cinap_lenrek [Wed, 28 Oct 2015 16:09:22 +0000 (17:09 +0100)]
devtls: reject SHA2_256 mac for SSL, but TLS is fine
sha256 is only defined for TLS1.2, however, technically, theres
no reason not to use it in TLS1.0/TLS1.1. the choice is up to
tlshand and pushtls, not the kernel.
cinap_lenrek [Thu, 15 Oct 2015 11:21:30 +0000 (13:21 +0200)]
lib9p: do not override Srv.end in listensrv(), simplify srvclose() and recounting
listensrv() used to override Srv.end() with its own handler
to free the malloc'd Srv structure and close the fd. this
makes it impossible to register your own cleanup handler.
instead, we introduce the private Srv.free() handler that
is used by listensrv to register its cleanup code. Srv.free()
is called once all the srv procs have been exited and all
requests on that srv have been responded to while Srv.end()
is called once all the procs exited the srv loop regardless
of the requests still being in flight.
cinap_lenrek [Fri, 9 Oct 2015 17:22:53 +0000 (19:22 +0200)]
cwfs: fix wstat() failing to mark block dirty when noatime is set
code assumed the accessdir() call would always mark the block dirty, but
this is not the case when noatime flag is enabled. this was reported by
michael in bug:
cinap_lenrek [Fri, 9 Oct 2015 03:10:47 +0000 (05:10 +0200)]
wifi: quote value of parsed ether options
introduce wificfg() function to convert ether->opt[] strings
to wifictl messages, which needs quoting for the value. so
etherX=type=iwl essid='something with spaces' works.
cinap_lenrek [Thu, 8 Oct 2015 15:14:07 +0000 (17:14 +0200)]
cc: restore side(), but do not consider OINDEX as side effect free
from charles forsuth:
because the previous version thought OINDEX might have a side effect, it
stopped it building a tower of them.
probably the best thing is to limit that anyway, since each one consumes
2-3 registers, so towering them can
keep even more active, and the x86 hasn't got that many.
the quick hack is to return that case to the earlier state by treating
OINDEX as a side-effect in side().
it's not a bad thing to do in the OSTRUCT case, for similar reasons: it's
better to collapse the indexed pointer
into a direct register, instead of repeating the indexing operation through
the copying of the value.
OINDEX isn't a machine-independent operation, so it doesn't affect the uses
in ../cc
cinap_lenrek [Tue, 6 Oct 2015 04:20:01 +0000 (06:20 +0200)]
5c/6c/8c/vc: import various changes from charles forsyth
- cover more cases that have no side effects
- ensure function has complex FNX
- pull operators out of OFUNC level
- rewrite OSTRUCT lhs to avoid all side-effects, use regalloc() instead of regret()
cinap_lenrek [Mon, 5 Oct 2015 02:31:31 +0000 (04:31 +0200)]
dossrv: handle file offsets > 2^31
maximum file size is 4GB-1 as the file length is stored in
a 32 bit long. make sure it doesnt overflow on write or
or truncate. interpret the file length as unsigned. pass
vlong to readfile()/writefile()/truncfile() so we can
handle overflows and not just ignore the upper bits.
- fix missing runlock(ifc) when ifcid != a->ifcid in rxmitsols() (thanks erik quanstro)
- don't leak packets when transfering blocks from arp entry hold list to droplist
- free rest of droplist when bwrite() errors in arpenter(), remove useless checks (ifc != nil)
- free arp entry hold list from cleanarpent()
- consistent use of nil for pointers
sprint() will replace invalid utf8 sequences with U+FFFD
which caused directory reads and stats to return the wrong
filename. just strcpy the name bytes.
cpp: fix memory corruption due to input buffer relocation
the dynamic input buffer resize code (fillbuf()) is broken as
the calling code assumes that memory wont relocate. instead
of trying to work out all the cases where this happens, i'm
getting rid of fillbuf() and just read the whole file into
memory in setsource().
the bug could be reproduced with something as simple as:
libsec: fix memory leaks in X509req() and X509gen() and return malloced pointer
X509req() and X509gen() used to leak memory, and had no way for
the caller to free the allocated certificate/certificate request
buffer returned. this is not critical as these functions are only
used in short lived rsa(2) helper programs. but i prefer to have
library routines not leak memory as one does not know in advance
where the code is going to be used.
libsec: implement client certificate authentication for tls1.2
we used to negotiate tls1.1 for client cert authentication because the
signature generation was not implemented for tls1.2. this is now fixed
and tls1.2 can be negotiated with client certs.
libsec: implement dh parameter signature verification, stop lying about non-rsa ciphers, fix memory leaks in X509 code
actually verify the diffie hellman parameter signature, this
comes in two flavours. TLS1.2 uses X509 signature with a
single hash specified by the signature algorithm field in
the signature itself and pre TLS1.2 where md5+sha1 hashes
of the signed blob are pkcs1 padded and encrypted with the
rsa private key.
stop advertizing non-rsa cipher suits (DSS and ECDSA), as
we have not implmenented them.
fix some memory leaks in X509 code while we'r at it.
libdraw, screenrc: bind devdraw and devmouse in screenrc instead of handling it in libdraw
libdraw was attempting to bind '#i' and '#m' to /dev when it could not find
/dev/mouse or /dev/draw. a library shouldnt be that clever and do namespace
manipulations on behalf of the caller. so instead, we setup the graphics
environment in screenrc on boot time.
libfis: fix wrong shift for lba40 (thanks geoff and charles)
---------- Forwarded message ----------
From: Charles Forsyth <charles.forsyth@gmail.com>
Date: 13 September 2015 at 12:38
Subject: fis bug
To: erik quanstrom <quanstro@quanstro.net>
geoff spotted something similar in sdiahci.c, but it's in libfis as well:
libsec: send sigature_algoritms extension for TLS1.2, order ciphers
TLS1.2 requires the client to send the list of supported
signature and hash algorithm pairs. some servers will simply
reject the client hello otherwise. note that we do not implement
any dh/ecdh param signature verification.
order the cipher list to strogest first. aes128 is actually more
secure than aes256.
tar: make z flag work, even when no file name was provided (thanks aiju)
tar used to infer compression type from the filenames extension, but when
no file name is given (stdin/stdout), the -z flag was ignored and no
compression filter applied. this changes tar to assume the default
gzip compression method when z is given and no file name is specified.
libthread: get rid of tprivalloc()/tprivfree()/tprivdata() and _workerdata() (thanks qrstuv)
these functions where undocumented and unused. especially
tprivfree() was buggy missing a unlock() call. theres not
much point in supporting these functions as theres
threaddata() and procdata().