From: Sergey Kaplun via Tarantool-patches <tarantool-patches@dev.tarantool.org> To: Maxim Kokryashkin <m.kokryashkin@tarantool.org>, Sergey Bronnikov <sergeyb@tarantool.org> Cc: tarantool-patches@dev.tarantool.org Subject: [Tarantool-patches] [PATCH luajit 1/2] Fix bit op coercion in DUALNUM builds. Date: Wed, 2 Oct 2024 11:09:05 +0300 [thread overview] Message-ID: <7efd063772135322abfcc8de1214d20ebea69bad.1727855711.git.skaplun@tarantool.org> (raw) In-Reply-To: <cover.1727855711.git.skaplun@tarantool.org> From: Mike Pall <mike> Thanks to Sergey Kaplun. (cherry picked from commit f5fd22203eadf57ccbaa4a298010d23974b22fc0) The `lj_carith_check64()` function coerces the given number value to the 32-bit wide value. In this case, the 64-bit-wide operands will lose upper bits. This patch removes the excess coercion for the DUALNUM mode. Sergey Kaplun: * added the description and the test for the problem Part of tarantool/tarantool#10199 --- src/lj_carith.c | 4 +--- .../lj-1273-dualnum-bit-coercion.test.lua | 19 +++++++++++++++++++ 2 files changed, 20 insertions(+), 3 deletions(-) create mode 100644 test/tarantool-tests/lj-1273-dualnum-bit-coercion.test.lua diff --git a/src/lj_carith.c b/src/lj_carith.c index 90b3220f..eb56d552 100644 --- a/src/lj_carith.c +++ b/src/lj_carith.c @@ -349,9 +349,7 @@ uint64_t lj_carith_check64(lua_State *L, int narg, CTypeID *id) if (LJ_LIKELY(tvisint(o))) { return (uint32_t)intV(o); } else { - int32_t i = lj_num2bit(numV(o)); - if (LJ_DUALNUM) setintV(o, i); - return (uint32_t)i; + return (uint32_t)lj_num2bit(numV(o)); } } diff --git a/test/tarantool-tests/lj-1273-dualnum-bit-coercion.test.lua b/test/tarantool-tests/lj-1273-dualnum-bit-coercion.test.lua new file mode 100644 index 00000000..e83dbbcd --- /dev/null +++ b/test/tarantool-tests/lj-1273-dualnum-bit-coercion.test.lua @@ -0,0 +1,19 @@ +local tap = require('tap') + +-- Test file to demonstrate LuaJIT misbehaviour on operating +-- for 64-bit operands in built-in `bit` library in DUALNUM mode. +-- See also, https://github.com/LuaJIT/LuaJIT/issues/1273. + +local test = tap.test('lj-1273-dualnum-bit-coercion') +test:plan(1) + +local bit = require('bit') + +-- The cdata value for 2 ^ 33. +local EXPECTED = 2 ^ 33 + 0LL +-- Same value is used as mask for `bit.band()`. +local MASK = EXPECTED + +test:is(bit.band(2 ^ 33, MASK), EXPECTED, 'correct bit.band result') + +test:done(true) -- 2.46.2
next prev parent reply other threads:[~2024-10-02 8:10 UTC|newest] Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top 2024-10-02 8:09 [Tarantool-patches] [PATCH luajit 0/2] Fixes for 64 bit operands of the bit library Sergey Kaplun via Tarantool-patches 2024-10-02 8:09 ` Sergey Kaplun via Tarantool-patches [this message] 2024-10-08 10:12 ` [Tarantool-patches] [PATCH luajit 1/2] Fix bit op coercion in DUALNUM builds Sergey Bronnikov via Tarantool-patches 2024-10-11 19:08 ` Maxim Kokryashkin via Tarantool-patches 2024-10-02 8:09 ` [Tarantool-patches] [PATCH luajit 2/2] FFI: Fix 64 bit shift fold rules Sergey Kaplun via Tarantool-patches 2024-10-08 12:07 ` Sergey Bronnikov via Tarantool-patches 2024-10-08 14:24 ` Sergey Kaplun via Tarantool-patches 2024-10-09 14:29 ` Sergey Bronnikov via Tarantool-patches 2024-10-11 19:12 ` Maxim Kokryashkin via Tarantool-patches 2024-10-18 15:17 ` [Tarantool-patches] [PATCH luajit 0/2] Fixes for 64 bit operands of the bit library Sergey Kaplun via Tarantool-patches
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=7efd063772135322abfcc8de1214d20ebea69bad.1727855711.git.skaplun@tarantool.org \ --to=tarantool-patches@dev.tarantool.org \ --cc=m.kokryashkin@tarantool.org \ --cc=sergeyb@tarantool.org \ --cc=skaplun@tarantool.org \ --subject='Re: [Tarantool-patches] [PATCH luajit 1/2] Fix bit op coercion in DUALNUM builds.' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox