All of lore.kernel.org
 help / color / mirror / Atom feed
From: david.laight.linux@gmail.com
To: "Willy Tarreau" <w@1wt.eu>,
	"Thomas Weißschuh" <linux@weissschuh.net>,
	linux-kernel@vger.kernel.org, "Cheng Li" <lechain@gmail.com>
Cc: David Laight <david.laight.linux@gmail.com>
Subject: [PATCH next 10/12] tools/nolibc/printf: Use bit-pattern for integral formats
Date: Tue,  3 Feb 2026 10:29:58 +0000	[thread overview]
Message-ID: <20260203103000.20206-11-david.laight.linux@gmail.com> (raw)
In-Reply-To: <20260203103000.20206-1-david.laight.linux@gmail.com>

From: David Laight <david.laight.linux@gmail.com>

Generates better code when code is common for multiple formats.
In particular it stops two calls to u64toa_r().

Signed-off-by: David Laight <david.laight.linux@gmail.com>
---
 tools/include/nolibc/stdio.h | 33 +++++++++++++++++++--------------
 1 file changed, 19 insertions(+), 14 deletions(-)

diff --git a/tools/include/nolibc/stdio.h b/tools/include/nolibc/stdio.h
index bad528921590..23fe8b8e7767 100644
--- a/tools/include/nolibc/stdio.h
+++ b/tools/include/nolibc/stdio.h
@@ -241,7 +241,7 @@ char *fgets(char *s, int size, FILE *stream)
 
 
 /* printf(). Supports most of the normal integer and string formats.
- *  - %[#0-+ ][width|*[.precision|*]][{l,t,z,ll,L,j,q}]{d,i,u,c,x,X,p,s,m}
+ *  - %[#0-+ ][width|*[.precision|*]][{l,t,z,ll,L,j,q}]{d,i,u,c,x,X,p,s,m,%}
  *  - %% generates a single %
  *  - %m outputs strerror(errno).
  *  - # only affects %x and prepends 0x to non-zero values.
@@ -258,7 +258,7 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 {
 	char c;
 	int len, written, width, precision;
-	unsigned int flags;
+	unsigned int flags, c_flag;
 	char tmpbuf[32 + 24];
 	const char *outstr;
 
@@ -326,8 +326,12 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 			if (!((c >= 'a' && c <= 'z') || (c == 'X' && (c = 'x'))))
 				goto bad_conversion_specifier;
 
-			/* Numeric and pointer conversion specifiers */
-			if (__PF_FLAG(c) & (__PF_FLAG('c') | __PF_FLAG('d') | __PF_FLAG('i') | __PF_FLAG('u') |
+			/* Numeric and pointer conversion specifiers.
+			 * We need to check for "%p" or "%#x" later, merging here gives better code.
+			 * But '#' collides with 'c' so shift right.
+			 */
+			c_flag = __PF_FLAG(c) | (flags & __PF_FLAG('#')) >> 1;
+			if (c_flag & (__PF_FLAG('c') | __PF_FLAG('d') | __PF_FLAG('i') | __PF_FLAG('u') |
 					    __PF_FLAG('x') | __PF_FLAG('p') | __PF_FLAG('s'))) {
 				unsigned long long v;
 				long long signed_v;
@@ -335,7 +339,7 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 				int sign = 0;
 
 				/* Annoying 'p' === '0' so mask from flags */
-				if ((__PF_FLAG(c) | (flags & ~__PF_FLAG('p'))) &
+				if ((c_flag | (flags & ~__PF_FLAG('p'))) &
 				    (__PF_FLAG('p') | __PF_FLAG('s') | __PF_FLAG('l') | __PF_FLAG('t') | __PF_FLAG('z'))) {
 					v = va_arg(args, unsigned long);
 					signed_v = (long)v;
@@ -347,13 +351,14 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 					signed_v = (int)v;
 				}
 
-				switch (c) {
-				case 'c':
+				if (c_flag & __PF_FLAG('c')) {
 					tmpbuf[0] = v;
 					len = 1;
 					outstr = tmpbuf;
 					goto do_output;
-				case 's':
+				}
+
+				if (c_flag & __PF_FLAG('s')) {
 					if (!v) {
 						outstr = "(null)";
 						/* Match glibc, nothing output if precision too small */
@@ -364,8 +369,9 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 do_strnlen_output:
 					len = strnlen(outstr, precision);
 					goto do_output;
-				case 'd':
-				case 'i':
+				}
+
+				if (c_flag & (__PF_FLAG('d') | __PF_FLAG('i'))) {
 					if (signed_v < 0) {
 						sign = '-';
 						v = -(signed_v + 1);
@@ -375,12 +381,11 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 					} else if (flags & __PF_FLAG(' ')) {
 						sign = ' ';
 					}
-					c = 'u';
 				}
 
 				if (v == 0) {
 					/* There are special rules for zero. */
-					if (c == 'p') {
+					if (c_flag & __PF_FLAG('p')) {
 						/* match glibc, precision is ignored */
 						outstr = "(nil)";
 						len = 5;
@@ -395,11 +400,11 @@ int __nolibc_printf(__nolibc_printf_cb cb, void *state, const char *fmt, va_list
 					*out = '0';
 					len = 1;
 				} else {
-					if (c == 'u') {
+					if (c_flag & (__PF_FLAG('d') | __PF_FLAG('i') | __PF_FLAG('u'))) {
 						len = u64toa_r(v, out);
 					} else {
 						len = u64toh_r(v, out);
-						if (c == 'p' || (flags & __PF_FLAG('#')))
+						if (c_flag & (__PF_FLAG('p') | __PF_FLAG('#' - 1)))
 							sign = 'x' | '0' << 8;
 					}
 				}
-- 
2.39.5


  parent reply	other threads:[~2026-02-03 10:30 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-03 10:29 [PATCH next 00/12] tools/nolibc: Enhance printf() david.laight.linux
2026-02-03 10:29 ` [PATCH next 01/12] tools/nolibc/printf: Move length check to snprintf callback david.laight.linux
2026-02-03 10:29 ` [PATCH next 02/12] tools/nolibc/printf: Add buffering to vfprintf() callback david.laight.linux
2026-02-03 10:29 ` [PATCH next 03/12] tools/nolibc/printf: output pad spaces in 16 byte chunks david.laight.linux
2026-02-03 10:29 ` [PATCH next 04/12] selftests/nolibc: Improve reporting of vfprintf() errors david.laight.linux
2026-02-03 10:29 ` [PATCH next 05/12] tools/nolibc/printf: Simplify __nolibc_printf() david.laight.linux
2026-02-03 10:29 ` [PATCH next 06/12] tools/nolibc/printf: Add support for left alignment and %[tzLq]d" david.laight.linux
2026-02-04  4:14   ` Willy Tarreau
2026-02-04 10:17     ` David Laight
2026-02-04 10:40       ` Willy Tarreau
2026-02-04 15:39         ` David Laight
2026-02-03 10:29 ` [PATCH next 07/12] tools/nolibc/printf: Prepend the sign after a numeric conversion david.laight.linux
2026-02-03 10:29 ` [PATCH next 08/12] tools/nolibc/printf: use bit-match to detect valid conversion characters david.laight.linux
2026-02-03 10:29 ` [PATCH next 09/12] tools/nolibc/printf: support precision and zero padding david.laight.linux
2026-02-03 10:29 ` david.laight.linux [this message]
2026-02-03 10:29 ` [PATCH 11/12] selftests/nolibc: Increase coverage of printf format tests david.laight.linux
2026-02-03 18:22   ` kernel test robot
2026-02-03 22:05     ` David Laight
2026-02-03 23:23   ` kernel test robot
2026-02-03 10:30 ` [PATCH 12/12] selftests/nolibc: Use printf("%.*s", n, "") to align output david.laight.linux
2026-02-03 17:39 ` [PATCH next 00/12] tools/nolibc: Enhance printf() David Laight

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=20260203103000.20206-11-david.laight.linux@gmail.com \
    --to=david.laight.linux@gmail.com \
    --cc=lechain@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@weissschuh.net \
    --cc=w@1wt.eu \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.