jca [Sun, 18 May 2014 21:43:11 +0000 (21:43 +0000)]
hid_start_parse can't fail - it calls malloc(M_WAITOK). ok mpi@
miod [Sun, 18 May 2014 21:18:07 +0000 (21:18 +0000)]
Correctly handle files smaller than an ELF header; reported by drahn@bitrig
mlarkin [Sun, 18 May 2014 20:16:29 +0000 (20:16 +0000)]
_PIC is an optional method, so we shouldn't be aborting the rest of the
MADT processing if it is missing. This, combined with an earlier diff
from kettenis@ to fix MADT interrupt polarity processing, should resolve
the issues reported with QEMU/KVM ACPI power button presses when using
OpenBSD guests.
ok kettenis@, pirofti@
tested by myself and kettenis for the past few weeks
miod [Sun, 18 May 2014 19:35:04 +0000 (19:35 +0000)]
If you need to allocate `a + b' bytes of memory, then don't allocate `a + b*2',
this is confusing and unnecessary.
Help (coz I got confused) and ok guenther@ beck@
miod [Sun, 18 May 2014 19:30:36 +0000 (19:30 +0000)]
Make sure UTF8_getc() is invoked with the proper buffer size.
ok beck@ guenther@
espie [Sun, 18 May 2014 17:50:11 +0000 (17:50 +0000)]
use reallocarray
okay ian@, chl@
tedu [Sun, 18 May 2014 17:49:47 +0000 (17:49 +0000)]
factor out a bit of the chunk index code and use it to make sure that a
freed chunk is actually freeable immediately. catch more errors.
hints/ok otto
mpi [Sun, 18 May 2014 17:10:27 +0000 (17:10 +0000)]
Do not pass an xfer pointer to the timeout routine checking for root hub
status changes because it might be freed when detaching the root uhub(4).
Also do not reschedule a timeout if the pipe is being aborted.
Finally do not add more code to retrieve the 'bInterval' value of the
root hub endpoint descriptor since this value is hardcoded in the uhci(4)
driver.
jsing [Sun, 18 May 2014 16:43:18 +0000 (16:43 +0000)]
Out of all the openssl apps, just one of them has code to mess around with
electric fence externs. Nuke from orbit.
ok miod@
espie [Sun, 18 May 2014 16:36:41 +0000 (16:36 +0000)]
use reallocarray
okay miod@, millert@
jsing [Sun, 18 May 2014 16:34:04 +0000 (16:34 +0000)]
W_OK comes via unistd.h - nuke several #ifndef W_OK that attempt to handle
missing defines.
jsing [Sun, 18 May 2014 16:26:02 +0000 (16:26 +0000)]
Remove unused include.
jsing [Sun, 18 May 2014 16:21:03 +0000 (16:21 +0000)]
Group and sort includes.
Currently "apps.h" needs to be included before any of the openssl headers
and there are several cases where the code depends on headers that are
included by something included by apps.h...
More untangling saved for later on.
miod [Sun, 18 May 2014 16:19:09 +0000 (16:19 +0000)]
Plug memory leaks upon error in OCSP_url_svcloc_new().
ok jsing@
miod [Sun, 18 May 2014 16:18:13 +0000 (16:18 +0000)]
In OCSP_sendreq_bio(), cope with OCSP_sendreq_new() returning NULL.
ok jsing@
miod [Sun, 18 May 2014 16:13:48 +0000 (16:13 +0000)]
In dtls1_reassemble_fragment() and dtls1_process_out_of_seq_message(), in case
of error, make sure we do not free pitem which is still linked into the
pqueue.
In the same vain, only free `frag' if we allocated it in this function.
Help and ok beck@
miod [Sun, 18 May 2014 16:10:26 +0000 (16:10 +0000)]
In ssl3_send_certificate_request(), when adding the extra payload if
NETSCAPE_HANG_BUG is defined, make sure we BUF_MEM_grow() the buffer to
accomodate for the payload size.
Issue reported by David Ramos; ok beck@
miod [Sun, 18 May 2014 16:08:37 +0000 (16:08 +0000)]
Make sure ssl3_setup_buffers() does not return upon error with a freed
pqueue still chained, by inserting it into the list only after all possible
failure conditions have been avoided.
Reported and fix proposed by David Ramos; ok beck@
stsp [Sun, 18 May 2014 15:27:28 +0000 (15:27 +0000)]
Fix tail packet check in elantech v3 touchpad code. Due to a typo this
code was masking out bits which were also tested in the same expression.
Reported by cppcheck via jsg@ (Expression '(X & 0xfc) != 0x2' is always true).
Tested with Elantech Touchpad, version 3, firmware 0x250f00.
krw [Sun, 18 May 2014 15:17:50 +0000 (15:17 +0000)]
Don't add a lease to the leases TAILQ more than once. It tends to
make writing out the leases file go into an infinite loop until
/var is full.
Reported by Roman Gorelov via bugs@
Fix tested & ok stsp@ uwe@
todd [Sun, 18 May 2014 14:06:23 +0000 (14:06 +0000)]
sync
jsing [Sun, 18 May 2014 11:36:24 +0000 (11:36 +0000)]
bio_err is initialised in main, hence there is no need to initialise it
in each application (since there is no longer a non-monolithic mode).
In typical OpenSSL fashion, the code is inconsistent and there are multiple
ways that bio_err was initialised - none of them actually checked to see if
the initialisation actually succeeded. Additionally, it is worth noting
that in at least two cases bio_err was also being used before it would have
been initialised.
ok miod@
jsing [Sun, 18 May 2014 11:28:41 +0000 (11:28 +0000)]
More KNF.
miod [Sun, 18 May 2014 11:20:08 +0000 (11:20 +0000)]
Omit -Werror if building with gcc3, for it triggers "redefinition of `pqueue'"
warnings now that pqueue.h has been removed from public scope.
stsp [Sun, 18 May 2014 10:52:16 +0000 (10:52 +0000)]
Make rtsx(4) attach to RTL8402 devices. Reported as working by Eric Huiban.
stsp [Sun, 18 May 2014 10:49:59 +0000 (10:49 +0000)]
regen
stsp [Sun, 18 May 2014 10:49:28 +0000 (10:49 +0000)]
Add PCI ID for Realtek RTL8402 card reader.
miod [Sun, 18 May 2014 09:39:18 +0000 (09:39 +0000)]
No need to check for NULL before invoking free(); use calloc() when
applicable; further simplify pqueue_find().
From Dimitris Papastamos on tech@
espie [Sun, 18 May 2014 09:29:54 +0000 (09:29 +0000)]
use reallocarray and friends. okay miod@
sthen [Sun, 18 May 2014 08:23:27 +0000 (08:23 +0000)]
sync openssl(1) with changes already made in src/lib/libssl/doc, reminded by
jmc: move from 1024 to 2048 bit key defaults; see genrsa.c 1.26 et al.
espie [Sun, 18 May 2014 08:10:00 +0000 (08:10 +0000)]
use reallocarray
okay chl@, tedu@
espie [Sun, 18 May 2014 08:08:50 +0000 (08:08 +0000)]
a bit more reallocarray (and kill ecalloc, which isn't used)
okay chl@
jsg [Sun, 18 May 2014 05:27:49 +0000 (05:27 +0000)]
There is no need to compile nginx with -O1 on vax anymore.
Since the change in rev 1.5 vax has switched to gcc3,
MAXDSIZ has increased and pcre_exec now builds with -O2.
ok miod@
chl [Sun, 18 May 2014 05:08:07 +0000 (05:08 +0000)]
replace some (type *)0 to NULL
ok krw@ millert@
miod [Sat, 17 May 2014 23:21:36 +0000 (23:21 +0000)]
One more mistake introduced in 1.65
chl [Sat, 17 May 2014 21:37:51 +0000 (21:37 +0000)]
remove unused variables
ok miod@
miod [Sat, 17 May 2014 21:32:19 +0000 (21:32 +0000)]
...and it needs a guard against NULL as well.
miod [Sat, 17 May 2014 21:21:17 +0000 (21:21 +0000)]
Fix previous commit by making sure we are not dereferencing uninitialized
pointers.
chl [Sat, 17 May 2014 20:31:07 +0000 (20:31 +0000)]
remove unused variable
ok tedu@
tedu [Sat, 17 May 2014 20:15:50 +0000 (20:15 +0000)]
does sunos not have memmove? sorry, i guess sunos will not have libressl.
ok guenther miod
chl [Sat, 17 May 2014 20:07:54 +0000 (20:07 +0000)]
add missing header needed by str* and mem* functions
ok gilles@
espie [Sat, 17 May 2014 20:05:07 +0000 (20:05 +0000)]
fix alloc idioms. watch out for emalloc that's actually a calloc in disguise
okay chl@ tedu@
miod [Sat, 17 May 2014 19:58:49 +0000 (19:58 +0000)]
unchecked malloc() return value in OCSP_sendreq_new().
miod [Sat, 17 May 2014 19:56:35 +0000 (19:56 +0000)]
KNF
chl [Sat, 17 May 2014 19:44:13 +0000 (19:44 +0000)]
add missing header needed for mem{cpy,cmp}()
ok tedu@
tedu [Sat, 17 May 2014 18:16:14 +0000 (18:16 +0000)]
correctly match size and buffer. from enh at google
tedu [Sat, 17 May 2014 17:35:04 +0000 (17:35 +0000)]
simple reallocarray to check multiplies. ok guenther
guenther [Sat, 17 May 2014 17:26:24 +0000 (17:26 +0000)]
When looking at another process, use the data from struct process
instead of peeking into the thread level
ok tedu@
tedu [Sat, 17 May 2014 15:57:52 +0000 (15:57 +0000)]
no no md2
tedu [Sat, 17 May 2014 15:48:14 +0000 (15:48 +0000)]
no no seed
tedu [Sat, 17 May 2014 15:44:28 +0000 (15:44 +0000)]
deleted seed some time ago
chl [Sat, 17 May 2014 15:19:17 +0000 (15:19 +0000)]
remove unused variable
ok tedu@
chl [Sat, 17 May 2014 15:18:06 +0000 (15:18 +0000)]
Add missing header needed by isdigit().
While there, sort headers.
ok tedu@
tedu [Sat, 17 May 2014 13:27:55 +0000 (13:27 +0000)]
refactor some of the des crypt code to use fewer globals.
ok miod
stsp [Sat, 17 May 2014 12:54:57 +0000 (12:54 +0000)]
Fix kernel build with EHCI_DEBUG by removing a non-compiling DPRINTF().
Since no DPRINTF() call remains also remove its definition (requested by mpi@).
ok mpi, and also reported by Jean-Philippe Ouellet
ratchov [Sat, 17 May 2014 12:45:53 +0000 (12:45 +0000)]
fix format mistakes visible if AZALIA_DEBUG is defined
ratchov [Sat, 17 May 2014 12:43:18 +0000 (12:43 +0000)]
fix format mistakes visible if UAUDIO_DEBUG is defined
ratchov [Sat, 17 May 2014 12:40:25 +0000 (12:40 +0000)]
fix format mistakes visible if AUICH_DEBUG is defined
ratchov [Sat, 17 May 2014 12:23:46 +0000 (12:23 +0000)]
unbreak build with AUDIO_DEBUG: use %lx for bus_addr_t
miod [Sat, 17 May 2014 12:22:09 +0000 (12:22 +0000)]
In pmap_remove_page(), do not keep wired mappings; the callers want them gone.
miod [Sat, 17 May 2014 12:20:55 +0000 (12:20 +0000)]
Remove "WARNING:" prefix from SYSFAIL and ACFAIL messages, so that, in case
of ACFAIL, what reaches the console is explicit enough.
ratchov [Sat, 17 May 2014 12:19:36 +0000 (12:19 +0000)]
unbreak build: exchange %b args order, from miod@
miod [Sat, 17 May 2014 12:13:44 +0000 (12:13 +0000)]
Lower VUPS value for delay() to behave closer to the intent.
This is temporary until vax grows code to compute a proper delay const
at runtime, but will help a bit, especially on KA46 where cd(4) end up
detected as sd(4) due to too short delays.
sthen [Sat, 17 May 2014 12:12:05 +0000 (12:12 +0000)]
switch RSA key generation default to 2048 bits (matching ssh); ok miod@
miod [Sat, 17 May 2014 12:08:37 +0000 (12:08 +0000)]
Print the rpb device type and unit values if the root device can't be found
by the kernel.
jmatthew [Sat, 17 May 2014 11:51:21 +0000 (11:51 +0000)]
When the firmware tells us a loop id is already in use, add the device
there to our port lists, then find the next available loop id and carry on
rather than restarting. This way, we're less likely to get stuck looping
when the firmware behaves inconsistently. We also don't need to treat domain
controller logins specially.
additional complications discovered by deraadt@
aoyama [Sat, 17 May 2014 10:06:43 +0000 (10:06 +0000)]
Change the return value of interrupt handler to 0 when nothing is
processed on this driver.
"This is correct" miod@
bluhm [Sat, 17 May 2014 08:12:53 +0000 (08:12 +0000)]
When parsing a numerical value for the TOS bits, make sure that it
is in a valid range.
OK henning@
jmc [Sat, 17 May 2014 06:26:44 +0000 (06:26 +0000)]
better description for -A and -a, with help from millert;
ok millert sobrado guenther
jmc [Fri, 16 May 2014 22:11:00 +0000 (22:11 +0000)]
update NAME and MLINKS for previous;
jmc [Fri, 16 May 2014 21:50:46 +0000 (21:50 +0000)]
more STANDARDS fleshing; ok sobrado millert
tedu [Fri, 16 May 2014 21:34:48 +0000 (21:34 +0000)]
document crypt_checkpass
tedu [Fri, 16 May 2014 21:28:15 +0000 (21:28 +0000)]
add prototypes for new crypt functions
todd [Fri, 16 May 2014 21:22:13 +0000 (21:22 +0000)]
sync
mpi [Fri, 16 May 2014 19:00:18 +0000 (19:00 +0000)]
Reduce the difference between HC drivers by always passing a generic
usbd_xfer pointer to the routines processing finished transfers to make
it crystal clear that the timeout and abort logic is the same everywhere.
tedu [Fri, 16 May 2014 18:35:01 +0000 (18:35 +0000)]
save up the failures for the end in the argc == 0.
help ok espie
mpi [Fri, 16 May 2014 18:17:03 +0000 (18:17 +0000)]
There is no need to remember which usb(4) device is the child of an USB
host controller because autoconf(9) already does it.
tedu [Fri, 16 May 2014 17:46:07 +0000 (17:46 +0000)]
a few changes for new code to match local style
tedu [Fri, 16 May 2014 17:42:24 +0000 (17:42 +0000)]
checksum mode can use magic pubkey finder too
millert [Fri, 16 May 2014 17:30:28 +0000 (17:30 +0000)]
Zero out grp before re-using it (new ones are already cleared by
calloc). We only reuse grp when there is an unresolvable host.
Fixes a bug where if a host in a netgroup is unresolvable then
entire netgroup is ignore. OK krw@
espie [Fri, 16 May 2014 16:18:33 +0000 (16:18 +0000)]
rework -C (perfect candidate for using hash tables, really):
insert files we want to check into a hash,
parse SHA256 message on the fly, delete entries whose checksum match,
then display entries that failed.
This completely avoids allocating temporary storage for file names and checksums
and removes the quadratic match (argv[i] vs line[n]).
okay tedu@
jsing [Fri, 16 May 2014 14:45:18 +0000 (14:45 +0000)]
Include openssl/evp.h for OPENSSL_add_all_algorithms_noconf().
From Brent Cook.
jsing [Fri, 16 May 2014 14:41:05 +0000 (14:41 +0000)]
Fix various issues that result in compiler warnings.
Prompted by a diff from Brent Cook.
jsing [Fri, 16 May 2014 14:24:36 +0000 (14:24 +0000)]
When OPENSSL_LOAD_CONF was added it ended up with more #if 0 code,
more #ifdefs and a new source file that contains a single function.
Nuke the #if 0 code that is now a macro and move the single function in
evp_acnf.c to c_all.c, which is where the other code lives. While here,
tidy evp.h slightly, remove an unnecessary #ifdef __OpenBSD__ and nuke
a comment that is now a lie.
ok miod@
sthen [Fri, 16 May 2014 14:12:19 +0000 (14:12 +0000)]
use $2b in the bcrypt example, ok tedu
sasano [Fri, 16 May 2014 14:05:39 +0000 (14:05 +0000)]
It seems that there is new and old revision of CH340.
Previous uchcom(4) driver targeted old one, and new one could not work
because of uchcom_set_line_control() broke the value of
UCHCOM_REG_LCR1(0x18).
To support new CH340, uchcom_set_line_control() and uchcom_reset_chip()
have been overhauled. Current uchcom(4) does not change the value of
UCHCOM_REG_LCR1 register, it means even/odd parity mode is no longer
supported with old CH340.
ok by mpi@
dcoppa [Fri, 16 May 2014 13:56:15 +0000 (13:56 +0000)]
regen
dcoppa [Fri, 16 May 2014 13:53:35 +0000 (13:53 +0000)]
Add Alcatel OneTouch L100V
Diff from Lukasz Ratajski
ok sthen@
krw [Fri, 16 May 2014 13:44:09 +0000 (13:44 +0000)]
Make the df calculations of available space the same as that done in
ffs_statfs().
In 1998, with /usr/src/sys/ufs/ffs/ffs_vfsops.c r1.16, mickey@
improved the calculation of available blocks in ffs_statfs().
Yesterday guenther@ noticed that this fix had not been applied to the
same calculation being done by df(8) when run against raw
devices. Which meant different values were displayed depending on
which device/file system you gave to df.
ok otto@
mikeb [Fri, 16 May 2014 08:46:23 +0000 (08:46 +0000)]
improve logging messages and style; requested by and ok reyk
mpi [Fri, 16 May 2014 08:21:54 +0000 (08:21 +0000)]
Do not put any link-layer address on the per-ifp lists or on the RB-
Tree.
Since interfaces only support one link-layer address accessible via the
if_sadl member, there's no need to have it elsewhere. This improves
various address lookups because the first element of the list, the link-
layer address, won't necessarily be discarded.
Finally remove the empty netmask associated to every link-layer address.
This hack was needed to (ab)use the address & netmask comparison code to
do a strcmp() on the interface name embedded in the sdl_data field.
ok henning@, claudio@
miod [Thu, 15 May 2014 21:07:10 +0000 (21:07 +0000)]
Fix memory leaks upon failure.
ok beck@
miod [Thu, 15 May 2014 21:06:10 +0000 (21:06 +0000)]
Replace ASN1_GENERALIZEDTIME_adj(), ASN1_UTCTIME_adj() and
ASN1_TIME_to_generalizedtime() with wrappers around their former
implementations, making sure memory allocated is freed in all failure cases.
help and ok from beck@ and Brendan MacDonell.
jmc [Thu, 15 May 2014 19:42:56 +0000 (19:42 +0000)]
more for STANDARDS: unsupported lists (noted by guenther) and env vars;
chl [Thu, 15 May 2014 19:40:42 +0000 (19:40 +0000)]
remove unused variable
ok espie@
miod [Thu, 15 May 2014 19:39:46 +0000 (19:39 +0000)]
No longer make pqueue.h a public (and installed) header file. Replace its
inclusion in <ssl/dtls1.h> by the benign
`struct _pqueue; typedef struct _pqueue *pqueue;'.
Note that said dtls1.h should probably be considered a private header as well...
ok beck@
jmc [Thu, 15 May 2014 19:38:06 +0000 (19:38 +0000)]
flesh out STANDARDS; ok guenther
jmc [Thu, 15 May 2014 19:36:45 +0000 (19:36 +0000)]
use <> for tables;
original report from creamy;
diff from Frank Brodbeck, tweaked
miod [Thu, 15 May 2014 19:23:14 +0000 (19:23 +0000)]
Get pqueue.h from the libssl sources rather than relying upon it being
installed.
chl [Thu, 15 May 2014 19:18:23 +0000 (19:18 +0000)]
- typo in comment
- memset() is not needed after a calloc()
- add missing DBG_ENTER;
- use SEEK_SET instead of 0 in an lseek() call
ok krw@