From: "Philippe Mathieu-Daudé" <philmd@linaro.org>
To: "Markus Armbruster" <armbru@redhat.com>,
"Daniel P. Berrangé" <berrange@redhat.com>
Cc: qemu-devel@nongnu.org,
"Pavel Dovgalyuk" <pavel.dovgaluk@ispras.ru>,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
"Michael Roth" <michael.roth@amd.com>,
"Stefan Berger" <stefanb@linux.vnet.ibm.com>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Marc-André Lureau" <marcandre.lureau@redhat.com>,
"Juan Quintela" <quintela@redhat.com>,
"Gerd Hoffmann" <kraxel@redhat.com>
Subject: Re: [PATCH v3 2/3] qapi: Do not generate empty enum
Date: Tue, 21 Mar 2023 15:48:12 +0100 [thread overview]
Message-ID: <d312145c-8af2-c8ca-ec45-4f6f060b3b07@linaro.org> (raw)
In-Reply-To: <873564spze.fsf@pond.sub.org>
On 16/3/23 15:57, Markus Armbruster wrote:
> Daniel P. Berrangé <berrange@redhat.com> writes:
> But all of this falls apart with conditional members!
>
> Example 1 (taken from qapi/block-core.json):
>
> { 'enum': 'BlockdevAioOptions',
> 'data': [ 'threads', 'native',
> { 'name': 'io_uring', 'if': 'CONFIG_LINUX_IO_URING' } ] }
>
> Generates now:
>
> typedef enum BlockdevAioOptions {
> BLOCKDEV_AIO_OPTIONS_THREADS,
> BLOCKDEV_AIO_OPTIONS_NATIVE,
> #if defined(CONFIG_LINUX_IO_URING)
> BLOCKDEV_AIO_OPTIONS_IO_URING,
> #endif /* defined(CONFIG_LINUX_IO_URING) */
> BLOCKDEV_AIO_OPTIONS__MAX,
> } BlockdevAioOptions;
>
> BLOCKDEV_AIO_OPTIONS__MAX is 3 if defined(CONFIG_LINUX_IO_URING), else
> 2.
>
> After the next patch:
>
> typedef enum BlockdevAioOptions {
> BLOCKDEV_AIO_OPTIONS_THREADS,
> BLOCKDEV_AIO_OPTIONS_NATIVE,
> #if defined(CONFIG_LINUX_IO_URING)
> BLOCKDEV_AIO_OPTIONS_IO_URING,
> #endif /* defined(CONFIG_LINUX_IO_URING) */
> #define BLOCKDEV_AIO_OPTIONS__MAX 3
> } BlockdevAioOptions;
>
> Now it's always 3.
>
> Example 2 (same with members reordered):
>
> { 'enum': 'BlockdevAioOptions',
> 'data': [ { 'name': 'io_uring', 'if': 'CONFIG_LINUX_IO_URING' },
> 'threads', 'native' ] }
>
> Same problem for __MAX, additional problem for __DUMMY:
>
> typedef enum BlockdevAioOptions {
> BLOCKDEV_AIO_OPTIONS__DUMMY = 0,
> #if defined(CONFIG_LINUX_IO_URING)
> BLOCKDEV_AIO_OPTIONS_IO_URING = 0,
> #endif /* defined(CONFIG_LINUX_IO_URING) */
> BLOCKDEV_AIO_OPTIONS_THREADS,
> BLOCKDEV_AIO_OPTIONS_NATIVE,
> #define BLOCKDEV_AIO_OPTIONS__MAX 3
> } BlockdevAioOptions;
>
> If CONFIG_LINUX_IO_URING is off, the enum starts at 1 instead of 0.
>
> Arrays indexed by the enum start with a hole. Code using them is
> probably not prepared for holes.
Can we meet half-way only generating the MAX definitions for
unconditional enums, keeping the conditional ones as is?
-- >8 --
diff --git a/scripts/qapi/types.py b/scripts/qapi/types.py
@@ -88,16 +88,18 @@ def gen_enum(name: str,
members: List[QAPISchemaEnumMember],
prefix: Optional[str] = None) -> str:
assert members
- # append automatically generated _MAX value
- enum_members = members + [QAPISchemaEnumMember('_MAX', None)]
-
ret = mcgen('''
typedef enum %(c_name)s {
''',
c_name=c_name(name))
- for memb in enum_members:
+ has_cond = any(memb.ifcond.is_present() for memb in members)
+ if has_cond:
+ # append automatically generated _MAX value
+ members += [QAPISchemaEnumMember('_MAX', None)]
+
+ for memb in members:
ret += memb.ifcond.gen_if()
ret += mcgen('''
%(c_enum)s,
@@ -105,6 +107,13 @@ def gen_enum(name: str,
c_enum=c_enum_const(name, memb.name, prefix))
ret += memb.ifcond.gen_endif()
+ if not has_cond:
+ ret += mcgen('''
+#define %(c_name)s %(c_length)s
+''',
+ c_name=c_enum_const(name, '_MAX', prefix),
+ c_length=len(members))
+
ret += mcgen('''
} %(c_name)s;
''',
---
next prev parent reply other threads:[~2023-03-21 14:48 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-15 11:28 [PATCH v3 0/3] qapi: Simplify enum generation Philippe Mathieu-Daudé
2023-03-15 11:28 ` [PATCH v3 1/3] scripts/git.orderfile: Display QAPI script changes before schema ones Philippe Mathieu-Daudé
2023-03-15 15:14 ` Philippe Mathieu-Daudé
2023-03-15 20:30 ` Juan Quintela
2023-03-15 11:28 ` [PATCH v3 2/3] qapi: Do not generate empty enum Philippe Mathieu-Daudé
2023-03-15 15:02 ` Richard Henderson
2023-03-16 12:31 ` Markus Armbruster
2023-03-16 13:54 ` Daniel P. Berrangé
2023-03-16 14:39 ` Juan Quintela
2023-03-16 14:42 ` Daniel P. Berrangé
2023-03-16 14:57 ` Markus Armbruster
2023-03-21 14:31 ` Philippe Mathieu-Daudé
2023-03-21 15:19 ` Daniel P. Berrangé
2023-03-21 21:43 ` Eric Blake
2023-03-22 5:45 ` Markus Armbruster
2023-03-21 14:48 ` Philippe Mathieu-Daudé [this message]
2023-03-21 19:00 ` Markus Armbruster
2023-03-15 11:28 ` [PATCH v3 3/3] qapi: Generate enum count as definition Philippe Mathieu-Daudé
2023-03-15 19:58 ` Dr. David Alan Gilbert
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=d312145c-8af2-c8ca-ec45-4f6f060b3b07@linaro.org \
--to=philmd@linaro.org \
--cc=armbru@redhat.com \
--cc=berrange@redhat.com \
--cc=dgilbert@redhat.com \
--cc=kraxel@redhat.com \
--cc=marcandre.lureau@redhat.com \
--cc=michael.roth@amd.com \
--cc=pavel.dovgaluk@ispras.ru \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
--cc=stefanb@linux.vnet.ibm.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).