[Tarantool-patches] [PATCH luajit] Throw any errors before stack changes in trace stitching.

[Sergey Bronnikov]

Max, thanks for fixes! LGTM

On 2/20/24 11:08, Sergey Bronnikov via Tarantool-patches wrote:
> Hello, Max,
>
> thanks for the patch! see my comments
>
> Sergey
>
> On 2/16/24 14:11, Maxim Kokryashkin wrote:
>> From: Mike Pall <mike>
>>
>> Thanks to doujiang24.
>>
>> (cherry-picked from commit 3f9389edc6cdf3f78a6896d550c236860aed62b2)
>>
>> The Lua stack is changed in the `lj_record_stop`, so if the trace
>> is aborted, for example, when the `maxsnap` is reached, and an
>> error is thrown from there, the execution continues with an
>> unbalanced stack.
>>
>> Since the only error that is thrown from the `lj_record_stop`
>> with modified stack is `LJ_TRERR_SNAPOV`, this patch adds a
>
> As I see lj_record_stop [1] throws only LJ_TRERR_RETRY or
>
> I got you wrong.
>
>
> 1. 
> https://github.com/LuaJIT/LuaJIT/blob/0d313b243194a0b8d2399d8b549ca5a0ff234db5/src/lj_record.c#L299
>
>> check for snapshot overflow to the `recff_stitch`.
>>
>> Maxim Kokryashkin:
>> * added the description and the test for the problem
>>
>> Part of tarantool/tarantool#9595
>> ---
>> Branch: 
>> https://github.com/tarantool/luajit/tree/fckxorg/lj-pr-720-errors-before-stitch
>> PR: https://github.com/tarantool/tarantool/pull/9703
>> Issues:
>> https://github.com/tarantool/tarantool/issues/9595
>> https://github.com/LuaJIT/LuaJIT/pull/720
>>
>>   src/lj_ffrecord.c                             |  4 ++++
>>   .../lj-pr-720-errors-before-stitch.test.lua   | 20 +++++++++++++++++++
>>   2 files changed, 24 insertions(+)
>>   create mode 100644 
>> test/tarantool-tests/lj-pr-720-errors-before-stitch.test.lua
>>
>> diff --git a/src/lj_ffrecord.c b/src/lj_ffrecord.c
>> index 1b2ddb72..e3ed80fb 100644
>> --- a/src/lj_ffrecord.c
>> +++ b/src/lj_ffrecord.c
>> @@ -107,6 +107,10 @@ static void recff_stitch(jit_State *J)
>>     const BCIns *pc = frame_pc(base-1);
>>     TValue *pframe = frame_prevl(base-1);
>>
>> +  /* Check for this now. Throwing in lj_record_stop messes up the 
>> stack. */
>> +  if (J->cur.nsnap >= (MSize)J->param[JIT_P_maxsnap])
>> +    lj_trace_err(J, LJ_TRERR_SNAPOV);
>> +
>>     /* Move func + args up in Lua stack and insert continuation. */
>>     memmove(&base[1], &base[-1-LJ_FR2], sizeof(TValue)*nslot);
>>     setframe_ftsz(nframe, ((char *)nframe - (char *)pframe) + 
>> FRAME_CONT);
>> diff --git 
>> a/test/tarantool-tests/lj-pr-720-errors-before-stitch.test.lua 
>> b/test/tarantool-tests/lj-pr-720-errors-before-stitch.test.lua
>> new file mode 100644
>> index 00000000..53b7b5ea
>> --- /dev/null
>> +++ b/test/tarantool-tests/lj-pr-720-errors-before-stitch.test.lua
>
> You could omit "-pr" in the name, anyway
>
> https://github.com/LuaJIT/LuaJIT/issues/720 redirects to 
> https://github.com/LuaJIT/LuaJIT/pull/720.
>
>> @@ -0,0 +1,20 @@
>> +local tap = require('tap')
>> +local test = tap.test('lj-pr-720-errors-before-stitch'):skipcond({
>> +  ['Test requires JIT enabled'] = not jit.status(),
>> +})
>> +test:plan(1)
>> +
>> +-- `math.modf` recording is NYI.
>> +local modf = require('math').modf
>> +jit.opt.start('hotloop=1', 'maxsnap=1')
>> +
>> +-- The loop has only two iterations: the first to detect its
>> +-- hotness and the second to record it. The snapshot limit is
>> +-- set to one and is certainly reached.
>> +for _ = 1, 2 do
>> +  -- Forcify stitch.
>> +  modf(1.2)
>> +end
>> +
>> +test:ok(true, 'stack is balanced')
>> +test:done(true)
>> -- 
>> 2.43.0
>>