From: "Alex Bennée" <alex.bennee@linaro.org>
To: Markus Armbruster <armbru@redhat.com>
Cc: Alistair Francis <alistair.francis@wdc.com>, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH v2 01/29] include: Make headers more self-contained
Date: Wed, 07 Aug 2019 22:28:54 +0100 [thread overview]
Message-ID: <878ss4is15.fsf@linaro.org> (raw)
In-Reply-To: <87woforc8d.fsf@dusky.pond.sub.org>
Markus Armbruster <armbru@redhat.com> writes:
> Alex Bennée <alex.bennee@linaro.org> writes:
>
>> Markus Armbruster <armbru@redhat.com> writes:
>>
>>> Back in 2016, we discussed[1] rules for headers, and these were
>>> generally liked:
>>>
>>> 1. Have a carefully curated header that's included everywhere first. We
>>> got that already thanks to Peter: osdep.h.
>>>
>>> 2. Headers should normally include everything they need beyond osdep.h.
>>> If exceptions are needed for some reason, they must be documented in
>>> the header. If all that's needed from a header is typedefs, put
>>> those into qemu/typedefs.h instead of including the header.
>>>
>>> 3. Cyclic inclusion is forbidden.
>>>
>>> This patch gets include/ closer to obeying 2.
>>>
>>> It's actually extracted from my "[RFC] Baby steps towards saner
>>> headers" series[2], which demonstrates a possible path towards
>>> checking 2 automatically. It passes the RFC test there.
>>>
>>> [1] Message-ID: <87h9g8j57d.fsf@blackfin.pond.sub.org>
>>> https://lists.nongnu.org/archive/html/qemu-devel/2016-03/msg03345.html
>>> [2] Message-Id: <20190711122827.18970-1-armbru@redhat.com>
>>> https://lists.nongnu.org/archive/html/qemu-devel/2019-07/msg02715.html
>>>
>>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>>> Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
>>> ---
>> <snip>
>>> include/exec/cputlb.h | 2 ++
>>> include/exec/exec-all.h | 1 +
>>> include/exec/ioport.h | 2 ++
>>> include/exec/memory-internal.h | 2 ++
>>> include/exec/ram_addr.h | 1 +
>>> include/exec/softmmu-semi.h | 2 ++
>>> include/exec/tb-hash.h | 2 ++
>>> include/exec/user/thunk.h | 1 +
>>> include/fpu/softfloat-macros.h | 2 ++
>> <snip>
>>>
>>> /*
>>> * bdrv_write_threshold_set:
>>> diff --git a/include/disas/disas.h b/include/disas/disas.h
>>> index 15da511f49..ba47e9197c 100644
>>> --- a/include/disas/disas.h
>>> +++ b/include/disas/disas.h
>>> @@ -1,6 +1,7 @@
>>> #ifndef QEMU_DISAS_H
>>> #define QEMU_DISAS_H
>>>
>>> +#include "exec/hwaddr.h"
>>>
>>> #ifdef NEED_CPU_H
>>> #include "cpu.h"
>>> diff --git a/include/exec/cputlb.h b/include/exec/cputlb.h
>>> index 5373188be3..23abd71579 100644
>>> --- a/include/exec/cputlb.h
>>> +++ b/include/exec/cputlb.h
>>> @@ -19,6 +19,8 @@
>>> #ifndef CPUTLB_H
>>> #define CPUTLB_H
>>>
>>> +#include "exec/cpu-common.h"
>>> +
>>> #if !defined(CONFIG_USER_ONLY)
>>> /* cputlb.c */
>>> void tlb_protect_code(ram_addr_t ram_addr);
>>> diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h
>>> index 16034ee651..135aeaab0d 100644
>>> --- a/include/exec/exec-all.h
>>> +++ b/include/exec/exec-all.h
>>> @@ -20,6 +20,7 @@
>>> #ifndef EXEC_ALL_H
>>> #define EXEC_ALL_H
>>>
>>> +#include "cpu.h"
>>> #include "exec/tb-context.h"
>>> #include "sysemu/cpus.h"
>>>
>>> diff --git a/include/exec/ioport.h b/include/exec/ioport.h
>>> index a298b89ce1..97feb296d2 100644
>>> --- a/include/exec/ioport.h
>>> +++ b/include/exec/ioport.h
>>> @@ -24,6 +24,8 @@
>>> #ifndef IOPORT_H
>>> #define IOPORT_H
>>>
>>> +#include "exec/memory.h"
>>> +
>>> #define MAX_IOPORTS (64 * 1024)
>>> #define IOPORTS_MASK (MAX_IOPORTS - 1)
>>>
>>> diff --git a/include/exec/memory-internal.h b/include/exec/memory-internal.h
>>> index d1a9dd1ec8..ef4fb92371 100644
>>> --- a/include/exec/memory-internal.h
>>> +++ b/include/exec/memory-internal.h
>>> @@ -20,6 +20,8 @@
>>> #ifndef MEMORY_INTERNAL_H
>>> #define MEMORY_INTERNAL_H
>>>
>>> +#include "cpu.h"
>>> +
>>> #ifndef CONFIG_USER_ONLY
>>> static inline AddressSpaceDispatch *flatview_to_dispatch(FlatView *fv)
>>> {
>>> diff --git a/include/exec/ram_addr.h b/include/exec/ram_addr.h
>>> index b7b2e60ff6..a327a80cfe 100644
>>> --- a/include/exec/ram_addr.h
>>> +++ b/include/exec/ram_addr.h
>>> @@ -20,6 +20,7 @@
>>> #define RAM_ADDR_H
>>>
>>> #ifndef CONFIG_USER_ONLY
>>> +#include "cpu.h"
>>> #include "hw/xen/xen.h"
>>> #include "sysemu/tcg.h"
>>> #include "exec/ramlist.h"
>>> diff --git a/include/exec/softmmu-semi.h b/include/exec/softmmu-semi.h
>>> index 970837992e..fbcae88f4b 100644
>>> --- a/include/exec/softmmu-semi.h
>>> +++ b/include/exec/softmmu-semi.h
>>> @@ -10,6 +10,8 @@
>>> #ifndef SOFTMMU_SEMI_H
>>> #define SOFTMMU_SEMI_H
>>>
>>> +#include "cpu.h"
>>> +
>>> static inline uint64_t softmmu_tget64(CPUArchState *env, target_ulong addr)
>>> {
>>> uint64_t val;
>>> diff --git a/include/exec/tb-hash.h b/include/exec/tb-hash.h
>>> index 4f3a37d927..805235d321 100644
>>> --- a/include/exec/tb-hash.h
>>> +++ b/include/exec/tb-hash.h
>>> @@ -20,6 +20,8 @@
>>> #ifndef EXEC_TB_HASH_H
>>> #define EXEC_TB_HASH_H
>>>
>>> +#include "exec/cpu-defs.h"
>>> +#include "exec/exec-all.h"
>>> #include "qemu/xxhash.h"
>>>
>>> #ifdef CONFIG_SOFTMMU
>>> diff --git a/include/exec/user/thunk.h b/include/exec/user/thunk.h
>>> index 8d3af5a3be..d05a8a4dab 100644
>>> --- a/include/exec/user/thunk.h
>>> +++ b/include/exec/user/thunk.h
>>> @@ -20,6 +20,7 @@
>>> #define THUNK_H
>>>
>>> #include "cpu.h"
>>> +#include "exec/user/abitypes.h"
>>>
>>> /* types enums definitions */
>>
>> These all seem OK.
>>
>>>
>>> diff --git a/include/fpu/softfloat-macros.h b/include/fpu/softfloat-macros.h
>>> index c55aa6d174..be83a833ec 100644
>>> --- a/include/fpu/softfloat-macros.h
>>> +++ b/include/fpu/softfloat-macros.h
>>> @@ -82,6 +82,8 @@ this code that are retained.
>>> #ifndef FPU_SOFTFLOAT_MACROS_H
>>> #define FPU_SOFTFLOAT_MACROS_H
>>>
>>> +#include "fpu/softfloat.h"
>>> +
>>
>> What does softfloat-macros actually need from the core softfloat API?
>> These are lower level functions used by softfloat itself (and m68k for
>> it's own bit fiddling).
>
> I extracted this patch out of "[PATCH RFC v5 0/3] Baby steps towards
> saner headers". PATCH 1/3 creates make target "check-source", which is
> what I used to find headers that aren't self-contained. In this case:
>
> CC cris-softmmu/tests/headers-tgt/include/fpu/softfloat.o
> In file included from tests/headers-tgt/include/fpu/softfloat-macros.c:21:
> /work/armbru/qemu/include/fpu/softfloat-macros.h: In function ‘estimateDiv128To64’:
> /work/armbru/qemu/include/fpu/softfloat-macros.h:623:27: error: implicit declaration of function ‘LIT64’ [-Werror=implicit-function-declaration]
> 623 | if ( b <= a0 ) return LIT64( 0xFFFFFFFFFFFFFFFF );
The LIT64 definition should be moved to softfloat-types.h which is
already included by softfloat.h unless we already have a QEMU expansion
we should be using. The softfloat-macros.h can include softfloat-types.h
as well and we should only include the full softfloat.h if they need it.
Do you want me to spin up a patch?
> | ^~~~~
> /work/armbru/qemu/include/fpu/softfloat-macros.h:623:27: error: nested extern declaration of ‘LIT64’ [-Werror=nested-externs]
> /work/armbru/qemu/include/fpu/softfloat-macros.h: At top level:
> /work/armbru/qemu/include/fpu/softfloat-macros.h:761:15: error: unknown type name ‘flag’
> 761 | static inline flag eq128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 )
> | ^~~~
> /work/armbru/qemu/include/fpu/softfloat-macros.h:774:15: error: unknown type name ‘flag’
> 774 | static inline flag le128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 )
> | ^~~~
> /work/armbru/qemu/include/fpu/softfloat-macros.h:787:15: error: unknown type name ‘flag’
> 787 | static inline flag lt128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 )
> | ^~~~
> /work/armbru/qemu/include/fpu/softfloat-macros.h:800:15: error: unknown type name ‘flag’
> 800 | static inline flag ne128( uint64_t a0, uint64_t a1, uint64_t b0, uint64_t b1 )
> | ^~~~
> cc1: all warnings being treated as errors
> make[1]: *** [/work/armbru/qemu/rules.mak:69: tests/headers-tgt/include/fpu/softfloat-macros.o] Error 1
--
Alex Bennée
next prev parent reply other threads:[~2019-08-07 21:29 UTC|newest]
Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-06 15:14 [Qemu-devel] [PATCH v2 00/29] Tame a few "touch this, recompile the world" headers Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 01/29] include: Make headers more self-contained Markus Armbruster
2019-08-07 15:03 ` Alex Bennée
2019-08-07 19:45 ` Markus Armbruster
2019-08-07 21:28 ` Alex Bennée [this message]
2019-08-08 4:21 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 02/29] Include generated QAPI headers less Markus Armbruster
2019-08-06 21:50 ` Eric Blake
2019-08-07 9:53 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 03/29] qapi: Split error.json off common.json Markus Armbruster
2019-08-07 9:57 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 04/29] memory: Fix type of IOMMUMemoryRegionClass member @parent_class Markus Armbruster
2019-08-07 10:11 ` Philippe Mathieu-Daudé
2019-08-07 10:16 ` Paolo Bonzini
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 05/29] queue: Drop superfluous #include qemu/atomic.h Markus Armbruster
2019-08-07 10:21 ` Philippe Mathieu-Daudé
2019-08-07 15:40 ` Alex Bennée
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 06/29] trace: Eliminate use of TARGET_FMT_plx Markus Armbruster
2019-08-07 9:29 ` Stefan Hajnoczi
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 07/29] trace: Do not include qom/cpu.h into generated trace.h Markus Armbruster
2019-08-07 9:32 ` Stefan Hajnoczi
2019-08-07 10:30 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 08/29] Include sysemu/reset.h a lot less Markus Armbruster
2019-08-07 10:38 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 09/29] Include migration/qemu-file-types.h " Markus Armbruster
2019-08-07 12:25 ` Philippe Mathieu-Daudé
2019-08-07 17:30 ` Philippe Mathieu-Daudé
2019-08-07 19:46 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 10/29] ide: Include hw/ide/internal a bit less outside hw/ide/ Markus Armbruster
2019-08-06 21:40 ` John Snow
2019-08-07 12:29 ` Philippe Mathieu-Daudé
2019-08-07 14:56 ` John Snow
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 11/29] typedefs: Separate incomplete types and function types Markus Armbruster
2019-08-07 12:31 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 12/29] Include hw/irq.h a lot less Markus Armbruster
2019-08-07 13:04 ` Philippe Mathieu-Daudé
2019-08-07 21:06 ` Eric Blake
2019-08-08 4:27 ` Markus Armbruster
2019-08-08 5:09 ` Richard Henderson
2019-08-08 11:48 ` Eric Blake
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 13/29] Clean up inclusion of exec/cpu-common.h Markus Armbruster
2019-08-07 14:20 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 14/29] migration: Move the VMStateDescription typedef to typedefs.h Markus Armbruster
2019-08-07 15:22 ` Alex Bennée
2019-08-07 17:13 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 15/29] Include migration/vmstate.h less Markus Armbruster
2019-08-07 14:44 ` Philippe Mathieu-Daudé
2019-08-08 11:36 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 16/29] Include exec/memory.h slightly less Markus Armbruster
2019-08-07 14:50 ` Philippe Mathieu-Daudé
2019-08-08 12:16 ` Markus Armbruster
2019-08-08 12:36 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 17/29] Include qom/object.h " Markus Armbruster
2019-08-07 17:11 ` Philippe Mathieu-Daudé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 18/29] Include hw/hw.h exactly where needed Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 19/29] Include qemu/queue.h slightly less Markus Armbruster
2019-08-07 15:57 ` Alex Bennée
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 20/29] Include qemu/main-loop.h less Markus Armbruster
2019-08-07 13:18 ` Philippe Mathieu-Daudé
2019-08-07 13:49 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 21/29] Include hw/qdev-properties.h less Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 22/29] Include hw/boards.h a bit less Markus Armbruster
2019-08-06 21:37 ` Alistair Francis
2019-08-07 17:26 ` Philippe Mathieu-Daudé
2019-08-07 17:57 ` Eduardo Habkost
2019-08-07 18:05 ` Philippe Mathieu-Daudé
2019-08-07 18:19 ` Eduardo Habkost
2019-08-08 11:57 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 23/29] numa: Don't include hw/boards.h into sysemu/numa.h Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 24/29] Include sysemu/hostmem.h less Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 25/29] numa: Move remaining NUMA declarations from sysemu.h to numa.h Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 26/29] Clean up inclusion of sysemu/sysemu.h Markus Armbruster
2019-08-06 21:36 ` Alistair Francis
2019-08-07 15:47 ` Alex Bennée
2019-08-07 20:10 ` Markus Armbruster
2019-08-08 14:21 ` Alex Bennée
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 27/29] Include sysemu/sysemu.h a lot less Markus Armbruster
2019-08-06 21:38 ` Alistair Francis
2019-08-07 9:28 ` Stefan Hajnoczi
2019-08-07 13:24 ` Philippe Mathieu-Daudé
2019-08-07 13:51 ` Markus Armbruster
2019-08-07 17:31 ` Philippe Mathieu-Daudé
2019-08-07 20:16 ` Markus Armbruster
2019-08-07 21:05 ` Philippe Mathieu-Daudé
2019-08-08 4:48 ` [Qemu-devel] Is network backend netmap worth keeping? (was: [PATCH v2 27/29] Include sysemu/sysemu.h a lot less) Markus Armbruster
2019-08-08 5:38 ` [Qemu-devel] Is network backend netmap worth keeping? Jason Wang
2019-08-08 13:36 ` Philippe Mathieu-Daudé
2019-08-08 17:27 ` Vincenzo Maffione
2019-08-08 7:27 ` Giuseppe Lettieri
2019-08-08 11:52 ` Markus Armbruster
2019-08-12 12:32 ` Philippe Mathieu-Daudé
2019-08-12 12:34 ` Philippe Mathieu-Daudé
2019-09-02 20:50 ` Giuseppe Lettieri
2019-09-13 13:04 ` Markus Armbruster
2019-09-16 14:45 ` Giuseppe Lettieri
2019-09-23 11:21 ` Markus Armbruster
2019-10-04 13:02 ` [PATCH] netmap: support git-submodule build otption Giuseppe Lettieri
2019-10-04 13:08 ` Peter Maydell
2019-10-07 10:49 ` Markus Armbruster
2019-10-07 11:58 ` Peter Maydell
2019-10-07 12:35 ` Markus Armbruster
2019-10-07 12:39 ` Peter Maydell
2019-10-08 9:17 ` Daniel P. Berrangé
2019-10-08 11:57 ` Markus Armbruster
2019-10-07 15:37 ` Thomas Huth
2019-10-07 17:53 ` Markus Armbruster
2019-10-07 15:44 ` Markus Armbruster
2019-08-08 8:12 ` [Qemu-devel] Is network backend netmap worth keeping? (was: [PATCH v2 27/29] Include sysemu/sysemu.h a lot less) Stefano Garzarella
2019-10-07 18:21 ` Is network backend vde worth keeping? (was: Is network backend netmap worth keeping?) Markus Armbruster
2019-10-09 9:13 ` Thomas Huth
2019-10-11 6:55 ` Thomas Huth
2019-10-10 17:07 ` Julia Suvorova
2019-10-10 17:39 ` Daniel P. Berrangé
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 28/29] sysemu: Move the VMChangeStateEntry typedef to qemu/typedefs.h Markus Armbruster
2019-08-07 15:26 ` Alex Bennée
2019-08-07 20:19 ` Markus Armbruster
2019-08-06 15:14 ` [Qemu-devel] [PATCH v2 29/29] sysemu: Split sysemu/runstate.h off sysemu/sysemu.h Markus Armbruster
2019-08-07 4:40 ` Markus Armbruster
2019-08-07 10:15 ` Paolo Bonzini
2019-08-06 16:02 ` [Qemu-devel] [PATCH v2 00/29] Tame a few "touch this, recompile the world" headers no-reply
2019-08-06 16:56 ` no-reply
2019-08-07 16:32 ` Alex Bennée
2019-08-07 20:21 ` Markus Armbruster
2019-08-08 13:21 ` Markus Armbruster
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=878ss4is15.fsf@linaro.org \
--to=alex.bennee@linaro.org \
--cc=alistair.francis@wdc.com \
--cc=armbru@redhat.com \
--cc=qemu-devel@nongnu.org \
/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).