From: Rich Felker Date: Mon, 30 Mar 2015 22:40:39 +0000 (-0400) Subject: release 1.0.5 X-Git-Tag: v1.0.5 X-Git-Url: https://git.librecmc.org/?a=commitdiff_plain;h=56be2e4c010944259bfa1b52b5044b517e31c200;p=oweals%2Fmusl.git release 1.0.5 --- diff --git a/VERSION b/VERSION index ee90284c..90a27f9c 100644 --- a/VERSION +++ b/VERSION @@ -1 +1 @@ -1.0.4 +1.0.5 diff --git a/WHATSNEW b/WHATSNEW index 7220fb45..f4762083 100644 --- a/WHATSNEW +++ b/WHATSNEW @@ -1267,3 +1267,39 @@ arch-specific bugs fixed: - missing barriers in atomics on mips, powerpc, and microblaze - TLS relocation bug in powerpc dynamic linker + + + +1.0.5 release notes + +compatibility: +- dynamic linker now honors DT_RUNPATH without DT_RPATH (new binutils) + +bugs fixed: +- stack-based buffer overflow in inet_pton (CVE-2015-1817) +- regcomp mishandling of high bytes after backslash +- regcomp miscompiled character class brace-repetitions +- use of uninitialized memory with application-provided thread stacks +- failure of dn_expand to null-terminate name for crafted DNS packets +- buffer overflow in swab with odd argument +- missing barrier in pthread_once fast path +- wordexp bad character checker mis-counted parentheses +- malloc init code could deadlock due to race condition +- getspnam_r returned results for partial username matches +- mishandling of negative non-whole-hour TZ offsets +- printf failed to report or stop on write errors +- syslog failed to check for connect error +- fchmodat was subject to fd leak race (missing O_CLOEXEC) +- fchmodat failed to report EOPNOTSUPP in race path +- wrong behavior in various zero-length stdio operations +- wrong return value on overflow in some strtoul-family functions +- incorrect sequence generation in the rand48 family of prng functions +- wrong printf formatting for %#.0o with value zero +- sched_getaffinity left uninitialized data in output bit array +- wrong return values for pthread_getaffinity_np and pthread_setaffinity_np +- broken CPU_EQUAL macro in sched.h + +arch-specific bugs fixed: +- usage of sahf instruction on x86_64 crashed on some early cpu models +- mips fesetenv did not handle FE_DFL_ENV +- mips POLLWRNORM and POLLWRBAND macros had wrong values