linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Al Viro <viro@zeniv.linux.org.uk>
To: Miklos Szeredi <mszeredi@redhat.com>
Cc: linux-fsdevel@vger.kernel.org, David Howells <dhowells@redhat.com>
Subject: Re: [PATCH 09/12] fs_parser: "string" with missing value is a "flag"
Date: Tue, 17 Dec 2019 18:31:43 +0000	[thread overview]
Message-ID: <20191217183143.GC4203@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20191217173259.GA4203@ZenIV.linux.org.uk>

On Tue, Dec 17, 2019 at 05:32:59PM +0000, Al Viro wrote:

> No.  This is simply wrong - as it is, there's no difference between
> "foo" and "foo=".  Passing NULL in the latter case is wrong, but
> this is not a good fix.
> 
> This
>         if (v_size > 0) {
>                 param.string = kmemdup_nul(value, v_size, GFP_KERNEL);
>                 if (!param.string)
>                         return -ENOMEM;
>         }
> should really be
> 	if (value) {
>                 param.string = kmemdup_nul(value, v_size, GFP_KERNEL);
>                 if (!param.string)
>                         return -ENOMEM;
>         }
> and your chunk should be conditional upon value, not v_size.  The
> same problem exists for rbd.c

How about this (completely untested):

Pass consistent param->type to fs_parse()

As it is, vfs_parse_fs_string() makes "foo" and "foo=" indistinguishable;
both get fs_value_is_string for ->type and NULL for ->string.  To make
it even more unpleasant, that combination is impossible to produce with
fsconfig().

Much saner rules would be
	"foo"		=> fs_value_is_flag, NULL
	"foo="		=> fs_value_is_string, ""
	"foo=bar"	=> fs_value_is_string, "bar"
All cases are distinguishable, all results are expressable by fsconfig(),
->has_value checks are much simpler that way (to the point of the field
being useless) and quite a few regressions go away (gfs2 has no business
accepting -o nodebug=, for example).

Partially based upon patches from Miklos.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index 2b184563cd32..9fc686be81ca 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -6433,7 +6433,7 @@ static int rbd_parse_options(char *options, struct rbd_parse_opts_ctx *pctx)
 		if (*key) {
 			struct fs_parameter param = {
 				.key	= key,
-				.type	= fs_value_is_string,
+				.type	= fs_value_is_flag,
 			};
 			char *value = strchr(key, '=');
 			size_t v_len = 0;
@@ -6443,14 +6443,11 @@ static int rbd_parse_options(char *options, struct rbd_parse_opts_ctx *pctx)
 					continue;
 				*value++ = 0;
 				v_len = strlen(value);
-			}
-
-
-			if (v_len > 0) {
 				param.string = kmemdup_nul(value, v_len,
 							   GFP_KERNEL);
 				if (!param.string)
 					return -ENOMEM;
+				param.type = fs_value_is_string;
 			}
 			param.size = v_len;
 
diff --git a/fs/fs_context.c b/fs/fs_context.c
index 138b5b4d621d..9097421cbba5 100644
--- a/fs/fs_context.c
+++ b/fs/fs_context.c
@@ -175,14 +175,15 @@ int vfs_parse_fs_string(struct fs_context *fc, const char *key,
 
 	struct fs_parameter param = {
 		.key	= key,
-		.type	= fs_value_is_string,
+		.type	= fs_value_is_flag,
 		.size	= v_size,
 	};
 
-	if (v_size > 0) {
+	if (value) {
 		param.string = kmemdup_nul(value, v_size, GFP_KERNEL);
 		if (!param.string)
 			return -ENOMEM;
+		param.type = fs_value_is_string;
 	}
 
 	ret = vfs_parse_fs_param(fc, &param);
diff --git a/fs/fs_parser.c b/fs/fs_parser.c
index d1930adce68d..65842cd2e320 100644
--- a/fs/fs_parser.c
+++ b/fs/fs_parser.c
@@ -85,7 +85,6 @@ int fs_parse(struct fs_context *fc,
 	const struct fs_parameter_enum *e;
 	int ret = -ENOPARAM, b;
 
-	result->has_value = !!param->string;
 	result->negated = false;
 	result->uint_64 = 0;
 
@@ -95,7 +94,7 @@ int fs_parse(struct fs_context *fc,
 		 * "xxx" takes the "no"-form negative - but only if there
 		 * wasn't an value.
 		 */
-		if (result->has_value)
+		if (param->type != fs_value_is_flag)
 			goto unknown_parameter;
 		if (param->key[0] != 'n' || param->key[1] != 'o' || !param->key[2])
 			goto unknown_parameter;
@@ -127,14 +126,13 @@ int fs_parse(struct fs_context *fc,
 	case fs_param_is_u64:
 	case fs_param_is_enum:
 	case fs_param_is_string:
-		if (param->type != fs_value_is_string)
-			goto bad_value;
-		if (!result->has_value) {
+		if (param->type == fs_value_is_string)
+			break;
+		if (param->type == fs_value_is_flag) {
 			if (p->flags & fs_param_v_optional)
 				goto okay;
-			goto bad_value;
 		}
-		/* Fall through */
+		goto bad_value;
 	default:
 		break;
 	}
@@ -144,8 +142,7 @@ int fs_parse(struct fs_context *fc,
 	 */
 	switch (p->type) {
 	case fs_param_is_flag:
-		if (param->type != fs_value_is_flag &&
-		    (param->type != fs_value_is_string || result->has_value))
+		if (param->type != fs_value_is_flag)
 			return invalf(fc, "%s: Unexpected value for '%s'",
 				      desc->name, param->key);
 		result->boolean = true;
@@ -206,9 +203,6 @@ int fs_parse(struct fs_context *fc,
 	case fs_param_is_fd: {
 		switch (param->type) {
 		case fs_value_is_string:
-			if (!result->has_value)
-				goto bad_value;
-
 			ret = kstrtouint(param->string, 0, &result->uint_32);
 			break;
 		case fs_value_is_file:
diff --git a/include/linux/fs_parser.h b/include/linux/fs_parser.h
index dee140db6240..45323203128b 100644
--- a/include/linux/fs_parser.h
+++ b/include/linux/fs_parser.h
@@ -72,7 +72,6 @@ struct fs_parameter_description {
  */
 struct fs_parse_result {
 	bool			negated;	/* T if param was "noxxx" */
-	bool			has_value;	/* T if value supplied to param */
 	union {
 		bool		boolean;	/* For spec_bool */
 		int		int_32;		/* For spec_s32/spec_enum */

  reply	other threads:[~2019-12-17 18:31 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-28 15:59 [PATCH 00/12] various vfs patches Miklos Szeredi
2019-11-28 15:59 ` [PATCH 01/12] aio: fix async fsync creds Miklos Szeredi
2019-12-13  9:32   ` Miklos Szeredi
2020-05-04  8:05   ` Avi Kivity
2019-11-28 15:59 ` [PATCH 02/12] fs_parse: fix fs_param_v_optional handling Miklos Szeredi
2019-11-29 11:31   ` Andrew Price
2019-11-29 14:43     ` Miklos Szeredi
2019-11-29 15:56       ` Andrew Price
2019-12-16 23:28   ` Al Viro
2019-12-17  1:18     ` Al Viro
2019-12-17  3:27       ` Al Viro
2019-11-28 15:59 ` [PATCH 03/12] vfs: verify param type in vfs_parse_sb_flag() Miklos Szeredi
2019-11-28 15:59 ` [PATCH 04/12] uapi: deprecate STATX_ALL Miklos Szeredi
2019-11-28 15:59 ` [PATCH 05/12] statx: don't clear STATX_ATIME on SB_RDONLY Miklos Szeredi
2019-11-28 15:59 ` [PATCH 06/12] utimensat: AT_EMPTY_PATH support Miklos Szeredi
2019-11-28 15:59 ` [PATCH 07/12] f*xattr: allow O_PATH descriptors Miklos Szeredi
2019-11-28 15:59 ` [PATCH 08/12] vfs: allow unprivileged whiteout creation Miklos Szeredi
2019-12-17  3:51   ` Al Viro
2019-12-17  4:22     ` Miklos Szeredi
2019-11-28 15:59 ` [PATCH 09/12] fs_parser: "string" with missing value is a "flag" Miklos Szeredi
2019-12-17 17:32   ` Al Viro
2019-12-17 18:31     ` Al Viro [this message]
2019-11-28 15:59 ` [PATCH 10/12] vfs: don't parse forbidden flags Miklos Szeredi
2019-11-28 15:59 ` [PATCH 11/12] vfs: don't parse "posixacl" option Miklos Szeredi
2019-12-17  3:42   ` Al Viro
2019-12-17  4:18     ` Miklos Szeredi
2019-12-17  4:28       ` Al Viro
2019-11-28 15:59 ` [PATCH 12/12] vfs: don't parse "silent" option Miklos Szeredi
2019-12-17  3:37   ` Al Viro
2019-12-17  4:12     ` Miklos Szeredi
2019-12-17  4:16       ` Miklos Szeredi
2019-12-17  4:19         ` Al Viro
2019-12-17  4:23           ` Miklos Szeredi
2019-12-17  4:28             ` Miklos Szeredi
2019-12-17  4:17       ` Al Viro
2019-12-13  9:33 ` [PATCH 00/12] various vfs patches Miklos Szeredi
2019-12-16 23:13   ` Al Viro

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=20191217183143.GC4203@ZenIV.linux.org.uk \
    --to=viro@zeniv.linux.org.uk \
    --cc=dhowells@redhat.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=mszeredi@redhat.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).