src/lib/libc/stdlib/malloc.c, branch releng/9.3 FreeBSD source tree https://cgit-dev.freebsd.org/src/atom?h=releng%2F9.3 2013-07-04T14:26:42Z Protect the #define MALLOC_PRODUCTION directive in case it was also 2013-07-04T14:26:42Z Dag-Erling Smørgrav des@FreeBSD.org 2013-07-04T14:26:42Z urn:sha1:42bdab678229c7020cfdc5bcc7988ba07dbebb15 defined on the command line (due to MALLOC_PRODUCTION=yes in src.conf). MFC r229848: 2012-11-22T15:19:53Z Ed Schouten ed@FreeBSD.org 2012-11-22T15:19:53Z urn:sha1:a48d580e689b940b0d6abc376310842a273cdcd8 Add aligned_alloc(3). The C11 folks reinvented the wheel by introducing an aligned version of malloc(3) called aligned_alloc(3), instead of posix_memalign(3). Instead of returning the allocation by reference, it returns the address, just like malloc(3). I'm MFCing this now, as it seems aligned_alloc(3) is needed to make the new version of libc++ work, which was merged back to FreeBSD 9 in r243376. Requested by: dim Ensure that the beginning of the DSS is aligned on a chunk boundary. 2012-06-13T19:04:39Z John Baldwin jhb@FreeBSD.org 2012-06-13T19:04:39Z urn:sha1:d587647ea598e402e986db400470640ff056a475 If the _end symbol used an address with the low bit set, then the initial arena could end up with corrupted rb trees causing a crash during the first call to malloc(). This is a direct commit to stable/9 as it does not affect the version of malloc in HEAD. Reviewed by: jasone MFC after: 1 week MFC r228540: 2011-12-31T12:58:21Z Dimitry Andric dim@FreeBSD.org 2011-12-31T12:58:21Z urn:sha1:6194f5339c63c5b841848e34100e11664b686912 Since clang does not support the tls_model attribute used in malloc.c yet (see LLVM PR 9788), and warns about it, rub it out for now. When clang grows support for this attribute, I will revert this again. Remove extra debuggin gsupport that is turned on for head but turned off 2011-10-15T21:23:04Z Ken Smith kensmith@FreeBSD.org 2011-10-15T21:23:04Z urn:sha1:b8975c9bfdbad79940e9b54e748e23520d6b571d for stable branches: - shift to MALLOC_PRODUCTION - turn off automatic crash dumps - remove kernel debuggers, INVARIANT*[1], WITNESS* from GENERIC kernel config files [1] INVARIANT* left on for ia64 at least temporarily, marcel@ will test to see if they are still required as they had been for stable/8. Approved by: re (implicit) Change sparc64 to use the initial exec TLS model, too. This avoids random 2011-06-21T19:15:23Z Marius Strobl marius@FreeBSD.org 2011-06-21T19:15:23Z urn:sha1:1e01d8915a4c0bdf21dad8be2ee6d14682df8594 assertion failures in _malloc_thread_cleanup(). Now that TLS generally is available on sparc64 since r219534 turn on 2011-03-11T21:35:38Z Marius Strobl marius@FreeBSD.org 2011-03-11T21:35:38Z urn:sha1:77ccab511d852afcfc759b6ef9c69739588f5396 support for it. Note that while sparc64 also supports the static TLS model and thus tls_model("initial-exec"), using the default model turned out to yield slightly better buildstone performance. Use aux vector to get values for SSP canary, pagesize, pagesizes array, 2010-08-17T09:13:26Z Konstantin Belousov kib@FreeBSD.org 2010-08-17T09:13:26Z urn:sha1:2793b018446552e71e9125dc39366f9dbf9042ba number of host CPUs and osreldate. This eliminates the last sysctl(2) calls from the dynamically linked image startup. No objections from: kan Tested by: marius (sparc64) MFC after: 1 month Provide 64-bit PowerPC support in libc. 2010-07-10T14:45:03Z Nathan Whitehorn nwhitehorn@FreeBSD.org 2010-07-10T14:45:03Z urn:sha1:840b91cc52085d26c112e99d733f71eb92c3ff54 Obtained from: projects/ppc64 Rewrite red-black trees to do lazy balance fixup. This improves 2010-02-28T22:57:13Z Jason Evans jasone@FreeBSD.org 2010-02-28T22:57:13Z urn:sha1:15f8d4975676d724c229bdd4121ea06e27cc1674 insert/remove speed by ~30%.