[<prev] [next>] [day] [month] [year] [list]
Message-ID: <20240630213626.GA23566@altlinux.org>
Date: Sun, 30 Jun 2024 21:36:26 +0000
From: "Dmitry V. Levin" <ldv@...ace.io>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: Vincent Donnefort <vdonnefort@...gle.com>,
David Hildenbrand <david@...hat.com>, linux-api@...r.kernel.org,
linux-trace-kernel@...r.kernel.org, linux-mm@...ck.org,
linux-kernel@...r.kernel.org
Subject: [PATCH] uapi: change TRACE_MMAP_IOCTL_GET_READER to avoid collision
with TCGETS
The number that was initially chosen for TRACE_MMAP_IOCTL_GET_READER,
unfortunately, collides with TCGETS on most of architectures.
For example, this is how strace output would look like when
support for this value of TRACE_MMAP_IOCTL_GET_READER is added:
$ strace -e ioctl stty
ioctl(0, TCGETS or TRACE_MMAP_IOCTL_GET_READER, {c_iflag=ICRNL|IXON, c_oflag=NL0|CR0|TAB0|BS0|VT0|FF0|OPOST|ONLCR,
+c_cflag=B38400|CS8|CREAD, c_lflag=ISIG|ICANON|ECHO|ECHOE|ECHOK|IEXTEN|ECHOCTL|ECHOKE, ...}) = 0
Even though ioctl numbers are inherently not unique, TCGETS
is a very traditional one, so let's change the value of
TRACE_MMAP_IOCTL_GET_READER a bit to avoid this collision.
Given that _IO('T', 0x1) is _IOC(_IOC_NONE, 'T', 0x1, 0),
something like _IOC(_IOC_NONE, 'T', 0x1, 0x1) should be OK.
Fixes: cf9f0f7c4c5bb ("tracing: Allow user-space mapping of the ring-buffer")
Signed-off-by: Dmitry V. Levin <ldv@...ace.io>
---
include/uapi/linux/trace_mmap.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/uapi/linux/trace_mmap.h b/include/uapi/linux/trace_mmap.h
index bd1066754220..cb858f1b8367 100644
--- a/include/uapi/linux/trace_mmap.h
+++ b/include/uapi/linux/trace_mmap.h
@@ -43,6 +43,6 @@ struct trace_buffer_meta {
__u64 Reserved2;
};
-#define TRACE_MMAP_IOCTL_GET_READER _IO('T', 0x1)
+#define TRACE_MMAP_IOCTL_GET_READER _IOC(_IOC_NONE, 'T', 0x1, 0x1)
#endif /* _TRACE_MMAP_H_ */
--
ldv
Powered by blists - more mailing lists