From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 290DDC433EF for ; Fri, 8 Oct 2021 18:56:35 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 05BA960FE8 for ; Fri, 8 Oct 2021 18:56:35 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S239678AbhJHS63 (ORCPT ); Fri, 8 Oct 2021 14:58:29 -0400 Received: from mail.kernel.org ([198.145.29.99]:43706 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231308AbhJHS6Z (ORCPT ); Fri, 8 Oct 2021 14:58:25 -0400 Received: by mail.kernel.org (Postfix) with ESMTPSA id 5B18360FE8; Fri, 8 Oct 2021 18:56:29 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1633719389; bh=hoAj05YuCRBBNrVjKODoIaxA5Jw6r/qy0A5fr49luFQ=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=FTedW6HxC0PQ40Ac5yGvKfkbP3O0QVCmM7/uLPjf3kz7pPvtu3Tn0JzbD5y8bhzyL OpVn2YKBHxIAK/43QAmCq3xdT85kXHHV+G3xQjBsux4Rsj3CJ+wNFXlaurQTHLp5KI 5BvpJqU0pwmisE0QhnjIRE1ClWiwQVLmsvJ0AFMRJvFzQRsCmuh7TstTC7u7xZ+mMX fHAYZsS7MsWiKifDBjBnMGBwXegLSl/g+FiIX8s5S1gd8fELcKv6j30aXbRaBfYvwl JLI6J6/pzmBssxupn6ZHssyJPL2UPt9i8c8S2UPRAWY12SjX0w6WPbit4I+8f2Fcfp JbmYaR8uTTNJA== Received: by quaco.ghostprotocols.net (Postfix, from userid 1000) id D609B410A1; Fri, 8 Oct 2021 15:56:26 -0300 (-03) Date: Fri, 8 Oct 2021 15:56:26 -0300 From: Arnaldo Carvalho de Melo To: Jiri Olsa Cc: James Clark , john.garry@huawei.com, ak@linux.intel.com, linux-perf-users@vger.kernel.org, Nick.Forrington@arm.com, Andrew.Kilroy@arm.com, Will Deacon , Mathieu Poirier , Leo Yan , Mark Rutland , Alexander Shishkin , Namhyung Kim , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 2/3] perf tools: Make the JSON parser more conformant when in strict mode Message-ID: References: <20211007110543.564963-1-james.clark@arm.com> <20211007110543.564963-3-james.clark@arm.com> <2e14963b-cb98-f508-7067-255fdbd36bdb@arm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-Url: http://acmel.wordpress.com Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Em Fri, Oct 08, 2021 at 03:12:03PM +0200, Jiri Olsa escreveu: > On Fri, Oct 08, 2021 at 11:08:25AM +0100, James Clark wrote: > > > > > > On 07/10/2021 18:52, Jiri Olsa wrote: > > > On Thu, Oct 07, 2021 at 12:05:41PM +0100, James Clark wrote: > > >> Return an error when a trailing comma is found or a new item is > > >> encountered before a comma or an opening brace. This ensures that the > > >> perf json files conform more closely to the spec at https://www.json.org > > >> > > >> Signed-off-by: James Clark > > >> --- > > >> tools/perf/pmu-events/jsmn.c | 42 ++++++++++++++++++++++++++++++++++-- > > >> 1 file changed, 40 insertions(+), 2 deletions(-) > > >> > > >> diff --git a/tools/perf/pmu-events/jsmn.c b/tools/perf/pmu-events/jsmn.c > > >> index 11d1fa18bfa5..8124d2d3ff0c 100644 > > >> --- a/tools/perf/pmu-events/jsmn.c > > >> +++ b/tools/perf/pmu-events/jsmn.c > > >> @@ -176,6 +176,14 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> jsmnerr_t r; > > >> int i; > > >> jsmntok_t *token; > > >> +#ifdef JSMN_STRICT > > > > > > I might have missed some discussion on this, but do we need the > > > JSMN_STRICT define, if you enable it in the next patch? > > > why can't we be more strict by default.. do you plan to disable > > > it in future? > > > > I didn't plan on disabling it, I was just trying to keep to the existing style of the > > jsmn project. > > > > I could have added the trailing comma detection by default and not inside any > > #ifdef JSMN_STRICT blocks, but I would like to enable JSMN_STRICT anyway, because it > > enables some additional built in checking that was already there. So I thought it > > made sense to put my new strict stuff inside the existing strict option. > > > > One option would be to remove all (including the existing) #ifdef JSMN_STRICT blocks > > and have everything strict by default. But it would be a further deviation from jsmn. > > ok, I think it makes sense to have JSMN_STRICT then.. > thanks for explanation > > Acked-by: Jiri Olsa So, is this for the whole patchset? b4 picked it just for this message. - Arnaldo > jirka > > > > > Thanks > > James > > > > > > > > thanks, > > > jirka > > > > > >> + /* > > >> + * Keeps track of whether a new object/list/primitive is expected. New items are only > > >> + * allowed after an opening brace, comma or colon. A closing brace after a comma is not > > >> + * valid JSON. > > >> + */ > > >> + int expecting_item = 1; > > >> +#endif > > >> > > >> for (; parser->pos < len; parser->pos++) { > > >> char c; > > >> @@ -185,6 +193,10 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> switch (c) { > > >> case '{': > > >> case '[': > > >> +#ifdef JSMN_STRICT > > >> + if (!expecting_item) > > >> + return JSMN_ERROR_INVAL; > > >> +#endif > > >> token = jsmn_alloc_token(parser, tokens, num_tokens); > > >> if (token == NULL) > > >> return JSMN_ERROR_NOMEM; > > >> @@ -196,6 +208,10 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> break; > > >> case '}': > > >> case ']': > > >> +#ifdef JSMN_STRICT > > >> + if (expecting_item) > > >> + return JSMN_ERROR_INVAL; > > >> +#endif > > >> type = (c == '}' ? JSMN_OBJECT : JSMN_ARRAY); > > >> for (i = parser->toknext - 1; i >= 0; i--) { > > >> token = &tokens[i]; > > >> @@ -219,6 +235,11 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> } > > >> break; > > >> case '\"': > > >> +#ifdef JSMN_STRICT > > >> + if (!expecting_item) > > >> + return JSMN_ERROR_INVAL; > > >> + expecting_item = 0; > > >> +#endif > > >> r = jsmn_parse_string(parser, js, len, tokens, > > >> num_tokens); > > >> if (r < 0) > > >> @@ -229,11 +250,15 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> case '\t': > > >> case '\r': > > >> case '\n': > > >> - case ':': > > >> - case ',': > > >> case ' ': > > >> break; > > >> #ifdef JSMN_STRICT > > >> + case ':': > > >> + case ',': > > >> + if (expecting_item) > > >> + return JSMN_ERROR_INVAL; > > >> + expecting_item = 1; > > >> + break; > > >> /* > > >> * In strict mode primitives are: > > >> * numbers and booleans. > > >> @@ -253,6 +278,9 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> case 'f': > > >> case 'n': > > >> #else > > >> + case ':': > > >> + case ',': > > >> + break; > > >> /* > > >> * In non-strict mode every unquoted value > > >> * is a primitive. > > >> @@ -260,6 +288,12 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> /*FALL THROUGH */ > > >> default: > > >> #endif > > >> + > > >> +#ifdef JSMN_STRICT > > >> + if (!expecting_item) > > >> + return JSMN_ERROR_INVAL; > > >> + expecting_item = 0; > > >> +#endif > > >> r = jsmn_parse_primitive(parser, js, len, tokens, > > >> num_tokens); > > >> if (r < 0) > > >> @@ -282,7 +316,11 @@ jsmnerr_t jsmn_parse(jsmn_parser *parser, const char *js, size_t len, > > >> return JSMN_ERROR_PART; > > >> } > > >> > > >> +#ifdef JSMN_STRICT > > >> + return expecting_item ? JSMN_ERROR_INVAL : JSMN_SUCCESS; > > >> +#else > > >> return JSMN_SUCCESS; > > >> +#endif > > >> } > > >> > > >> /* > > >> -- > > >> 2.28.0 > > >> > > > > > -- - Arnaldo