public inbox for
 help / color / mirror / Atom feed
From: Sam James <>
To: Mike Frysinger <>
Cc: Mark Wielaard <>,
Subject: Re: [PATCH] sim: common, microblaze, mn10300: handle signal.h defining REC_PC.
Date: Sun, 15 Jan 2023 00:58:24 +0000	[thread overview]
Message-ID: <> (raw)
In-Reply-To: <Y8NHRQHwYtS4joYo@vapier>

[-- Attachment #1: Type: text/plain, Size: 2462 bytes --]

> On 15 Jan 2023, at 00:22, Mike Frysinger <> wrote:
> On 15 Jan 2023 00:28, Mark Wielaard wrote:
>> signal.h isn't needed in microblaze and mn10300 interp.c
>> so don't include it.
> these changes are fine to merge if you want to split it out
>> In common sim-events.c, sim-signal.c, nrun.c and dv-sockser.c we
>> do need signal.h, but check whether REG_PC is defined (and then
>> undefine it) before including the sim headers.
>> It breaks the build on sparc because signal.h indirectly
>> includes /usr/include/sys/ucontext.h and defines REG_PC,
>> which is also defined in microblaze-opcm.h
> i don't think this is correct.  none of the files quoted use REG_PC,
> so undefining a random symbol in them doesn't make sense.  nothing in
> sim/common/ uses REG_PC for that matter.

The original error ( is:
In file included from ../../binutils-gdb/sim/mn10300/sim-main.h:41,
from ../../binutils-gdb/sim/common/dv-sockser.c:42:
../../binutils-gdb/sim/mn10300/mn10300-sim.h:68: error: "REG_PC" redefined [-Werror]
68 | #define REG_PC 9
In file included from /usr/include/signal.h:316,
from ../gnulib/import/signal.h:52,
from ../../binutils-gdb/sim/common/dv-sockser.c:29:
/usr/include/sys/ucontext.h:111: note: this is the location of the previous definition
111 | # define REG_PC (1)

There's history of just ducking this in other projects, and I can't really blame them:

Overall, we have:
$ grep -rsin "#define.*REG_PC"
sim/mn10300/mn10300_sim.h:57:#define PC (State.regs[REG_PC])
sim/mn10300/mn10300_sim.h:74:#define REG_PC 9
gas/config/tc-arm.c:744:#define REG_PC  15
gprofng/libcollector/unwind.c:111:#define GET_PC(ctx) (((ucontext_t*)ctx)->uc_mcontext.gregs[REG_PC])
opcodes/microblaze-opcm.h:77:#define REG_PC_MASK 0x8000
opcodes/microblaze-opcm.h:101:#define REG_PC  32 /* PC.  */
include/opcode/cris.h:35:#define REG_PC (15)
include/opcode/cris.h:143:#define BDAP_PC_LOW     (BDAP_INDIR_LOW + REG_PC)

What do you prefer?

1. Rename all the REG_* (ugly)
2. #undef hack in each of the consumers where there's a #define for it?
3. What Mark did in some misc. top-level sim place
4. Beg every vendor to change their ucontext.h
5. Something else?


[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 358 bytes --]

  reply	other threads:[~2023-01-15  0:58 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-14 23:28 Mark Wielaard
2023-01-15  0:22 ` Mike Frysinger
2023-01-15  0:58   ` Sam James [this message]
2023-01-15  8:07     ` Mike Frysinger
2023-01-15 17:48   ` Mark Wielaard

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:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \ \ \ \ \ \

* 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 read-only IMAP folder(s) and NNTP newsgroup(s).