lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20211109123621.52474-1-arnd@kernel.org>
Date:   Tue,  9 Nov 2021 13:36:08 +0100
From:   Arnd Bergmann <arnd@...nel.org>
To:     Jan Kara <jack@...e.com>
Cc:     Arnd Bergmann <arnd@...db.de>, linux-kernel@...r.kernel.org
Subject: [PATCH] udf: shut up pointer cast warning

From: Arnd Bergmann <arnd@...db.de>

On 32-bit architectures, the workaround of storing the directory position
in the private_data pointer causes a warning, as loff_t does not fit in
a pointer:

fs/udf/dir.c: In function 'udf_readdir':
fs/udf/dir.c:78:25: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
   78 |         if (ctx->pos != (uintptr_t)file->private_data) {
      |                         ^
fs/udf/dir.c:211:30: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
  211 |         file->private_data = (void *)(uintptr_t)ctx->pos;
      |                              ^

An extra cast to uintptr_t shuts up the warning. This is of course
still broken if the position is ever beyond the first 2^32 bytes (4GB).

I have not found a clear information on whether directories this
large are allowed on UDF, but it seems unlikely.

Fixes: 39a464de961f ("udf: Fix crash after seekdir")
Signed-off-by: Arnd Bergmann <arnd@...db.de>
---
 fs/udf/dir.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/fs/udf/dir.c b/fs/udf/dir.c
index 8ae501d27eff..5aaa82be420a 100644
--- a/fs/udf/dir.c
+++ b/fs/udf/dir.c
@@ -75,7 +75,7 @@ static int udf_readdir(struct file *file, struct dir_context *ctx)
 	 * identifying beginning of dir entry (names are under user control),
 	 * we need to scan the directory from the beginning.
 	 */
-	if (ctx->pos != (loff_t)file->private_data) {
+	if (ctx->pos != (uintptr_t)file->private_data) {
 		emit_pos = nf_pos;
 		nf_pos = 0;
 	}
@@ -208,7 +208,7 @@ static int udf_readdir(struct file *file, struct dir_context *ctx)
 
 out:
 	/* Store position where we've ended */
-	file->private_data = (void *)ctx->pos;
+	file->private_data = (void *)(uintptr_t)ctx->pos;
 	if (fibh.sbh != fibh.ebh)
 		brelse(fibh.ebh);
 	brelse(fibh.sbh);
-- 
2.29.2

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ