Skip to contentSkip to conversation

binder: convert `BINDER_*` ioctl `#define`s into an `enum`

PATCH Superseded Miguel Ojeda 4 participants 5 messages May 16, 2022 – May 18, 2022 Rust

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Miguel Ojeda

author maintainer
#0 lore raw

PATCH 1/1 binder: convert `BINDER_*` ioctl `#define`s into an `enum`

bindgen (a tool which generates the "raw" C bindings for Rust) only works (so far) with "simple" C #defines. In order to avoid having to manually maintain these constants in the (potential) Rust side, this patch converts them into an enum.

There may be support in the future for expanding macros that end up in a "numeric" one: https://github.com/rust-lang/rust-bindgen/issues/753.

---
Two notes:
- Let me know if you prefer that I base this on top of a binder branch.
- Wedson should OK the changes for his Signed-off-by tag (and probably he should be the author of the patch).

 include/uapi/linux/android/binder.h | 28 +++++++++++++++-------------
 1 file changed, 15 insertions(+), 13 deletions(-)
include/uapi/linux/android/binder.h +15 -13
index 11157fae8a8e..ed4d11a0bb99 100644
@@ -236,19 +236,21 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
-#define BINDER_WRITE_READ		_IOWR('b', 1, struct binder_write_read)
-#define BINDER_SET_IDLE_TIMEOUT		_IOW('b', 3, __s64)
-#define BINDER_SET_MAX_THREADS		_IOW('b', 5, __u32)
-#define BINDER_SET_IDLE_PRIORITY	_IOW('b', 6, __s32)
-#define BINDER_SET_CONTEXT_MGR		_IOW('b', 7, __s32)
-#define BINDER_THREAD_EXIT		_IOW('b', 8, __s32)
-#define BINDER_VERSION			_IOWR('b', 9, struct binder_version)
-#define BINDER_GET_NODE_DEBUG_INFO	_IOWR('b', 11, struct binder_node_debug_info)
-#define BINDER_GET_NODE_INFO_FOR_REF	_IOWR('b', 12, struct binder_node_info_for_ref)
-#define BINDER_SET_CONTEXT_MGR_EXT	_IOW('b', 13, struct flat_binder_object)
-#define BINDER_FREEZE			_IOW('b', 14, struct binder_freeze_info)
-#define BINDER_GET_FROZEN_INFO		_IOWR('b', 15, struct binder_frozen_status_info)
-#define BINDER_ENABLE_ONEWAY_SPAM_DETECTION	_IOW('b', 16, __u32)
+enum {
+	BINDER_WRITE_READ			= _IOWR('b', 1, struct binder_write_read),
+	BINDER_SET_IDLE_TIMEOUT			= _IOW('b', 3, __s64),
+	BINDER_SET_MAX_THREADS			= _IOW('b', 5, __u32),
+	BINDER_SET_IDLE_PRIORITY		= _IOW('b', 6, __s32),
+	BINDER_SET_CONTEXT_MGR			= _IOW('b', 7, __s32),
+	BINDER_THREAD_EXIT			= _IOW('b', 8, __s32),
+	BINDER_VERSION				= _IOWR('b', 9, struct binder_version),
+	BINDER_GET_NODE_DEBUG_INFO		= _IOWR('b', 11, struct binder_node_debug_info),
+	BINDER_GET_NODE_INFO_FOR_REF		= _IOWR('b', 12, struct binder_node_info_for_ref),
+	BINDER_SET_CONTEXT_MGR_EXT		= _IOW('b', 13, struct flat_binder_object),
+	BINDER_FREEZE				= _IOW('b', 14, struct binder_freeze_info),
+	BINDER_GET_FROZEN_INFO			= _IOWR('b', 15, struct binder_frozen_status_info),
+	BINDER_ENABLE_ONEWAY_SPAM_DETECTION	= _IOW('b', 16, __u32),
+};
 
 /*
  * NOTE: Two special error codes you should check for when calling

base-commit: 42226c989789d8da4af1de0c31070c96726d990c

Signature · 1 line
2.36.1
#1 lore raw
Quoted 8 lines · Miguel Ojeda

bindgen (a tool which generates the "raw" C bindings for Rust) only works (so far) with "simple" C #defines. In order to avoid having to manually maintain these constants in the (potential) Rust side, this patch converts them into an enum.

There may be support in the future for expanding macros that end up in a "numeric" one: https://github.com/rust-lang/rust-bindgen/issues/753.

Quoted 6 lines

---
Two notes:
- Let me know if you prefer that I base this on top of a binder branch.

Yes, this does not apply to my char-misc.git tree in the char-misc-next branch on git.kernel.org as I think we have added some new binder ioctls recently.

Or you can make it against linux-next with the rust stuff removed, that also would work as well. But as-is, this patch does not work.

thanks,

greg k-h

#2 lore raw
Quoted 4 lines · Greg Kroah-Hartman

Yes, this does not apply to my char-misc.git tree in the char-misc-next branch on git.kernel.org as I think we have added some new binder ioctls recently.

Yeah, exactly, I thought you might want the other base to avoid the conflicts -- sending it in a bit!

Cheers,
Miguel

#3 lore raw
In reply to Miguel Ojeda
Quoted 12 lines · Miguel Ojeda
include/uapi/linux/android/binder.h +0 -11
-#define BINDER_WRITE_READ              _IOWR('b', 1, struct binder_write_read)
-#define BINDER_SET_IDLE_TIMEOUT                _IOW('b', 3, __s64)
-#define BINDER_SET_MAX_THREADS         _IOW('b', 5, __u32)
-#define BINDER_SET_IDLE_PRIORITY       _IOW('b', 6, __s32)
-#define BINDER_SET_CONTEXT_MGR         _IOW('b', 7, __s32)
-#define BINDER_THREAD_EXIT             _IOW('b', 8, __s32)
-#define BINDER_VERSION                 _IOWR('b', 9, struct binder_version)
-#define BINDER_GET_NODE_DEBUG_INFO     _IOWR('b', 11, struct binder_node_debug_info)
-#define BINDER_GET_NODE_INFO_FOR_REF   _IOWR('b', 12, struct binder_node_info_for_ref)
-#define BINDER_SET_CONTEXT_MGR_EXT     _IOW('b', 13, struct flat_binder_object)
-#define BINDER_FREEZE                  _IOW('b', 14, struct binder_freeze_info)

include/uapi/linux/android/binder.h struct binder_frozen_status_info { · line 246

Quoted 10 lines
include/uapi/linux/android/binder.h +8 -2
-#define BINDER_GET_FROZEN_INFO         _IOWR('b', 15, struct binder_frozen_status_info)
-#define BINDER_ENABLE_ONEWAY_SPAM_DETECTION    _IOW('b', 16, __u32)
+enum {
+       BINDER_WRITE_READ                       = _IOWR('b', 1, struct binder_write_read),
+       BINDER_SET_IDLE_TIMEOUT                 = _IOW('b', 3, __s64),
+       BINDER_SET_MAX_THREADS                  = _IOW('b', 5, __u32),
+       BINDER_SET_IDLE_PRIORITY                = _IOW('b', 6, __s32),
+       BINDER_SET_CONTEXT_MGR                  = _IOW('b', 7, __s32),
+       BINDER_THREAD_EXIT                      = _IOW('b', 8, __s32),
+       BINDER_VERSION                          = _IOWR('b', 9, struct binder_version),

I wonder if that breaks any tools that extract ioctl command number definitions from kernel headers. I see one other header using enum, but everything else has the #define. The main user of ioctl command definitions that comes to mind is 'strace', but I don't know where they get the numbers from.

It's probably not a big deal as long as it's limited to binder, but it may become
more of a problem if we do this for more subsystems over time.

Arnd

#4 lore raw
Quoted 9 lines · Arnd Bergmann

I wonder if that breaks any tools that extract ioctl command number definitions from kernel headers. I see one other header using enum, but everything else has the #define. The main user of ioctl command definitions that comes to mind is 'strace', but I don't know where they get the numbers from.

It's probably not a big deal as long as it's limited to binder, but it may become
more of a problem if we do this for more subsystems over time.

Good point. For public headers, we may just want to do this on the Rust side, after all. And it should be as automated as possible too -- I will take a look.

In any case, I also took at strace. They indeed parse the C files -- with regexes, not an actual C preprocessor/compiler, and for some files, including binder.h, they have custom code where they convert the enums into #defines and then call the common path on that. But it does that only for enums that match a regexp. So this particular change does not work as-is, though would work if we give a suitable name to the enum (I double-checked running the scripts with the changes applied, and it seemed to work).

But who knows what other projects are doing anyway, so it may be best to just hold on this change, since anyway Rust Binder is still an RFC -- thus if Linus/Greg decide to merge the Rust support, this change can still wait. And we may not need it anyway if we do it on the Rust side.

In conclusion, I would drop this patch for the moment, and I will fix it on our side.

(By the way, please note there was a v2:
https://lore.kernel.org/lkml/[email protected]/)

Cheers,
Miguel