Граф коммитов

437 Коммитов

Автор SHA1 Сообщение Дата
Gian-Carlo Pascutto 3b43637e56 Bug 1537781 - Trailing guard pages for normal allocations. r=glandium
Differential Revision: https://phabricator.services.mozilla.com/D27912

--HG--
extra : moz-landing-system : lando
2019-05-02 14:19:17 +00:00
Chris Martin ef8ae4fcb3 Bug 1052579 - Add ability to query ArenaID to mozjemalloc_ptr_info r=glandium
To ensure that any new JSString has its char buffer allocated in the new arena,
it is useful to be able to query a pointer and assert that it is in the
correct arena (at-least in Debug Build).

This adds the required functionality to mozjemalloc, and JSString can use it
for its new assertion in a later change.

Differential Revision: https://phabricator.services.mozilla.com/D25711

--HG--
extra : moz-landing-system : lando
2019-04-23 12:39:58 +00:00
Narcis Beleuzu cd1fcbc396 Backed out changeset 497561b76737 (bug 1052579) for bustages on mozjemalloc_types.h . CLOSED TREE 2019-04-22 19:00:58 +03:00
Chris Martin cb3f697d99 Bug 1052579 - Add ability to query ArenaID to mozjemalloc_ptr_info. r=glandium
To ensure that any new JSString has its char buffer allocated in the new arena,
it is useful to be able to query a pointer and assert that it is in the
correct arena (at-least in Debug Build).

This adds the required functionality to mozjemalloc, and JSString can use it
for its new assertion in a later change.

Differential Revision: https://phabricator.services.mozilla.com/D25711
2019-04-02 03:55:06 +00:00
Gian-Carlo Pascutto bfcbc83c13 Bug 1542290 - Decommit unused base_page allocations. r=glandium
Differential Revision: https://phabricator.services.mozilla.com/D26327

--HG--
extra : moz-landing-system : lando
2019-04-16 11:04:21 +00:00
Sylvestre Ledru 7f60810d86 Bug 1519636 - Reformat recent changes to the Google coding style r=Ehsan
# ignore-this-changeset

Differential Revision: https://phabricator.services.mozilla.com/D27245

--HG--
extra : moz-landing-system : lando
2019-04-12 13:14:25 +00:00
Petr Sumbera a65de26d20 Bug 1542758 - mozjemalloc.cpp should use posix_madvise on Solaris r=glandium
Differential Revision: https://phabricator.services.mozilla.com/D26529

--HG--
extra : moz-landing-system : lando
2019-04-09 08:20:48 +00:00
Gian-Carlo Pascutto d735d8aa70 Bug 1529922 - Add guard pages following huge allocations. r=glandium
Differential Revision: https://phabricator.services.mozilla.com/D23292

--HG--
extra : moz-landing-system : lando
2019-03-20 10:45:15 +00:00
David Major e5773183d6 Bug 1528074 - Remove MSVC warning flags that clang-cl doesn't understand r=chmanchester
Per the previous patch, clang-cl only understands five MSVC-style warning flags: 7219c7e9af/clang/include/clang/Driver/CLCompatOptions.td (L188-L197)

This patch removes the flags that clang-cl doesn't understand.

Differential Revision: https://phabricator.services.mozilla.com/D22588

--HG--
extra : moz-landing-system : lando
2019-03-13 20:19:08 +00:00
Masatoshi Kimura e59e3b3e20 Bug 1528651 - Fix some trivial warnings and re-enable warnings-as-errors on Windows in some directories. r=dmajor
--HG--
extra : source : 5ee9efeda9e67f6af0521be85266703763cc4c50
2019-02-19 20:42:11 +09:00
Mike Hommey ef3ad686ee Bug 1512504 - Remove support for MSVC. r=froydnj
Consequently, this removes:
- MOZ_LIBPRIO, which is now always enabled.
- non_msvc_compiler, which is now always true.
- The cl.py wrapper, since it's not used anymore.
- CL_INCLUDES_PREFIX, which was only used for the cl.py wrapper.
- NONASCII, which was only there to ensure CL_INCLUDES_PREFIX still
  worked in non-ASCII cases.

This however keeps a large part of detecting and configuring for MSVC,
because we still do need it for at least headers, libraries, and midl.

Depends on D19614

Differential Revision: https://phabricator.services.mozilla.com/D19615

--HG--
extra : moz-landing-system : lando
2019-02-14 21:45:27 +00:00
Ehsan Akhgari e5e885ae31 Bug 1521000 - Part 2: Adjust our clang-format rules to include spaces after the hash for nested preprocessor directives r=sylvestre
# ignore-this-changeset

--HG--
extra : amend_source : 7221c8d15a765df71171099468e7c7faa648f37c
extra : histedit_source : a0cce6015636202bff09e35a13f72e03257a7695
2019-01-18 10:16:18 +01:00
Mike Hommey a8d4234310 Bug 1511251 - Remove redundant and costly assert. r=njn
The diagnostic assert (so fortunately, it doesn't impact release builds)
as added in bug 1405159, but is costly because it uses the modulus of
the division with a variable integer, which is a slow operation.
However, in arena_run_reg_dalloc, we end up doing the same diagnostic
assert, in a different form: after performing the division in a faster
manner, we assert that the result, multiplied by the diviser, returns
the original number.

Differential Revision: https://phabricator.services.mozilla.com/D13501

--HG--
extra : moz-landing-system : lando
2018-12-05 14:45:52 +00:00
Sylvestre Ledru 265e672179 Bug 1511181 - Reformat everything to the Google coding style r=ehsan a=clang-format
# ignore-this-changeset

--HG--
extra : amend_source : 4d301d3b0b8711c4692392aa76088ba7fd7d1022
2018-11-30 11:46:48 +01:00
Chris Martin adde9e8556 Bug 1402282 - Change jemalloc to use secure random private arena ids r=glandium
Previously the id for a new arena was just a counter that increased by one
every time. For hardening purposes, we want to make private arenas use a secure
random ID, so an attacker will have a more difficult time finding the memory
they are looking for.

Differential Revision: https://phabricator.services.mozilla.com/D10158

--HG--
extra : moz-landing-system : lando
2018-11-21 01:52:26 +00:00
Mike Hommey c4ea7f7d5a Bug 1507035 - Fix run sizes for size classes >= 16KB on systems with large pages. r=njn
Differential Revision: https://phabricator.services.mozilla.com/D11836

--HG--
extra : moz-landing-system : lando
2018-11-14 06:58:53 +00:00
arthur.iakab 27754a7d12 Backed out 2 changesets (bug 1402282) for turning multiple browser chrome bugs into permafail
Backed out changeset db7059b57f92 (bug 1402282)
Backed out changeset cea1d44ac776 (bug 1402282)
2018-11-05 17:56:37 +02:00
Chris Martin 3824b0d43e Bug 1402282 - Change jemalloc to use secure random arena ids r=glandium
Previously the id for a new arena was just a counter that increased by one
every time. For hardening purposes, we want to make the new counter a secure
random ID, so an attacker will have a more difficult time finding the memory
they are looking for.

Differential Revision: https://phabricator.services.mozilla.com/D10158

--HG--
extra : moz-landing-system : lando
2018-11-05 00:27:31 +00:00
Chris Martin a1d5bd1909 Bug 1402282 - Clang-format on mozjemalloc r=glandium
Differential Revision: https://phabricator.services.mozilla.com/D10794

--HG--
extra : moz-landing-system : lando
2018-11-05 00:26:00 +00:00
Randell Jesup 0f390bf1a5 Bug 1480430: Modify jemalloc to allow dynamic replacement r=glandium 2018-10-09 22:28:37 -04:00
Mike Hommey 5f59918688 Bug 1482797 - Don't use MADV_FREE on Linux until we support it properly. r=njn 2018-08-15 21:33:57 +09:00
Masatoshi Kimura 3b21b7868b Bug 1090497 - Re-enable warnings as errors on clang-cl. r=froydnj
--HG--
extra : rebase_source : c09366fb93e5b0f72abe1e99d3094e3d96a934fb
extra : intermediate-source : 5950c9d63c3b4fd63a25464a7b50944aaec7079f
extra : source : ca1b9a2bcc4381795f556fea2fb59066567c30f3
2018-07-31 22:10:07 +09:00
Brian Hackett 29bb91306b Bug 1465452 Part 5 - Don't record some jemalloc atomics, r=njn.
--HG--
extra : rebase_source : a1dacd30546372e836b69e51f200e4c3e1295930
2018-07-21 14:30:33 +00:00
Jon Coppeard 858a13bce3 Bug 1468767 - Check result of calling vm_copy() r=njn 2018-06-14 14:58:45 -07:00
Mike Hommey 7c246fac68 Bug 1460838 - Avoid static initializers in mozjemalloc with MSVC. r=njn
--HG--
extra : rebase_source : dd2106192a90fbade6f89dfa1169c6e9ab3a553b
2018-05-24 11:23:10 +09:00
Tom Ritter 4e3daa47c1 Bug 1460720 Do not define _aligned_malloc - instead define _aligned_malloc_impl and export _aligned_malloc r=glandium
MozReview-Commit-ID: 3EwAd81Iz7r

--HG--
extra : rebase_source : 899303e4c5db39b24451692f59a9d3bd1f9fd5a2
2018-05-15 11:10:48 -05:00
Masatoshi Kimura e98b2c42f0 Bug 1445601 - Stop using LoadLibraryA in replace_malloc. r=glandium
MozReview-Commit-ID: 8EzDtCIlg7F

--HG--
extra : rebase_source : cf909f472c1c0007b2ff759d011435b8b6bc0f37
2018-03-25 13:12:03 +09:00
Tom Ritter 5fda6df793 Bug 1446466 Remove Nightly-only restriction on jemalloc arena implementation r=glandium
MozReview-Commit-ID: CC2cftngmli

--HG--
extra : rebase_source : 5cc5d5b0638b29074cc0e497f4669ebabcf6578a
2018-03-21 20:53:46 -05:00
Tom Ritter 51a8daef9d Bug 1446466 Crash if moz_dispose_arena is called, and comment out all callers r=glandium
Bug 1364359 is to fix a leaked arena. Until that is fixed; it is unsafe to
call moz_dispose_arena more than once.

MozReview-Commit-ID: KIby1RLtrPK

--HG--
extra : rebase_source : 6ea41001e9f0c4d5eb24ee678d6c1c0218991ac3
2018-03-21 20:49:35 -05:00
Nathan Froyd b68dfdbc2d Bug 1435407 - declare our wrapped delete definitions with noexcept(true); r=glandium
This behavior matches what gets used in mozalloc.h to define these
wrappers, and is particularly necessary for newer versions of clang to
not complain about our definitions.
2018-03-13 11:10:06 -05:00
Mike Hommey 5f11951b9b Bug 1441335 - Fix base allocator commit evaluation. r=njn
Base allocator commit stats were added in bug 515556, along other commit
stats, but they have actually been wrong since then: the committed count
is updated with the difference between pbase_next_addr and
base_next_decommitted *after* the latter is set to the former, making
the difference always 0.

--HG--
extra : rebase_source : a2aed523314549a37a61bd4ab300c98f198f9252
2018-02-27 07:39:34 +09:00
Mike Hommey ef4741aa26 Bug 1439470 - Remove some now unnecessary checks. r=njn
Since TreeNode::{Left,Right,Color} is always a valid call to make, we
don't need to check if for nullity before calling those functions.

This effectively kind of reverts some parts of bug 1412722.

--HG--
extra : rebase_source : 3deb316f463b51fdbb3aebc2e57e437018b3a829
2018-02-15 20:25:57 +09:00
Mike Hommey 6442687426 Bug 1439470 - Turn TreeNode(nullptr) into a "virtual" sentinel. r=njn
The code before bug 1412722 benefitted from the sentinel being an actual
node object, and some code paths ended up checking its color (always
black) or getting its right and left node, that always returned to the
sentinel.

When TreeNode currently contains a nullptr, all those lead to a null
deref if the calling code is not doing the right checks, which happens
to be the case in at least some places. Instead of relying on the
callers doing the right thing, make the TreeNode do the right thing when
it contains a nullptr, effectively acting as the sentinel in that case.

We additionally ensure that nothing in the calling code will be trying
to change the color or left/right pointers on the sentinel, which is an
extra safe net compared to the code before bug 1412722.

--HG--
extra : rebase_source : 09ab0bf8682092ef6d9a0a5921be3da787d0d548
2018-02-15 20:20:11 +09:00
Mike Hommey 1c0141e333 Bug 1439470 - Turn TreeNode into a smart pointer-like type. r=njn
This will allow the upcoming changes to add some safety back to the code
after bug 1412722.

--HG--
extra : rebase_source : 5033b8034cabaf5a7fdd578459588d5099402d02
2018-02-15 20:15:00 +09:00
Andreea Pavel e7ca112682 Backed out 3 changesets (bug 1439470) for failing automation.py on a CLOSED TREE
Backed out changeset c43ee00c3e6b (bug 1439470)
Backed out changeset cf9d00862149 (bug 1439470)
Backed out changeset f95559ae3134 (bug 1439470)
2018-02-20 13:39:28 +02:00
Mike Hommey a70ca542aa Bug 1439470 - Remove some now unnecessary checks. r=njn
Since TreeNode::{Left,Right,Color} is always a valid call to make, we
don't need to check if for nullity before calling those functions.

This effectively kind of reverts some parts of bug 1412722.

--HG--
extra : rebase_source : 172f1c042bdbb4d500e1afb4d57774ab76826876
2018-02-15 20:25:57 +09:00
Mike Hommey b4f9300f41 Bug 1439470 - Turn TreeNode(nullptr) into a "virtual" sentinel. r=njn
The code before bug 1412722 benefitted from the sentinel being an actual
node object, and some code paths ended up checking its color (always
black) or getting its right and left node, that always returned to the
sentinel.

When TreeNode currently contains a nullptr, all those lead to a null
deref if the calling code is not doing the right checks, which happens
to be the case in at least some places. Instead of relying on the
callers doing the right thing, make the TreeNode do the right thing when
it contains a nullptr, effectively acting as the sentinel in that case.

We additionally ensure that nothing in the calling code will be trying
to change the color or left/right pointers on the sentinel, which is an
extra safe net compared to the code before bug 1412722.

--HG--
extra : rebase_source : ac61ea259ac49bf76e2f8f6f54dda991498d4664
2018-02-15 20:20:11 +09:00
Mike Hommey 4489841597 Bug 1439470 - Turn TreeNode into a smart pointer-like type. r=njn
This will allow the upcoming changes to add some safety back to the code
after bug 1412722.

--HG--
extra : rebase_source : c906e9b3168fe738cba8a3de3fdf4efee1f0d4df
2018-02-15 20:15:00 +09:00
Mike Hommey c60ee1a610 Bug 1439474 - Make double-free crashes more identifiable. r=njn
- Turn MOZ_DIAGNOSTIC_ASSERTs related to double-frees into
MOZ_RELEASE_ASSERTs with a crash message making them more identifiable
than the asserted condition.
- In huge_dalloc, MOZ_RELEASE_ASSERT early, instead of letting
RedBlackTree::Remove end up crashing because the node is not in the
tree.

--HG--
extra : rebase_source : e051caaf371e88a9db6b5153f58c8a4aa4cde757
2018-02-20 11:39:04 +09:00
Mike Hommey 5c18ecad6b Bug 1438427 - Fix wrong change from bug 1412722 in RedBlackTree::Remove. r=njn
Before bug 1412722, which removed the sentinels, the code looked like:

  if (rbp_r_c->Right()->Left()->IsBlack()) {

At that point in the code, rbp_r_c is the root node of the tree. If
rbp_r_c->Right() was the sentinel, ->Right()->Left() would be the
sentinel too, and the sentinel is black. Which means the condition would
be true.

The code after was:

  if (rbp_r_c->Right() && (!rbp_r_c->Right()->Left() ||
                           rbp_r_c->Right()->Left()->IsBlack())) {

The second half correctly deals with the case of
rbp_r_c->Right()->Left() being the sentinel. But the first half now
makes things different: ->Right() being null would correspond to the
previous case where it was the sentinel, and the test would not return
true in that case when it would have before. When ->Right() is not null,
things are normal again.

The correct check is to make the branch taken when ->Right() is null.

Now, looking under which conditions we may get in that branch wrongly:
- The root node's right link must be empty, which means a very small tree.
- The comparison between the removed key and the root node must indicate
  the key is greater than the value of the root node.
- There's another case where the comparison result (rbp_r_cmp) can be
  eGreater, when it is reassigned under one of the branches under the
  eEqual test, and that branch is only taken when ->Right() on the root
  node was non-null, which was the non-broken case.

So it would seem we can't reach that code when rbp_r_c->Right() is null
anyways, so it /should/ practically make no difference. Better safe than
sorry, though. It's hard to tell anything from crash stats, because
since the templatization in bug 1403444, all crashes fit in one bucket,
when there used to be 5 functions before :(

While here, add a missing include in rb.h.

--HG--
extra : rebase_source : 2ebcb84345ad52059b0c081b9e2e1af1d0bbb7bc
2018-02-15 14:38:52 +09:00
Bobby Holley 3d8a3d0d6d Bug 1436541 - Don't clobber the thread-local arenas when we happen to hit a large allocation. r=glandium
MozReview-Commit-ID: 9i5B76vkNfr
2018-02-07 18:17:48 -08:00
Mike Hommey e483ecaab0 Bug 1424709 - Force disable the OSX system "nano allocator". r=spohl
We're not actually using it, and it messes up with the zone allocator in
mozjemalloc after fork(). See the lengthy analysis in
https://bugzilla.mozilla.org/show_bug.cgi?id=1424709#c34 and following.

--HG--
extra : rebase_source : c58e13b897dde7b32d83c43fbb2a04a0db3a5dc9
2018-01-31 17:18:01 +09:00
Sylvestre Ledru 5de63ef061 Bug 1394734 - Replace CONFIG['MSVC'] by CONFIG['CC_TYPE'] r=glandium
MozReview-Commit-ID: 5orfnoude7h

--HG--
extra : rebase_source : 1ed9a6b56e1d27221a07624767a7fb0e6147117f
2017-12-08 13:46:13 +01:00
Mike Hommey 3f1f5c9fc2 Bug 1423114 - Remove moz_xposix_memalign and moz_xvalloc. r=njn
They are both infallible wrappers of posix_memalign and valloc.
There is also moz_xmemalign, which wraps memalign, which is mostly
always available as of bug 1402647.

None of them are actually used, but it's still desirable to at least
have one infallible variant, so keep moz_xmemalign and remove the other
two.

While here, we actually make both memalign and moz_xmemalign always
available.

--HG--
extra : rebase_source : 1c3ca4b3e3310543145f3181dfa4e764be1d6ff8
2017-12-05 17:34:19 +09:00
Mike Hommey 3dc24c3f35 Bug 1423000 - Re-run clang-format on memory/build. r=njn
Most adjustements come from some recent .clang-format changes. A few
were overlooked from changes to the code.

--HG--
extra : rebase_source : c397762ea739fec05256a8c0132541bf4c727c32
2017-12-03 14:22:05 +09:00
Mike Hommey ab9b4f676d Bug 1423000 - Move mozjemalloc mutexes to a separate header. r=njn
Also change the definition of the StaticMutex constructor to unconfuse clang-format.

--HG--
extra : rebase_source : aa2aa07c8c324e1253c20b34d850230579d45632
2017-12-03 14:21:19 +09:00
Mike Hommey 2b2a874b47 Bug 1420355 - Statically link DMD. r=njn
--HG--
extra : rebase_source : 8e7cf975d096116b666532f3fe8aa5a7f61b5725
2017-11-28 08:10:48 +09:00
Mike Hommey 50d09b60b0 Bug 1420355 - Allow to statically link replace-malloc libraries. r=njn
And statically link logalloc.

Statically linking is the default, except when building with
--enable-project=memory, allowing to use the generated libraries from
such builds with Firefox.

--HG--
extra : rebase_source : efe9edce8db6a6264703e0105c2192edc5ca8415
2017-11-23 17:24:19 +09:00
Mike Hommey 449973411b Bug 1420355 - Don't declare replace_* functions in replace_malloc.h. r=njn
The original purpose of those declarations was to avoid the function
definitions being wrong, as well as forcing them being exported
properly (as extern "C", as weak symbols when necessary, etc.), but:

- The implementations being C++, function overloads simply allowed
  functions with the same name to have a different signature.

- As of bug 1420353, the functions don't need to be exported anymore,
  nor do we care whether their symbols are mangled. Furthermore, they're
  now being assigned to function table fields, meaning there is type
  checking in place, now.

So all in all, these declarations can be removed.

Also, as further down the line we're going to statically link the
replace-malloc libraries, avoid symbol conflicts by making those
functions static.

--HG--
extra : rebase_source : 0dbb15f2c85bc873e7eb662b8d757f99b0732270
2017-11-28 07:18:15 +09:00
Csoregi Natalia 4ce8d0124c Backed out 7 changesets (bug 1420355) for mass failures on OS X and Android. r=backout on a CLOSED TREE
Backed out changeset a7ed89e13a4c (bug 1420355)
Backed out changeset fd6702e6e0a0 (bug 1420355)
Backed out changeset 0479dda078a2 (bug 1420355)
Backed out changeset e69357ccca9e (bug 1420355)
Backed out changeset 3742a4b69ba2 (bug 1420355)
Backed out changeset 451cd087922f (bug 1420355)
Backed out changeset d80b5c4e1dd0 (bug 1420355)
2017-11-29 03:08:46 +02:00