Hi, Sergey,

thanks for the patch! See comments below.


On 16.01.2025 16:35, Sergey Kaplun wrote:


<snipped>

> diff --git a/test/tarantool-tests/lj-1116-redzones-checks.test.lua b/test/tarantool-tests/lj-1116-redzones-checks.test.lua
> new file mode 100644
> index 00000000..70062ec9
> --- /dev/null
> +++ b/test/tarantool-tests/lj-1116-redzones-checks.test.lua
> @@ -0,0 +1,118 @@
> +local tap = require('tap')
> +-- Test file to demonstrate mcode area overflow during recording a
> +-- trace with the high FPR pressure.
> +-- See also,https://github.com/LuaJIT/LuaJIT/issues/1116.
> +--
> +-- XXX: Test fails only with GC64 enabled before the commit.
I would rephrase: XXX: Test fails with reverted fix and enabled GC64.
> +local test = tap.test('lj-1116-redzones-checks'):skipcond({
> +  ['Test requires JIT enabled'] = not jit.status(),
> +})
> +
> +test:plan(1)
> +
> +jit.opt.start('hotloop=1')
> +
> +-- XXX: This test snippet was originally created by the fuzzer.
> +-- Seehttps://oss-fuzz.com/testcase-detail/5622965122170880.
> +--
> +-- Unfortunately, it's impossible to reduce the testcase further.
> +-- Before the patch, assembling some instructions (like `IR_CONV
> +-- int.num`, for example) with many mcode to be emitted may
> +-- overflow the `MCLIM_REDZONE` (64) at once due to the huge
> +-- mcode emitting.
> +-- For example `IR_CONV` in this test requires 66 bytes of the
> +-- machine code:
> +-- |  cvttsd2si r15d, xmm5
> +-- |  xorps xmm9, xmm9
> +-- |  cvtsi2sd xmm9, r15d
> +-- |  ucomisd xmm5, xmm9
> +-- |  jnz 0x11edb00e5       ->37
> +-- |  jpe 0x11edb00e5       ->37
> +-- |  mov [rsp+0x80], r15d
> +-- |  mov r15, [rsp+0xe8]
> +-- |  movsd xmm9, [rsp+0xe0]
> +-- |  movsd xmm5, [rsp+0xd8]
> +--
> +-- The reproducer needs sufficient register pressure as to
> +-- immediately spill the result of the instruction to the stack
> +-- and then reload the three registers used by the instruction,
> +-- and to have chosen enough registers with numbers >=8 (because
> +-- shaving off a REX prefix [1] or two would get 66 back down
> +-- to <= `MCLIM_REDZONE`), and to be using lots of spill slots
> +-- (because memory offsets <= 0x7f are shorter to encode compared
> +-- to those >= 0x80. So, each reload instruction consumes 9 bytes.
> +-- This makes this reproducer unstable (regarding the register
> +-- allocator changes). So, lets use this as a regression test.
> +--
> +-- [1]:https://wiki.osdev.org/X86-64_Instruction_Encoding#REX_prefix
> +
> +_G.a = 0
> +_G.b = 0
> +_G.c = 0
> +_G.d = 0
> +_G.e = 0
> +_G.f = 0
> +_G.g = 0
> +_G.h = 0
> +-- Skip `i`.

I didn't get it.


<snipped>