 10 Mar 2016 18:03:28 +0000 (19:03 +0100)]
merge

8 years agoaan: handle case when reader closes netfd
cinap_lenrek [Thu, 10 Mar 2016 17:50:29 +0000 (18:50 +0100)]
aan: handle case when reader closes netfd

when netfd is closed and set to -1 by the fromnet()
reader proc, handle it like network failure and try
reconnecting.

8 years agodevcons: nil vs 0
cinap_lenrek [Thu, 10 Mar 2016 02:28:36 +0000 (03:28 +0100)]
devcons: nil vs 0

8 years agokernel: make fversion()/mntversion() types consistent
cinap_lenrek [Thu, 10 Mar 2016 02:02:28 +0000 (03:02 +0100)]
kernel: make fversion()/mntversion() types consistent

8 years agodevtls: fix wrong iounit
cinap_lenrek [Wed, 9 Mar 2016 18:54:33 +0000 (19:54 +0100)]
devtls: fix wrong iounit

devtls writes are only atomic up to MaxRecLen as this is the
maximum payload size we put in a record application message.

8 years agorio, libdraw: experimental removal of redundant flushimage() calls for roundtrip...
cinap_lenrek [Tue, 8 Mar 2016 15:45:29 +0000 (16:45 +0100)]
rio, libdraw: experimental removal of redundant flushimage() calls for roundtrip latency reduction

- remove redundant flushimage() calls before readmouse()
- remove flushimage() calls for allocimage(),freeimage() and originwindow()

this is experimental. it will break allocimage() error handling unless the
caller does explicit flushimage() calls, tho the gains
in usability over high latency connections is huge. in most cases, programs
will just terminate when encountering these errors.

8 years agoaan: use sync messages as keep alives
cinap_lenrek [Mon, 7 Mar 2016 19:27:34 +0000 (20:27 +0100)]
aan: use sync messages as keep alives

both server and client need to be convinced that the connection
is broken for a connection reestablishment to happen as the server
will only start looking for incoming clients when the connection
already broke. so use the 8 second interval sync messages
to check for connection lifeness. if we miss two syncs in time,
we declare the connecton to be broken and will try to reconnect.

8 years agovncs: fix dead Kend key
cinap_lenrek [Mon, 7 Mar 2016 18:20:05 +0000 (19:20 +0100)]
vncs: fix dead Kend key

8 years ago9/mkfile: update ARCH
ftrvxmtrx [Sat, 5 Mar 2016 10:38:54 +0000 (11:38 +0100)]
9/mkfile: update ARCH

8 years agolibFLAC/mkfile: -DFLAC__NO_ASM
ftrvxmtrx [Sat, 5 Mar 2016 10:00:00 +0000 (11:00 +0100)]
libFLAC/mkfile: -DFLAC__NO_ASM

8 years agolibFLAC: update to 1.3.1
ftrvxmtrx [Sat, 5 Mar 2016 09:32:47 +0000 (10:32 +0100)]
libFLAC: update to 1.3.1

8 years agoauth/pemencode: fix usage()
BurnZeZ [Wed, 2 Mar 2016 02:27:37 +0000 (21:27 -0500)]
auth/pemencode: fix usage()

8 years agodict: similar fixes, different files. also remove runescpy
BurnZeZ [Wed, 2 Mar 2016 02:24:35 +0000 (21:24 -0500)]
dict: similar fixes, different files. also remove runescpy

8 years agodict mkindex: code cleanup
BurnZeZ [Wed, 2 Mar 2016 02:21:06 +0000 (21:21 -0500)]
dict mkindex: code cleanup
handle malloc failure
check if open failed _before_ seeking
better arg handling

8 years agoplot: handle create() failure
BurnZeZ [Wed, 2 Mar 2016 00:40:14 +0000 (19:40 -0500)]
plot: handle create() failure

8 years agomap/libmap: remove unused function/definitions
BurnZeZ [Wed, 2 Mar 2016 00:36:24 +0000 (19:36 -0500)]
map/libmap: remove unused function/definitions

8 years agovt: add -r flag to start in raw mode
BurnZeZ [Wed, 2 Mar 2016 00:17:54 +0000 (19:17 -0500)]
vt: add -r flag to start in raw mode

8 years agosort: code cleanup
BurnZeZ [Wed, 2 Mar 2016 00:09:40 +0000 (19:09 -0500)]
sort: code cleanup

8 years agocc: populate the flag list comment in lex.c
BurnZeZ [Wed, 2 Mar 2016 00:00:41 +0000 (19:00 -0500)]
cc: populate the flag list comment in lex.c
This way the various compiler flags are documented
somewhere, even if not in full.

Also fixed a typo. 'r' should have been 'R'

8 years agocc: sort the lex.c comment listing debug flags
BurnZeZ [Tue, 1 Mar 2016 23:12:04 +0000 (18:12 -0500)]
cc: sort the lex.c comment listing debug flags

8 years agopppoe: Avoid double free (thanks k0ga)
cinap_lenrek [Tue, 1 Mar 2016 22:54:37 +0000 (23:54 +0100)]
pppoe: Avoid double free (thanks k0ga)

k0ga reports:

Hello,

While I was setting my pppoe conexion with my ISP
I discovered several problems in ip/pppoe. I used
the command line ip/pppoe -A '' ether0 and I got
this:

...
dropping unwanted pkt: wrong ac name
panic: D2B called on non-block dc10 (double-free?)
note rcved: sys: trap: fault read addr=0x0 pc=0x000066e1
pppoe 1013: suicide: sys: trap: fault read addr=0x0 pc=0x000066e1
cpu% acid 1013
/proc/1013/text:386 plan 9 executable
/sys/lib/acid/port
/sys/lib/acid/386
acid: stk()
abort()+0x0 /sys/src/libc/9sys/abort.c:6
ppanic(p=0xd1b8,fmt=0xc7f9)+0x146 /sys/src/libc/port/malloc.c:166
D2B(p=0xd1b8,v=0xdc10)+0x57 /sys/src/libc/port/pool.c:926
poolfreel(v=0xdc10,p=0xd1b8)+0x20 /sys/src/libc/port/pool.c:1152
poolfree(p=0xd1b8,v=0xdc10)+0x3b /sys/src/libc/port/pool.c:1287
free(v=0xdc18)+0x23 /sys/src/libc/port/malloc.c:250
clearstate()+0x1b /sys/src/cmd/ip/pppoe.c:328
pppoe(ether=0xdfffefc1)+0x123 /sys/src/cmd/ip/pppoe.c:426
main(argv=0xdfffefa0,argc=0x1)+0x89 /sys/src/cmd/ip/pppoe.c:100
_main+0x31 /sys/src/libc/386/main9.s:16
acid:

clearstate() is called in pppoe.c:424, and it frees acname and sets it
to nil.  pktread() is called in pppoe.c:434 with parameter wantoffer,
which frees acname again in line pppoe.c:360 but doesn't set it to
nil, so clearstate() makes a double free in the next iteration.

8 years agofix /lib/keyboard spacing
BurnZeZ [Tue, 1 Mar 2016 22:30:42 +0000 (17:30 -0500)]
fix /lib/keyboard spacing

8 years agoadd a couple runes to /lib/keyboard
BurnZeZ [Tue, 1 Mar 2016 22:29:38 +0000 (17:29 -0500)]
add a couple runes to /lib/keyboard

8 years agoadd bug to yesterday(1), fix /sys/man/3/cons permissions
BurnZeZ [Tue, 1 Mar 2016 20:27:39 +0000 (15:27 -0500)]
add bug to yesterday(1), fix /sys/man/3/cons permissions

8 years agomanpage references/typos
BurnZeZ [Tue, 1 Mar 2016 20:24:07 +0000 (15:24 -0500)]
manpage references/typos

8 years agolibsec: fix verifyDHparams() for version <= TLS1.1
cinap_lenrek [Tue, 1 Mar 2016 10:30:01 +0000 (11:30 +0100)]
libsec: fix verifyDHparams() for version <= TLS1.1

for version <= TLS1.1, there is no sigalg field in the ServerKeyExchange
message and the signature digest algorithm is fixed to md5+sha1 and we
only support RSA signatures (TLS1.1 doesnt know about ECDSA).

8 years agoglenda's default profile: remove /n/other mount, done by /lib/namespace
cinap_lenrek [Sun, 28 Feb 2016 23:25:14 +0000 (00:25 +0100)]
glenda's default profile: remove /n/other mount, done by /lib/namespace

8 years agoauthsrv(2): document _asgetpakkey(), authpak_hash(), authpak_new(), authpak_finish()
cinap_lenrek [Sun, 28 Feb 2016 22:23:05 +0000 (23:23 +0100)]
authsrv(2): document _asgetpakkey(), authpak_hash(), authpak_new(), authpak_finish()

8 years agogames/doom: handle allocimage() failure
cinap_lenrek [Sun, 28 Feb 2016 10:44:53 +0000 (11:44 +0100)]
games/doom: handle allocimage() failure

8 years agolibsec: have 16 32-bit words in DigestState to avoid out of bounds warnings for poly1305
cinap_lenrek [Sun, 28 Feb 2016 08:09:41 +0000 (09:09 +0100)]
libsec: have 16 32-bit words in DigestState to avoid out of bounds warnings for poly1305

8 years agodevdraw: remove unused Edepth[]
cinap_lenrek [Sun, 28 Feb 2016 02:06:42 +0000 (03:06 +0100)]
devdraw: remove unused Edepth[]

8 years agolibdraw: remove unused static log2[] array
cinap_lenrek [Sun, 28 Feb 2016 00:38:29 +0000 (01:38 +0100)]
libdraw: remove unused static log2[] array

8 years agorio: exit rio when /dev/cons or /dev/kbd read loop terminates
cinap_lenrek [Sat, 27 Feb 2016 01:39:45 +0000 (02:39 +0100)]
rio: exit rio when /dev/cons or /dev/kbd read loop terminates

8 years agokbdfs: read outer /dev/kbd file and use it just like /dev/kbdin
cinap_lenrek [Sat, 27 Feb 2016 01:36:54 +0000 (02:36 +0100)]
kbdfs: read outer /dev/kbd file and use it just like /dev/kbdin

this allows running kbdfs under kbdfs :-)

going use this in new drawterm. drawterm provides the initial
/dev/kbd, but only sends rune up/down messages (keeps it simple).
the servers kbdfs reads that and exports itself the full
set of files, similar to what we do in vncs. this also
provides note processing.

8 years agorcpu: post hangup note to remote when connection breaks
cinap_lenrek [Sat, 27 Feb 2016 01:08:00 +0000 (02:08 +0100)]
rcpu: post hangup note to remote when connection breaks

when the cat that forwards notes on the remote exits,
send a hangup note so processes wont keep hanging arround.

8 years agorcpu(1): fix manpage reference
cinap_lenrek [Mon, 22 Feb 2016 22:26:35 +0000 (23:26 +0100)]
rcpu(1): fix manpage reference

8 years agolibauthsrv: can write p = 2^448 - 2^224 - 1 now, as mpc has constant folding.
cinap_lenrek [Mon, 22 Feb 2016 21:56:55 +0000 (22:56 +0100)]
libauthsrv: can write p = 2^448 - 2^224 - 1 now, as mpc has constant folding.