]> git.lizzy.rs Git - plan9front.git/log
plan9front.git
11 years agomothra: implement link buttons
cinap_lenrek [Sat, 1 Jun 2013 00:42:02 +0000 (02:42 +0200)]
mothra: implement link buttons

a button embedded in a <a> link tag should follow the link
when clicked.

11 years agoether82557: panic() if receive buffer allocation fails on init
cinap_lenrek [Fri, 31 May 2013 23:42:39 +0000 (01:42 +0200)]
ether82557: panic() if receive buffer allocation fails on init

11 years agoether83815: dont raise error() from etherreset
cinap_lenrek [Fri, 31 May 2013 23:40:56 +0000 (01:40 +0200)]
ether83815: dont raise error() from etherreset

can't raise error() without process context. use panic() instead.

11 years ago/lib/1oct1993: Reprinted with corrections, October 2012
stanley lieber [Fri, 31 May 2013 17:49:17 +0000 (13:49 -0400)]
/lib/1oct1993: Reprinted with corrections, October 2012

11 years agopool: do poolcheck when free nodes get corrupted instead of assert
cinap_lenrek [Fri, 31 May 2013 16:55:01 +0000 (18:55 +0200)]
pool: do poolcheck when free nodes get corrupted instead of assert

doing poolcheck should give us better context to figure out
what memory blocks / owners corrupted the blocks after free.

this is for hunting down a memory corruption seen in connection
with usb serial.

11 years agofixed a segfault in auth/as because it didn't check to see if argv[0] and argv[1...
mveety [Thu, 30 May 2013 02:25:57 +0000 (22:25 -0400)]
fixed a segfault in auth/as because it didn't check to see if argv[0] and argv[1] actually had things in them

11 years agohjfs: deal with ORCLOSE in chancreat()
cinap_lenrek [Wed, 29 May 2013 23:58:09 +0000 (01:58 +0200)]
hjfs: deal with ORCLOSE in chancreat()

ORCLOSE was effectively ignored for Tcreate. just setting
CHRCLOSE flag on chan should fix it.

11 years agomerge
cinap_lenrek [Thu, 30 May 2013 21:27:28 +0000 (23:27 +0200)]
merge

11 years agomerge
cinap_lenrek [Thu, 30 May 2013 21:27:07 +0000 (23:27 +0200)]
merge

11 years agokernel: do all fp state fork from procfork() (like pc kernel)
cinap_lenrek [Thu, 30 May 2013 21:26:21 +0000 (23:26 +0200)]
kernel: do all fp state fork from procfork() (like pc kernel)

this simplifies the arm ports and keeps all the stuff in one place
instead of spreading it thru notify(), trap() and syscall() functions
and prevents useless fp state copying for kernel procs.

also make sure to save fp in notify while still splhi().

11 years agopc kernel: simulate FXSAVE/FXRSTOR tag-byte from x87 tag-word, copy fp state across...
cinap_lenrek [Thu, 30 May 2013 21:16:22 +0000 (23:16 +0200)]
pc kernel: simulate FXSAVE/FXRSTOR tag-byte from x87 tag-word, copy fp state across fork

the tag-word is not compatible between x87 and sse, have to convert
properly for emulation to work.

we now copy fp state across fork (again!) to preserve FCW and MXCSR
registers. this might not be neccesary as we could probably just
get the current value for the curren process and avoid the fpsave()
call, but become conservative again.

11 years agolibmach: fix reversed F0-F7 register order
cinap_lenrek [Thu, 30 May 2013 21:01:59 +0000 (23:01 +0200)]
libmach: fix reversed F0-F7 register order

11 years agoircrc: target is more important (especially in winwatch)
ftrvxmtrx [Wed, 29 May 2013 09:37:30 +0000 (11:37 +0200)]
ircrc: target is more important (especially in winwatch)

11 years agoio, seg: we should exits(nil) at the end of main
ppatience0 [Wed, 29 May 2013 00:06:53 +0000 (20:06 -0400)]
io, seg: we should exits(nil) at the end of main

11 years agomothra: clicking on the url scroll bar, snarfing, or saving a hit when no urls have...
ppatience0 [Mon, 27 May 2013 04:26:19 +0000 (00:26 -0400)]
mothra: clicking on the url scroll bar, snarfing, or saving a hit when no urls have been loaded (right after running mothra) would crash. this was due to use of w->url and selection when they are actually nil.

to fix this, make genwww return 0 if w->url is nil and output an error message when attempting to access selection if it's nil. the problem with the w->url fix is that after clicking on the url scroll bar, an empty url history row could be highlighted by clicking on it. fix this by making sure lp->lo is not less than zero.

11 years agotojpg: GREY[1248] images should be grayscale
ppatience0 [Mon, 27 May 2013 23:32:28 +0000 (19:32 -0400)]
tojpg: GREY[1248] images should be grayscale

11 years agomerge
cinap_lenrek [Tue, 28 May 2013 21:42:59 +0000 (23:42 +0200)]
merge

11 years agokernel: sysrfork abortion
cinap_lenrek [Tue, 28 May 2013 21:41:54 +0000 (23:41 +0200)]
kernel: sysrfork abortion

when we fail to fork resources for the child due to resource
exhaustion, make the half forked child process call pexit()
to free the resources that where allocated and error out.

11 years agotojpg: add jpeg encoder
ppatience0 [Mon, 27 May 2013 01:56:56 +0000 (21:56 -0400)]
tojpg: add jpeg encoder

11 years agokernel: make imagereclaim() skip notext pages (prevent it from blowing the mount...
cinap_lenrek [Mon, 27 May 2013 01:26:53 +0000 (03:26 +0200)]
kernel: make imagereclaim() skip notext pages (prevent it from blowing the mount cache)

imagereclaim would happily uncache pages from the mountcache (port/cache.c)
without ever getting a Image* released from it. simple fix, just check for
p->image->notext but make sure todo it under the page lock :)

11 years agomerge
cinap_lenrek [Sun, 26 May 2013 23:20:46 +0000 (01:20 +0200)]
merge

11 years agokernel: dont copy fpsave on fork, simplify freeing waitq in pexit(), remove unused...
cinap_lenrek [Sun, 26 May 2013 23:17:11 +0000 (01:17 +0200)]
kernel: dont copy fpsave on fork, simplify freeing waitq in pexit(), remove unused semlock from Proc sturcure

11 years agodevfs/devsd: fix waserror() and unused variable compiler problem
cinap_lenrek [Sun, 26 May 2013 23:12:21 +0000 (01:12 +0200)]
devfs/devsd: fix waserror() and unused variable compiler problem

the compiler optimizes setting unused variables out, which is
problematic if they are used in waserror() handler which the
compiler isnt aware of. rearrange the code to avoid this problem.

11 years agokernel: closechanq error
cinap_lenrek [Sun, 26 May 2013 23:09:34 +0000 (01:09 +0200)]
kernel: closechanq error

catch potential interrupt error from kproc(). this can happen when
we run out of processes, then newproc() will call rsrcwait()
which does tsleep(). if the process gets a note, this might
raise a interrupt error.

11 years agokernel: image reclaim pauses
cinap_lenrek [Sun, 26 May 2013 23:04:53 +0000 (01:04 +0200)]
kernel: image reclaim pauses

get a bit more verbose about process image exhaustion and make
imagreclaim() try to get at least one image on the freelist.

use rsrcwait() to notify the state, and call freebroken() in
case imagereclaim() couldnt free any images.

11 years agokernel: fix sysexec() error handling compiler problem, sysrendez() busyloop
cinap_lenrek [Sun, 26 May 2013 22:59:43 +0000 (00:59 +0200)]
kernel: fix sysexec() error handling compiler problem, sysrendez() busyloop

the variables elem and file0 and commited are explicitely
set to avoid that they get freed in ther waserror() handlers.

but it turns out the compiler optimizes this out as he
thinks the variables arent used any further. (the compiler
is not aware of the waserror() / longjmp() semantics).

rearrange the code to account for this. instead of using
a local variable to check for point of no return (commited),
we use up->seg[SSEG] to figure it out.

for file0 and elem, we just rearrange the code. elem can be
checked in the error handler if it was already assigned to
up->text, and file0 is just free()'d after the poperror().

remove silly busy loop in sysrendez. it is not needed.
dequeueproc() will make sure that the process has come to
rest.

11 years agokeep fpregs always in sse (FXSAVE) format, adapt libmach and acid files for new format
cinap_lenrek [Sun, 26 May 2013 20:41:40 +0000 (22:41 +0200)]
keep fpregs always in sse (FXSAVE) format, adapt libmach and acid files for new format

we now always use the new FXSAVE format in FPsave structure and fpregs
file, converting back and forth in fpx87save() and fpx87restore().

document that fprestore() is a destructive operation now.

change fp register definition in libmach and adapt fpr() acid funciton.

avoid unneccesary copy of fpstate and fpsave in sysfork(). functions
including syscalls do not preserve the fp registers and copying fpstate
from the current process would mean we had to fpsave(&up->fpsave); first.
simply not doing it, new process starts in FPinit state.

11 years agoadd leading dot for explicit domain in webcookies, bump redirection limit to 12 in...
cinap_lenrek [Sun, 26 May 2013 01:33:39 +0000 (03:33 +0200)]
add leading dot for explicit domain in webcookies, bump redirection limit to 12 in webfs

11 years agographics(2): remove non existing char *mousedir argument from geninitdraw (from sourc...
cinap_lenrek [Sat, 25 May 2013 02:39:53 +0000 (04:39 +0200)]
graphics(2): remove non existing char *mousedir argument from geninitdraw (from sources / geninitdraw-no-mousedir patch)

11 years agolibdraw: fix agefont crash when display == nil (from sources / agefont-nodisplay...
cinap_lenrek [Sat, 25 May 2013 02:36:16 +0000 (04:36 +0200)]
libdraw: fix agefont crash when display == nil (from sources / agefont-nodisplay patch)

If you run this:
{for (i in `{seq 1 70000}) echo '½•α·'} | mc | wc
mc will eventually suicide. The error comes from
/sys/src/libdraw/font.c:/^agefont which assumes it has a
display. Normally, this is set up by geninitdraw (usually
via initdraw), but mc.c avoids that to avoid the screen
redraws (flicker, at a minimum, but worse in long-running
cases).

graphics(2) does not prohibit this use, so I made the least
intrusive change to make it work.

11 years agokernel: fix floating point exceptions (was broken by sse kernel changes)
cinap_lenrek [Sat, 25 May 2013 20:44:19 +0000 (22:44 +0200)]
kernel: fix floating point exceptions (was broken by sse kernel changes)

the fpenv() instruction stores in x87 format, using mathstate()
would interpret fpsave as FPssestate in case it was enabled!
instead, pass the status word and fppc explicitely to mathnote()
in matherror().

get rid of m->fpsavealign buffer, as we can just use FPssesave struct
which has enougth padding so rounding up base pointer will not overflow.

11 years agokernel: cleanup /sys/src/9/pc/mkfile PORT= and OBJ= lists
cinap_lenrek [Tue, 21 May 2013 17:45:40 +0000 (19:45 +0200)]
kernel: cleanup /sys/src/9/pc/mkfile PORT= and OBJ= lists

11 years agokernel: prevent alarm(2) from returning instantly when the tick timer wraps. (from...
cinap_lenrek [Tue, 21 May 2013 17:31:41 +0000 (19:31 +0200)]
kernel: prevent alarm(2) from returning instantly when the tick timer wraps. (from erik quanstros alarmwrap patch)

11 years agowifi: properly size internal wifi input processing queue
cinap_lenrek [Tue, 21 May 2013 17:29:57 +0000 (19:29 +0200)]
wifi: properly size internal wifi input processing queue

we only used wifi->iq to process management frames only, now its also used
to forward encrypted data frames, so size it accordingly.

11 years agomerge
cinap_lenrek [Tue, 21 May 2013 16:58:13 +0000 (18:58 +0200)]
merge

11 years agomerge
jpathy [Wed, 22 May 2013 16:25:08 +0000 (21:55 +0530)]
merge

11 years agosse kernel support (sources)
jpathy [Wed, 22 May 2013 18:17:05 +0000 (23:47 +0530)]
sse kernel support (sources)

11 years agoadd 6(a|l) sse support to 8(a|l)
jpathy [Tue, 21 May 2013 17:45:13 +0000 (23:15 +0530)]
add 6(a|l) sse support to 8(a|l)

11 years agokernel: make allocb() wait instead of panic() when possible
cinap_lenrek [Tue, 21 May 2013 00:29:46 +0000 (02:29 +0200)]
kernel: make allocb() wait instead of panic() when possible

as long as our process doesnt hold any locks or ilocks, we
can try to wait for the memory to become available instead of
panicing.

11 years agouarti8250: print on malloc failure
cinap_lenrek [Mon, 20 May 2013 15:43:30 +0000 (17:43 +0200)]
uarti8250: print on malloc failure

11 years agomerge
cinap_lenrek [Mon, 20 May 2013 15:33:46 +0000 (17:33 +0200)]
merge

11 years agopcuart: malloc error handling, cleanup
cinap_lenrek [Mon, 20 May 2013 15:32:48 +0000 (17:32 +0200)]
pcuart: malloc error handling, cleanup

11 years agokbmap/ua: fix uppercase ь
ftrvxmtrx [Mon, 20 May 2013 11:48:09 +0000 (13:48 +0200)]
kbmap/ua: fix uppercase ь

11 years agodevether: handle malloc error in etherprobe()
cinap_lenrek [Sun, 19 May 2013 23:17:38 +0000 (01:17 +0200)]
devether: handle malloc error in etherprobe()

11 years agowifi: handle malloc errors in wifiattach()
cinap_lenrek [Sun, 19 May 2013 23:03:20 +0000 (01:03 +0200)]
wifi: handle malloc errors in wifiattach()

11 years agouse resrcwait() when waiting for memory to become available
cinap_lenrek [Mon, 20 May 2013 21:55:38 +0000 (23:55 +0200)]
use resrcwait() when waiting for memory to become available

use resrcwait() when waiting for memory to become available. randomize
the sleep time and properly restore old process status in case tsleep()
gets interrupted.

11 years agodevsd: don't raise Enomem error if sdmalloc() fails, instead wait for the memory...
cinap_lenrek [Sun, 19 May 2013 18:59:55 +0000 (20:59 +0200)]
devsd: don't raise Enomem error if sdmalloc() fails, instead wait for the memory to become available

filesystems do not handle i/o errors well (cwfs will abandon the blocks),
and temporary exhaustion of kernel memory (because of too many i/o's in
parallel) causes read and write on the partition to fail.

i think it is better to wait for the memory to become available in
this case. the single allocation is at max SDmaxio bytes, which makes
it likely to become available. if we havnt even enought fo that, then
rebooting the machine would be the best option. (aux/reboot)

11 years agokbmap/ru: fix uppercase ь
ftrvxmtrx [Sun, 19 May 2013 15:50:51 +0000 (17:50 +0200)]
kbmap/ru: fix uppercase ь

11 years agonusb/kb: simplify repeat logic as scancode now contains the info about esc1 extension
cinap_lenrek [Fri, 17 May 2013 23:46:25 +0000 (01:46 +0200)]
nusb/kb: simplify repeat logic as scancode now contains the info about esc1 extension

11 years agonusb/kb: oops, fixup
ftrvxmtrx [Fri, 17 May 2013 23:03:13 +0000 (01:03 +0200)]
nusb/kb: oops, fixup

11 years agonusb/kb: use a flag to indicate extended code; fix the most unix-friendly key on...
ftrvxmtrx [Fri, 17 May 2013 23:01:34 +0000 (01:01 +0200)]
nusb/kb: use a flag to indicate extended code; fix the most unix-friendly key on se/fi kbd layout

11 years agofaces: just say no to dejavu
stanley lieber [Wed, 15 May 2013 19:38:42 +0000 (15:38 -0400)]
faces: just say no to dejavu

11 years agonetaudit: check if ipgw= is an ip address
cinap_lenrek [Tue, 14 May 2013 16:54:42 +0000 (18:54 +0200)]
netaudit: check if ipgw= is an ip address

11 years agomake clear that the "ipgw" attribute has to be an ip address
cinap_lenrek [Tue, 14 May 2013 16:31:24 +0000 (18:31 +0200)]
make clear that the "ipgw" attribute has to be an ip address

11 years agofortunes: Subject: [9fans] [gsoc] Dart9P
stanley lieber [Tue, 14 May 2013 14:18:49 +0000 (10:18 -0400)]
fortunes: Subject: [9fans]  [gsoc] Dart9P

11 years agorio: place text at beginning of line before cursor and host point
cinap_lenrek [Mon, 13 May 2013 02:40:02 +0000 (04:40 +0200)]
rio: place text at beginning of line before cursor and host point

11 years agorio: fix completion and readback and make it possible to complete history
cinap_lenrek [Mon, 13 May 2013 02:19:10 +0000 (04:19 +0200)]
rio: fix completion and readback and make it possible to complete history

11 years agotoppm: no need to initialize err
ppatience0 [Sun, 12 May 2013 22:00:47 +0000 (18:00 -0400)]
toppm: no need to initialize err

11 years agotoppm: checking for err is useless; it was initialized to nil earlier
ppatience0 [Sun, 12 May 2013 21:52:35 +0000 (17:52 -0400)]
toppm: checking for err is useless; it was initialized to nil earlier

11 years agotoppm: multichan converts to RGB24, not RGBV
ppatience0 [Sun, 12 May 2013 21:46:39 +0000 (17:46 -0400)]
toppm: multichan converts to RGB24, not RGBV

11 years agowriteppm: replace WriteGIF by WritePPM
ppatience0 [Sun, 12 May 2013 21:08:52 +0000 (17:08 -0400)]
writeppm: replace WriteGIF by WritePPM

11 years agowriteppm: simplify
cinap_lenrek [Mon, 13 May 2013 21:03:23 +0000 (23:03 +0200)]
writeppm: simplify

11 years agowriteppm: for bitmaps, 1 is black, 0 is white
ppatience0 [Sun, 12 May 2013 20:36:29 +0000 (16:36 -0400)]
writeppm: for bitmaps, 1 is black, 0 is white

11 years agotoppm: fix style
ppatience0 [Sun, 12 May 2013 20:17:09 +0000 (16:17 -0400)]
toppm: fix style

11 years agotoppm: add -r flag for raw ppm
ppatience0 [Sun, 12 May 2013 20:11:43 +0000 (16:11 -0400)]
toppm: add -r flag for raw ppm

11 years agowriteppm: fix tab/spaces
cinap_lenrek [Sun, 12 May 2013 17:12:21 +0000 (19:12 +0200)]
writeppm: fix tab/spaces

11 years agomerge
cinap_lenrek [Sun, 12 May 2013 17:05:51 +0000 (19:05 +0200)]
merge

11 years agodevaoe: avoid panic on malloc() error by using smalloc() instead.
cinap_lenrek [Sun, 12 May 2013 17:04:07 +0000 (19:04 +0200)]
devaoe: avoid panic on malloc() error by using smalloc() instead.

11 years agofile: recognize bitmap and graymap ppm images
ppatience0 [Sun, 12 May 2013 16:53:43 +0000 (12:53 -0400)]
file: recognize bitmap and graymap ppm images

writeppm: do not print extra space after each pixel for graymap formats to save space

11 years agoether8169: do phywakeup magic only for specific mac versions (from openbsd)
cinap_lenrek [Sun, 12 May 2013 16:25:24 +0000 (18:25 +0200)]
ether8169: do phywakeup magic only for specific mac versions (from openbsd)

11 years agoether8169: and phy wakeup and reset phy before enabling auto negotiation
cinap_lenrek [Sun, 12 May 2013 00:58:02 +0000 (02:58 +0200)]
ether8169: and phy wakeup and reset phy before enabling auto negotiation

this brings up the 8198B on BurnZeZ's ASRock Z68 Professional Gen3 board.

11 years agodevmnt: fix mount device leak and allocation error handling in mntversion()
cinap_lenrek [Sat, 11 May 2013 18:54:50 +0000 (20:54 +0200)]
devmnt: fix mount device leak and allocation error handling in mntversion()

the fist problem is that qopen() might return nil and that kstrdup() will
sleep, so we should try to avoid holding the mntalloc lock. so we move
the kstrdup() and qopen() calls before the Mnt allocation, and properly
recover the memory if we fail later.

the second problem was that we error(Eshort) after we already created the Mnt
when returnlen < sizeof(f.version). this check has to happen *before* we
even attempt to allocate the Mnt structures. note that we only copy the
version string once everything is in the clear, so the semantics of the
user buffer not being modified in case of error is not changed.

a little cleanup in muxclose(), getting rid of mntptfree()...

11 years agosdide: add Intel SERIES 6 SATA controller pci ids
cinap_lenrek [Sat, 11 May 2013 16:45:46 +0000 (18:45 +0200)]
sdide: add Intel SERIES 6 SATA controller pci ids

11 years agoacme: apply nemos acmediskread patch (from sources)
cinap_lenrek [Sat, 11 May 2013 16:43:03 +0000 (18:43 +0200)]
acme: apply nemos acmediskread patch (from sources)

pread does not guarantee that it would read all the data asked for.
But acme usage of disk assumes that. This issues as many reads as
needed to make acme work when read returns less data than it wanted.

11 years agoape: change types from unsigned long to void* for rendezvous() and segattach()/segbrk...
cinap_lenrek [Sat, 11 May 2013 16:34:40 +0000 (18:34 +0200)]
ape: change types from unsigned long to void* for rendezvous() and segattach()/segbrk() (erik), use uniqueue rendezvous tags for _buf and listen

11 years agoawk: proctab.c is always regenerated from maketab, so it is unnecessary to keep it
ppatience0 [Fri, 10 May 2013 00:48:23 +0000 (20:48 -0400)]
awk: proctab.c is always regenerated from maketab, so it is unnecessary to keep it

11 years agosrv: be nice, close the previously duped pipe filedescriptor
cinap_lenrek [Thu, 9 May 2013 23:24:11 +0000 (01:24 +0200)]
srv: be nice, close the previously duped pipe filedescriptor

11 years agofortunes: all the world's a social vax cloud
khm [Thu, 9 May 2013 01:07:02 +0000 (21:07 -0400)]
fortunes: all the world's a social vax cloud

11 years agosamterm: revert previous changes.
ppatience0 [Thu, 9 May 2013 00:34:50 +0000 (20:34 -0400)]
samterm: revert previous changes.

11 years agosamterm: move cursor to next tabstop or non-space character before backspacing in...
ppatience0 [Wed, 8 May 2013 21:47:48 +0000 (17:47 -0400)]
samterm: move cursor to next tabstop or non-space character before backspacing in spaces mode. this makes it less annoying to backspace (it's more like tabs).

11 years agoAdd RDRAND Support for /dev/random
jpathy [Wed, 8 May 2013 17:57:58 +0000 (23:27 +0530)]
Add RDRAND Support for /dev/random

11 years agoape: remove unused variables
ppatience0 [Tue, 7 May 2013 00:51:18 +0000 (20:51 -0400)]
ape: remove unused variables

11 years agoape: fix uninitialized variable bug in getfields()
ppatience0 [Tue, 7 May 2013 00:24:19 +0000 (20:24 -0400)]
ape: fix uninitialized variable bug in getfields()

11 years agomerge
ppatience0 [Mon, 6 May 2013 23:37:51 +0000 (19:37 -0400)]
merge

11 years agoaccess: unused variable
ppatience0 [Mon, 6 May 2013 23:35:03 +0000 (19:35 -0400)]
access: unused variable

11 years agopbs: allow 9bootfat start cluster beyond 0xFFFF
cinap_lenrek [Sun, 5 May 2013 23:15:46 +0000 (01:15 +0200)]
pbs: allow 9bootfat start cluster beyond 0xFFFF

pbs failed to load (silently loading garbage) 9BOOTFAT from start
cluster beyond FFFF because we ignored the low word from the
directory entry. now taking the high 16 bit of the directory's
start cluster into account.

11 years agodevip: dont raise error() out of Fsprotocone()
cinap_lenrek [Sun, 5 May 2013 02:28:50 +0000 (04:28 +0200)]
devip: dont raise error() out of Fsprotocone()

Fsprotoclone() is not supposed to raise error, but return nil.
ipopen() seemed to assume otherwise as it setup error label
before calling Fsprotoclone(). fix ipopen(), make Fsprotoclone()
return nil instead of raising error.

11 years agodevip: handle malloc errors, fix queue leaks
cinap_lenrek [Sun, 5 May 2013 01:56:11 +0000 (03:56 +0200)]
devip: handle malloc errors, fix queue leaks

Fsprotocone():

qopen() and qbypass() can fail and return nil, so make sure
the connection was not partially created by checking if read
and write queues have been setup by the protocol create hanler.
on error, free any resources of the partial connection and
error out.

netlogopen(): check malloc() error.

11 years agolibdraw: fix font f->cacheimage nil dereference
cinap_lenrek [Sat, 4 May 2013 18:36:28 +0000 (20:36 +0200)]
libdraw: fix font f->cacheimage nil dereference

the initial fontresize() might fail but its error code is ignored
potentially leaving f->cacheimage == nil. make sure we call
fontresize() in loadchar() when theres no cacheimage and check the
return value to avoid nil pointer dereference.

11 years agolibmach: update mips disassembler (from sources)
cinap_lenrek [Sat, 4 May 2013 16:40:05 +0000 (18:40 +0200)]
libmach: update mips disassembler (from sources)

11 years agosort: handle write error in debug mode
cinap_lenrek [Sat, 4 May 2013 16:33:52 +0000 (18:33 +0200)]
sort: handle write error in debug mode

11 years agobio: use UTFmax for Bungetsize and fix libmach to include <libc.h> before <bio.h...
cinap_lenrek [Sat, 4 May 2013 16:30:49 +0000 (18:30 +0200)]
bio: use UTFmax for Bungetsize and fix libmach to include <libc.h> before <bio.h> (from sources)

11 years agonedmail: recognize image/jpg mimetype
cinap_lenrek [Sat, 4 May 2013 05:03:37 +0000 (07:03 +0200)]
nedmail: recognize image/jpg mimetype

11 years agou9fs: fix compilation warnings
ftrvxmtrx [Sat, 4 May 2013 00:53:05 +0000 (02:53 +0200)]
u9fs: fix compilation warnings

11 years agou9fs: make it compile
ftrvxmtrx [Sat, 4 May 2013 00:43:27 +0000 (02:43 +0200)]
u9fs: make it compile

11 years agoape: get rid of fixed MUXADDR for buffered i/o
cinap_lenrek [Fri, 3 May 2013 17:34:48 +0000 (19:34 +0200)]
ape: get rid of fixed MUXADDR for buffered i/o

instead of trying to resize the segment (which will not work when
the kernel picks the address as it will allocate right before
the base of the topmost segment), we create the mux segment with the
maximum size needed (arround 1.4MB) for OPEN_MAX filedescriptors.

buf slots will be reused and slots get demand paged once used.

11 years agomkpaqfs(8): correct maximum block size (from erik quanstros mkpaqfsman patch)
cinap_lenrek [Fri, 3 May 2013 11:57:00 +0000 (13:57 +0200)]
mkpaqfs(8): correct maximum block size (from erik quanstros mkpaqfsman patch)

11 years agoape: remove unused variables from _buf (from erik qunastros _bufupd patch)
cinap_lenrek [Fri, 3 May 2013 11:46:27 +0000 (13:46 +0200)]
ape: remove unused variables from _buf (from erik qunastros _bufupd patch)

11 years agofixed operand size for ADD SP, n
qeed [Thu, 2 May 2013 14:40:45 +0000 (10:40 -0400)]
fixed operand size for ADD SP, n

11 years agodevsd: initialize unit->sense[0] in sdsetsense() (from erik quanstroms 9atom)
cinap_lenrek [Thu, 2 May 2013 21:12:37 +0000 (23:12 +0200)]
devsd: initialize unit->sense[0] in sdsetsense() (from erik quanstroms 9atom)