Linux-Sparse Archive on lore.kernel.org
 help / color / Atom feed
From: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
To: linux-sparse@vger.kernel.org
Cc: Ben Dooks <ben.dooks@codethink.co.uk>,
	Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Subject: [PATCH 02/13] format-check: more complete parsing of the length & type modifiers
Date: Wed, 14 Oct 2020 01:22:20 +0200
Message-ID: <20201013232231.10349-3-luc.vanoostenryck@gmail.com> (raw)
In-Reply-To: <20201013232231.10349-1-luc.vanoostenryck@gmail.com>

Reorganize the parsing the length modifiers to follow more closely
the specs and handle some missing cases like 'hh' or '%[ieEaA]'.
Also, treats 'L' and 'll' synonymously as done by GCC.

Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
---
 verify-format.c | 265 +++++++++++++++++++++++++++---------------------
 1 file changed, 151 insertions(+), 114 deletions(-)

diff --git a/verify-format.c b/verify-format.c
index 4b4730285237..ae5bb2e6e985 100644
--- a/verify-format.c
+++ b/verify-format.c
@@ -44,6 +44,14 @@
 #include "expression.h"
 #include "verify-format.h"
 
+enum length_mod {
+	LEN_none,
+	LEN_hh, LEN_h,
+	LEN_l, LEN_ll,
+	LEN_L,
+	LEN_j, LEN_z, LEN_t,
+};
+
 struct format_type {
 	const char	*format;
 	int		(*test)(struct format_type *fmt,
@@ -113,140 +121,169 @@ static struct format_type printf_fmt_ptr_ref = {
 	.test = printf_fmt_pointer,
 };
 
-static int is_float_spec(char t)
-{
-	return t == 'f' || t == 'g' || t == 'F' || t == 'G';
-}
-
 static struct format_type *parse_printf_get_fmt(struct format_type *type,
 						const char *msg, const char **msgout)
 {
 	const char *ptr = msg;
-	int szmod=0;
+	int szmod = LEN_none;
 
 	type->test = NULL;
 	*msgout = ptr;
 
-	if (*ptr == 's') {
-		ptr++;
-		type->test = printf_fmt_string;
-		type->type = &const_string_ctype;
-	} else if (*ptr == 'c') {
-		ptr++;
-		type->test = printf_fmt_numtype;
-		type->type = &char_ctype;
-	} else if (*ptr == 'p') {
-		ptr++;
-		type->test = printf_fmt_print_pointer;
-		/* check for pointer being printed as hex explicitly */
-		if (*ptr == 'x' || *ptr == 'X') {
+	switch (*ptr++) {
+	case 'h':
+		szmod = LEN_h;
+		if (*ptr == 'h') {
+			szmod = LEN_hh;
 			ptr++;
-		} else if (isalpha(*ptr)) {
-			/* probably some extra specifiers after %p */
+		}
+		break;
+	case 'l':
+		szmod = LEN_l;
+		if (*ptr == 'l') {
+			szmod = LEN_ll;
 			ptr++;
-			type->test = printf_fmt_pointer;
 		}
-	} else if (*ptr == 'z') {
-		// todo - we should construct pointer to int/etc //
+		break;
+	case 'q':
+		szmod = LEN_ll;
+		break;
+	case 'L':
+		szmod = LEN_L;
+		break;
+	case 'j':
+		szmod = LEN_j;
+		break;
+	case 'z':
+		szmod = LEN_z;
+		break;
+	case 't':
+		szmod = LEN_t;
+		break;
+	default:
+		ptr--;
+		break;
+	}
 
-		ptr++;
-		if (*ptr == 'd' || *ptr == 'i') {
-			ptr++;
-			type->test = printf_fmt_numtype;
+	switch (*ptr++) {
+	case 'd': case 'i':
+		type->test = printf_fmt_numtype;
+		switch (szmod) {
+		case LEN_hh:
+		case LEN_h:
+		case LEN_none:
+			type->type = &int_ctype;
+			break;
+		case LEN_l:
+			type->type = &long_ctype;
+			break;
+		case LEN_L:
+		case LEN_ll:
+			type->type = &llong_ctype;
+			break;
+		case LEN_j:
+			type->type = intmax_ctype;
+			break;
+		case LEN_z:
 			type->type = ssize_t_ctype;
-		} else if (*ptr == 'u' || *ptr == 'x' || *ptr == 'X' ||
-			   *ptr == 'o') {
-			ptr++;
-			type->test = printf_fmt_numtype;
-			type->type = size_t_ctype;
+			break;
+		case LEN_t:
+			type->type = ptrdiff_ctype;
+			break;
+		default:
+			type->test = NULL;
 		}
-	} else {
-		if (*ptr == 'l') {
-			szmod++;
-			ptr++;
-			if (*ptr == 'l') {
-				szmod++;
-				ptr++;
-			}
-		} else if (*ptr == 'L') {
-			szmod++;
-			ptr++;
-		} else {
-			if (*ptr == 'h') { // short/char to int
-				szmod = -1;
-				ptr++;
-				if (*ptr == 'h')  // promotion from char
-					ptr++;
-			}
-			if (*ptr == 't') {  // ptrdiff_t
-				szmod = 2;
-				ptr++;
-			}
-			if (*ptr == 'j') { // intmax_t
-				// todo - replace iwth intmax_ctype when added
-				szmod = 1;
-				ptr++;
-			}
+		break;
+	case 'u': case 'o': case 'x': case 'X':
+		type->test = printf_fmt_numtype;
+		switch (szmod) {
+		case LEN_hh:
+		case LEN_h:
+		case LEN_none:
+			type->type = &uint_ctype;
+			break;
+		case LEN_l:
+			type->type = &ulong_ctype;
+			break;
+		case LEN_L:
+		case LEN_ll:
+			type->type = &ullong_ctype;
+			break;
+		case LEN_j:
+			type->type = uintmax_ctype;
+			break;
+		case LEN_z:
+			type->type = size_t_ctype;
+			break;
+		case LEN_t:
+			type->type = ptrdiff_ctype;
+			break;
+		default:
+			type->test = NULL;
 		}
-
-		if (*ptr == 'x' || *ptr == 'X' || *ptr == 'u' || *ptr == 'o') {
-			ptr++;
-			type->test = printf_fmt_numtype;
-			switch (szmod) {
-			case -1:
-				type->type = &ushort_ctype;
-				break;
-			case 0:
-				type->type = &uint_ctype;
-				break;
-			case 1:
-				type->type = &ulong_ctype;
-				break;
-			case 2:
-				type->type = &ullong_ctype;
-				break;
-			default:
-				type->test = NULL;
-			}
-		} else if (*ptr == 'i' || *ptr == 'd') {
-			ptr++;
-			type->test = printf_fmt_numtype;
-			switch (szmod) {
-			case -1:
-				type->type = &short_ctype;
-				break;
-			case 0:
-				type->type = &int_ctype;
-				break;
-			case 1:
-				type->type = &long_ctype;
-				break;
-			case 2:
-				type->type = &llong_ctype;
-				break;
-			default:
-				type->test = NULL;
-			}
-		} else if (*ptr == 'L' && is_float_spec(ptr[1])) {
-			type->test = printf_fmt_numtype;
+		break;
+	case 'e': case 'E': case 'f': case 'F': case 'g': case 'G':
+	case 'a': case 'A':
+		type->test = printf_fmt_numtype;
+		switch (szmod) {
+		case LEN_none:
+			type->type = &double_ctype;
+			break;
+		case LEN_L:
 			type->type = &ldouble_ctype;
-			ptr += 2;
-		} else if (is_float_spec(*ptr)) {
-			type->test = printf_fmt_numtype;
-			type->type = szmod == 1 ? &ldouble_ctype :  &double_ctype;
-			ptr++;
-		} else if (*ptr == 's') {
-			type->test = printf_fmt_string;
+			break;
+		default:
+			break;
+		}
+		break;
+	case 'c':
+		type->test = printf_fmt_numtype;
+		switch (szmod) {
+		case LEN_none:
+			type->type = &int_ctype;
+			break;
+		case LEN_L:
+		case LEN_l:
+			type->type = wint_ctype;
+			break;
+		default:
+			break;
+		}
+		break;
+	case 's':
+		type->test = printf_fmt_string;
+		switch (szmod) {
+		case LEN_none:
+			type->type = &const_string_ctype;
+			break;
+		case LEN_L:
+		case LEN_l:
 			type->type = &const_wstring_ctype;
+			break;
+		default:
+			break;
+		}
+		break;
+	case 'p':
+		type->test = printf_fmt_print_pointer;
+		/* check for pointer being printed as hex explicitly */
+		if (*ptr == 'x' || *ptr == 'X') {
 			ptr++;
-		} else if (*ptr == 'n') {	/* pointer to an de-referenced int/etc */
-			// todo - we should construct pointer to int/etc //
-			// also should not have any flags or widths for this
-			type->test = printf_fmt_pointer;
+		} else if (isalpha(*ptr)) {
+			/* probably some extra specifiers after %p */
 			ptr++;
-		} else {
-			// anything else here?
+			type->test = printf_fmt_pointer;
 		}
+		break;
+	case 'n':
+		/* pointer to an de-referenced int/etc */
+		// todo - we should construct pointer to int/etc //
+		// also should not have any flags or widths for this
+		type->test = printf_fmt_pointer;
+		break;
+	default:
+		// anything else here?
+		break;
 	}
 
 	if (type->test == NULL)
-- 
2.28.0


  parent reply index

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-13 23:22 [PATCH 00/13] format-check: add specific type checking Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 01/13] format-check: void * is not OK for strings, fix the test Luc Van Oostenryck
2020-10-13 23:22 ` Luc Van Oostenryck [this message]
2020-10-13 23:22 ` [PATCH 03/13] format-check: add helper type_class() Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 04/13] format-check: merge 'fmt_string' & 'string' Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 05/13] format-check: remove unneeded member: target Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 06/13] format-check: add a function to check to type of strings Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 07/13] format-check: add a function to check to type of 'n' arguments Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 08/13] format-check: add a function to check to type of pointers Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 09/13] format-check: remove printf_fmt_print_pointer() Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 10/13] format-check: add a function to check the type of floats Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 11/13] format-check: add a function to check the type of integers Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 12/13] format-check: remove wrappers around type checking methods Luc Van Oostenryck
2020-10-13 23:22 ` [PATCH 13/13] format-check: simplify calling of parse_printf_get_fmt() Luc Van Oostenryck

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=20201013232231.10349-3-luc.vanoostenryck@gmail.com \
    --to=luc.vanoostenryck@gmail.com \
    --cc=ben.dooks@codethink.co.uk \
    --cc=linux-sparse@vger.kernel.org \
    /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

Linux-Sparse Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-sparse/0 linux-sparse/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-sparse linux-sparse/ https://lore.kernel.org/linux-sparse \
		linux-sparse@vger.kernel.org
	public-inbox-index linux-sparse

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-sparse


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git