Re: linux-next: build warning after merge of the block tree

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 2022-07-15 at 22:33 +1000, Stephen Rothwell wrote:
> Hi all,
> 
> On Fri, 15 Jul 2022 13:52:32 +1000 Stephen Rothwell
> <sfr@xxxxxxxxxxxxxxxx> wrote:
> > 
> > After merging the block tree, today's linux-next build (arm
> > multi_v7_defconfig) produced this warning:
> > 
> > In file included from include/linux/slab.h:16,
> >                  from io_uring/net.c:5:
> > io_uring/net.c: In function 'io_recvmsg_multishot_overflow':
> > include/linux/overflow.h:67:22: warning: comparison of distinct
> > pointer types lacks a cast
> >    67 |         (void) (&__a == &__b);                  \
> >       |                      ^~
> > io_uring/net.c:332:13: note: in expansion of macro
> > 'check_add_overflow'
> >   332 |         if (check_add_overflow(sizeof(struct
> > io_uring_recvmsg_out),
> >       |             ^~~~~~~~~~~~~~~~~~
> > include/linux/overflow.h:68:22: warning: comparison of distinct
> > pointer types lacks a cast
> >    68 |         (void) (&__a == __d);                   \
> >       |                      ^~
> > io_uring/net.c:332:13: note: in expansion of macro
> > 'check_add_overflow'
> >   332 |         if (check_add_overflow(sizeof(struct
> > io_uring_recvmsg_out),
> >       |             ^~~~~~~~~~~~~~~~~~
> > include/linux/overflow.h:67:22: warning: comparison of distinct
> > pointer types lacks a cast
> >    67 |         (void) (&__a == &__b);                  \
> >       |                      ^~
> > io_uring/net.c:335:13: note: in expansion of macro
> > 'check_add_overflow'
> >   335 |         if (check_add_overflow(hdr, iomsg->controllen,
> > &hdr))
> >       |             ^~~~~~~~~~~~~~~~~~
> > 
> > Introduced by commit
> > 
> >   a8b38c4ce724 ("io_uring: support multishot in recvmsg")
> 
> This became a build failure in my i386 defconfig build.  So I have
> applied the following (probably not correct) patch that makes it
> build.
> 

Thanks for this. Your patch will probably not cause much damage,
however I have suggested the below patch to io_uring since I noticed
that the current logic was a bit off

Dylan


From: Dylan Yudaken <dylany@xxxxxx>
Date: Fri, 15 Jul 2022 05:55:02 -0700
Subject: [PATCH for-next] io_uring: fix types in
io_recvmsg_multishot_overflow

io_recvmsg_multishot_overflow had incorrect types on non x64 system.
But also it had an unnecessary INT_MAX check, which could just be done
by changing the type of the accumulator to int (also simplifying the
casts).

Reported-by: Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>
Fixes: a8b38c4ce724 ("io_uring: support multishot in recvmsg")
Signed-off-by: Dylan Yudaken <dylany@xxxxxx>
---
 io_uring/net.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/io_uring/net.c b/io_uring/net.c
index 616d5f04cc74..6b7d5f33e642 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -327,14 +327,14 @@ int io_send(struct io_kiocb *req, unsigned int
issue_flags)
 
 static bool io_recvmsg_multishot_overflow(struct io_async_msghdr
*iomsg)
 {
-	unsigned long hdr;
+	int hdr;
 
-	if (check_add_overflow(sizeof(struct io_uring_recvmsg_out),
-			       (unsigned long)iomsg->namelen, &hdr))
+	if (iomsg->namelen < 0)
 		return true;
-	if (check_add_overflow(hdr, iomsg->controllen, &hdr))
+	if (check_add_overflow((int)sizeof(struct
io_uring_recvmsg_out),
+			       iomsg->namelen, &hdr))
 		return true;
-	if (hdr > INT_MAX)
+	if (check_add_overflow(hdr, (int)iomsg->controllen, &hdr))
 		return true;
 
 	return false;

base-commit: a8b38c4ce7240d869c820d457bcd51e452149510
-- 
2.30.2







[Index of Archives]     [Linux Kernel]     [Linux USB Development]     [Yosemite News]     [Linux SCSI]

  Powered by Linux