From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from [87.239.111.99] (localhost [127.0.0.1]) by dev.tarantool.org (Postfix) with ESMTP id 4F0A86687CD; Fri, 20 Oct 2023 16:18:15 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 dev.tarantool.org 4F0A86687CD DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=tarantool.org; s=dev; t=1697807895; bh=QJIB8l5Qj2nPoqScgqPVv62PaQXfcuIwkuXavWz0vK0=; h=To:Date:In-Reply-To:References:Subject:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=pI60IzfOW/eUoqkRqyNKcjnzx0Qg+29RIhFAJGSGov9ITfv7Z8l6X6dNwiGbik/xe eWqgZRsv7HviupyAQcVdcmzyca+s+uXgkB5KFMpfhBG8WYnSP4XNJ53xPQKCfVO64C TmspNpIZXCrNcfCbcSBT/RsVwTlZjugynrp+EfTY= Received: from mail-ej1-f52.google.com (mail-ej1-f52.google.com [209.85.218.52]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by dev.tarantool.org (Postfix) with ESMTPS id 4799D6687CD for ; Fri, 20 Oct 2023 16:18:14 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 dev.tarantool.org 4799D6687CD Received: by mail-ej1-f52.google.com with SMTP id a640c23a62f3a-9adca291f99so123489766b.2 for ; Fri, 20 Oct 2023 06:18:14 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1697807893; x=1698412693; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=74l6EFrn9o7LUPyFWe6unOswHnBSVMNNh/yXDelZgec=; b=ozm9Fb/ZKOS07yB5l2xsGdo+NbeiqKbkvYxrWz9Gk9ML43jbXxWXVjhyVl/ezB45+G UJ0s5UWbw/YBTGA4kUIkxCDkMhGxgp2l+jdpgrlc8M/xDrzPxZ5ZHjrLu1p/7RqqSDKE 54zUiEICMmn5y+2kr2gIhs+yAksT6rjPilkPE7u/KkVJtiHdwo69zjP/6xpfQLVgbr1w 7pPJWHRSI+N6KNbaJ7WJRophuWXYVHMcGYJsT3xTBUIQYg2zo0vQQqZ78OEu9wndrIi2 KQ0TJPWO/P8S9dW63yk/nDrqolzj6c3N56Lm5vEByD9kypGg6rQvduhSMJ6zSOnoYakq 91kA== X-Gm-Message-State: AOJu0YyzxLQbpyBnFX5FNpF7+rQU38XvxQJuTvrTxbS7od3WMUUjteEl jwWIVIHMyH4XQIICdI2rd5f0PGqYA8+Kyg== X-Google-Smtp-Source: AGHT+IEAqMuh1PVKdRmh+CP1iCF6bhXrMazjbmpDPRX4mo3PlLRBs4uLf+owKfb928x48SMi9Z7f5Q== X-Received: by 2002:a17:907:724e:b0:9b2:bdaf:25da with SMTP id ds14-20020a170907724e00b009b2bdaf25damr1224582ejc.17.1697807892943; Fri, 20 Oct 2023 06:18:12 -0700 (PDT) Received: from pony.. ([185.6.247.97]) by smtp.gmail.com with ESMTPSA id qh12-20020a170906ecac00b009b9af27d98csm1478598ejb.132.2023.10.20.06.18.11 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 20 Oct 2023 06:18:12 -0700 (PDT) To: tarantool-patches@dev.tarantool.org, Sergey Kaplun , max.kokryashkin@gmail.com, Igor Munkin Date: Fri, 20 Oct 2023 16:17:21 +0300 Message-Id: <821f9aeee78a0b91e895b4c4ff191dc26f5175bb.1697806970.git.sergeyb@tarantool.org> X-Mailer: git-send-email 2.34.1 In-Reply-To: References: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Subject: [Tarantool-patches] [PATCH luajit 1/2][v3] Fix embedded bytecode loader. X-BeenThere: tarantool-patches@dev.tarantool.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: Tarantool development patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , From: Sergey Bronnikov via Tarantool-patches Reply-To: Sergey Bronnikov Errors-To: tarantool-patches-bounces@dev.tarantool.org Sender: "Tarantool-patches" From: Mike Pall (cherry-picked from commit 820339960123dc78a7ce03edf53fcf4fdae0e55d) The original problem is specific to x32 and is as follows: when a chunk with a bytecode library is loaded into memory, and the address is higher than 0x80000100, the `LexState->pe`, that contains an address of the end of the bytecode chunk in the memory, will wrap around and become smaller than the address in `LexState->p`, that contains an address of the beginning of bytecode chunk in the memory. In `bcread_fill()` called by `bcread_want()`, `memcpy()` is called with a very large size and causes the bus error on x86 and the segmentation fault on ARM Android. The problem cannot be reproduced on platforms supported by Tarantool (ARM64, x86_64), so test doesn't reproduce a problem without a patch and tests the patch partially. Sergey Bronnikov: * added the description and the test Part of tarantool/tarantool#9145 --- src/lib_package.c | 4 +-- src/lj_bcread.c | 10 +++---- src/lj_lex.c | 6 +++++ src/lj_lex.h | 1 + test/tarantool-tests/CMakeLists.txt | 1 + .../lj-549-bytecode-loader.test.lua | 27 +++++++++++++++++++ .../lj-549-bytecode-loader/CMakeLists.txt | 20 ++++++++++++++ .../lj-549-bytecode-loader/script.lua | 3 +++ 8 files changed, 65 insertions(+), 7 deletions(-) create mode 100644 test/tarantool-tests/lj-549-bytecode-loader.test.lua create mode 100644 test/tarantool-tests/lj-549-bytecode-loader/CMakeLists.txt create mode 100644 test/tarantool-tests/lj-549-bytecode-loader/script.lua diff --git a/src/lib_package.c b/src/lib_package.c index b49f0209..12603038 100644 --- a/src/lib_package.c +++ b/src/lib_package.c @@ -260,7 +260,7 @@ static int ll_loadfunc(lua_State *L, const char *path, const char *name, int r) const char *bcdata = ll_bcsym(*reg, mksymname(L, name, SYMPREFIX_BC)); lua_pop(L, 1); if (bcdata) { - if (luaL_loadbuffer(L, bcdata, LJ_MAX_BUF, name) != 0) + if (luaL_loadbuffer(L, bcdata, ~(size_t)0, name) != 0) return PACKAGE_ERR_LOAD; return 0; } @@ -431,7 +431,7 @@ static int lj_cf_package_loader_preload(lua_State *L) if (lua_isnil(L, -1)) { /* Not found? */ const char *bcname = mksymname(L, name, SYMPREFIX_BC); const char *bcdata = ll_bcsym(NULL, bcname); - if (bcdata == NULL || luaL_loadbuffer(L, bcdata, LJ_MAX_BUF, name) != 0) + if (bcdata == NULL || luaL_loadbuffer(L, bcdata, ~(size_t)0, name) != 0) lua_pushfstring(L, "\n\tno field package.preload['%s']", name); } return 1; diff --git a/src/lj_bcread.c b/src/lj_bcread.c index cddf6ff1..48ec15e4 100644 --- a/src/lj_bcread.c +++ b/src/lj_bcread.c @@ -79,6 +79,7 @@ static LJ_NOINLINE void bcread_fill(LexState *ls, MSize len, int need) ls->c = -1; /* Only bad if we get called again. */ break; } + if (sz >= LJ_MAX_BUF - n) lj_err_mem(ls->L); if (n) { /* Append to buffer. */ n += (MSize)sz; p = lj_buf_need(&ls->sb, n < len ? len : n); @@ -90,20 +91,20 @@ static LJ_NOINLINE void bcread_fill(LexState *ls, MSize len, int need) ls->p = buf; ls->pe = buf + sz; } - } while (ls->p + len > ls->pe); + } while ((MSize)(ls->pe - ls->p) < len); } /* Need a certain number of bytes. */ static LJ_AINLINE void bcread_need(LexState *ls, MSize len) { - if (LJ_UNLIKELY(ls->p + len > ls->pe)) + if (LJ_UNLIKELY((MSize)(ls->pe - ls->p) < len)) bcread_fill(ls, len, 1); } /* Want to read up to a certain number of bytes, but may need less. */ static LJ_AINLINE void bcread_want(LexState *ls, MSize len) { - if (LJ_UNLIKELY(ls->p + len > ls->pe)) + if (LJ_UNLIKELY((MSize)(ls->pe - ls->p) < len)) bcread_fill(ls, len, 0); } @@ -463,8 +464,7 @@ GCproto *lj_bcread(LexState *ls) setprotoV(L, L->top, pt); incr_top(L); } - if ((int32_t)(2*(uint32_t)(ls->pe - ls->p)) > 0 || - L->top-1 != bcread_oldtop(L, ls)) + if ((ls->pe != ls->p && !ls->endmark) || L->top-1 != bcread_oldtop(L, ls)) bcread_error(ls, LJ_ERR_BCBAD); /* Pop off last prototype. */ L->top--; diff --git a/src/lj_lex.c b/src/lj_lex.c index cef3c683..6291705f 100644 --- a/src/lj_lex.c +++ b/src/lj_lex.c @@ -48,6 +48,11 @@ static LJ_NOINLINE LexChar lex_more(LexState *ls) size_t sz; const char *p = ls->rfunc(ls->L, ls->rdata, &sz); if (p == NULL || sz == 0) return LEX_EOF; + if (sz >= LJ_MAX_BUF) { + if (sz != ~(size_t)0) lj_err_mem(ls->L); + sz = ~(uintptr_t)0 - (uintptr_t)p; + ls->endmark = 1; + } ls->pe = p + sz; ls->p = p + 1; return (LexChar)(uint8_t)p[0]; @@ -408,6 +413,7 @@ int lj_lex_setup(lua_State *L, LexState *ls) ls->lookahead = TK_eof; /* No look-ahead token. */ ls->linenumber = 1; ls->lastline = 1; + ls->endmark = 0; lex_next(ls); /* Read-ahead first char. */ if (ls->c == 0xef && ls->p + 2 <= ls->pe && (uint8_t)ls->p[0] == 0xbb && (uint8_t)ls->p[1] == 0xbf) { /* Skip UTF-8 BOM (if buffered). */ diff --git a/src/lj_lex.h b/src/lj_lex.h index ae05a954..a26e504a 100644 --- a/src/lj_lex.h +++ b/src/lj_lex.h @@ -73,6 +73,7 @@ typedef struct LexState { BCInsLine *bcstack; /* Stack for bytecode instructions/line numbers. */ MSize sizebcstack; /* Size of bytecode stack. */ uint32_t level; /* Syntactical nesting level. */ + int endmark; /* Trust bytecode end marker, even if not at EOF. */ } LexState; LJ_FUNC int lj_lex_setup(lua_State *L, LexState *ls); diff --git a/test/tarantool-tests/CMakeLists.txt b/test/tarantool-tests/CMakeLists.txt index c15d6037..55226143 100644 --- a/test/tarantool-tests/CMakeLists.txt +++ b/test/tarantool-tests/CMakeLists.txt @@ -63,6 +63,7 @@ add_subdirectory(gh-5813-resolving-of-c-symbols/stripped) add_subdirectory(gh-6098-fix-side-exit-patching-on-arm64) add_subdirectory(gh-6189-cur_L) add_subdirectory(lj-416-xor-before-jcc) +add_subdirectory(lj-549-bytecode-loader) add_subdirectory(lj-601-fix-gc-finderrfunc) add_subdirectory(lj-727-lightuserdata-itern) add_subdirectory(lj-802-panic-at-mcode-protfail) diff --git a/test/tarantool-tests/lj-549-bytecode-loader.test.lua b/test/tarantool-tests/lj-549-bytecode-loader.test.lua new file mode 100644 index 00000000..490d94fc --- /dev/null +++ b/test/tarantool-tests/lj-549-bytecode-loader.test.lua @@ -0,0 +1,27 @@ +local tap = require('tap') +local test = tap.test('lj-549-bytecode-loader') + +test:plan(2) + +-- Test creates a shared library with LuaJIT bytecode, +-- loads shared library as a Lua module and checks, +-- that no crashes eliminated. +-- +-- Manual steps for reproducing are the following: +-- +-- $ make HOST_CC='gcc -m32' TARGET_CFLAGS='-m32' \ +-- TARGET_LDFLAGS='-m32' \ +-- TARGET_SHLDFLAGS='-m32' \ +-- -f Makefile.original +-- $ echo 'print("test")' > a.lua +-- $ LUA_PATH="src/?.lua;;" luajit -b a.lua a.c +-- $ gcc -m32 -fPIC -shared a.c -o a.so +-- $ luajit -e "require('a')" +-- Program received signal SIGBUS, Bus error + +local module_name = 'script' +local ok, module = pcall(require, module_name) +test:is(ok, true, 'bytecode loader works') +test:is(module.msg, 'Lango team', 'message is ok') + +test:done(true) diff --git a/test/tarantool-tests/lj-549-bytecode-loader/CMakeLists.txt b/test/tarantool-tests/lj-549-bytecode-loader/CMakeLists.txt new file mode 100644 index 00000000..6431c682 --- /dev/null +++ b/test/tarantool-tests/lj-549-bytecode-loader/CMakeLists.txt @@ -0,0 +1,20 @@ +set(LIB_NAME "script") +set(LUA_FILE ${CMAKE_CURRENT_SOURCE_DIR}/${LIB_NAME}.lua) +set(C_FILE ${LIB_NAME}.c) + +make_lua_path(LUA_PATH + PATHS + ${PROJECT_SOURCE_DIR}/src/?.lua + ${PROJECT_SOURCE_DIR}/src/jit/?.lua +) + +add_custom_target(export_bc + COMMAND ${CMAKE_COMMAND} -E env LUA_PATH=${LUA_PATH} ${LUAJIT_BINARY} -b ${LUA_FILE} ${C_FILE} + DEPENDS luajit-main ${LUA_FILE} + BYPRODUCTS ${C_FILE} + COMMENT "Exporting bytecode to a C file" + VERBATIM +) + +BuildTestCLib(${LIB_NAME} ${C_FILE}) +add_dependencies(${LIB_NAME} export_bc) diff --git a/test/tarantool-tests/lj-549-bytecode-loader/script.lua b/test/tarantool-tests/lj-549-bytecode-loader/script.lua new file mode 100644 index 00000000..e72162af --- /dev/null +++ b/test/tarantool-tests/lj-549-bytecode-loader/script.lua @@ -0,0 +1,3 @@ +return { + msg = 'Lango team', +} -- 2.34.1