All of lore.kernel.org
 help / color / mirror / Atom feed
From: Warner Losh <imp@bsdimp.com>
To: qemu-devel@nongnu.org
Cc: Warner Losh <imp@bsdimp.com>, Kyle Evans <kevans@freebsd.org>,
	Richard Henderson <richard.henderson@linaro.org>
Subject: [PULL 03/11] bsd-user: Flag shared memory as needing atomics
Date: Tue, 22 Sep 2026 13:21:01 -0600	[thread overview]
Message-ID: <20260922192109.98467-4-imp@bsdimp.com> (raw)
In-Reply-To: <20260922192109.98467-1-imp@bsdimp.com>

Catch up with linux-user 228168cbb7c:
    Signal the translator to use host atomic instructions for
    guest operations, insofar as it is possible.  This is the
    best we can do to allow the guest to interact atomically
    with other processes.

as well as subsequent tweaks and adjustmentst for bsd-user's code
structure (we don't have one exit path linux-user does and could frankly
stand to resturcture a bit to match the evolved code in linux-user). On
non-private mmap(2) sharing and all successful shmat(2) calls we tell
tcg that we're in a parallel context, which changes how it does atomics.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
---
 bsd-user/bsd-mem.h | 8 ++++++++
 bsd-user/mmap.c    | 9 +++++++++
 2 files changed, 17 insertions(+)

diff --git a/bsd-user/bsd-mem.h b/bsd-user/bsd-mem.h
index 3086143d40..65270e2cdf 100644
--- a/bsd-user/bsd-mem.h
+++ b/bsd-user/bsd-mem.h
@@ -389,6 +389,14 @@ static inline abi_long do_bsd_shmat(int shmid, abi_ulong shmaddr, int shmflg)
         }
     }
 
+    /*
+     * We're mapping shared memory, so ensure we generate code for parallel
+     * execution and flush old translations.  This will work up to the level
+     * supported by the host -- anything that requires EXCP_ATOMIC will not
+     * be atomic with respect to an external process.
+     */
+    begin_parallel_context(thread_cpu);
+
     return raddr;
 }
 
diff --git a/bsd-user/mmap.c b/bsd-user/mmap.c
index fe77eceb48..548b4824df 100644
--- a/bsd-user/mmap.c
+++ b/bsd-user/mmap.c
@@ -662,6 +662,15 @@ abi_long target_mmap(abi_ulong start, abi_ulong len, int prot,
     printf("\n");
 #endif
     mmap_unlock();
+    /*
+     * If we're mapping shared memory, ensure we generate code for parallel
+     * execution and flush old translations.  This will work up to the level
+     * supported by the host -- anything that requires EXCP_ATOMIC will not
+     * be atomic with respect to an external process.
+     */
+    if ((flags & MAP_TYPE) != MAP_PRIVATE) {
+        begin_parallel_context(thread_cpu);
+    }
     return start;
 fail:
     mmap_unlock();
-- 
2.55.0



  parent reply	other threads:[~2026-09-22 19:22 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-22 19:20 [PULL 00/11] Bsd user 2026q3 patches Warner Losh
2026-09-22 19:20 ` [PULL 01/11] bsd-user: HOST_BIG_ENDIAN and TARGET_BIG_ENDIAN are always Warner Losh
2026-09-22 19:21 ` [PULL 02/11] bsd-user: Drop support for FreeBSD 12 Warner Losh
2026-09-22 19:21 ` Warner Losh [this message]
2026-09-22 19:21 ` [PULL 04/11] bsd-user: Initialize per-arch child registers Warner Losh
2026-09-22 19:21 ` [PULL 05/11] bsd-user: Fix crazy bug with mmap Warner Losh
2026-09-22 19:21 ` [PULL 06/11] bsd-user: Move tb_flush stuff to begin_parallel_context Warner Losh
2026-09-22 19:21 ` [PULL 07/11] bsd-user: Implement exterrctl(2) Warner Losh
2026-09-22 19:21 ` [PULL 08/11] bsd-user: Implement sigfastblock Warner Losh
2026-09-22 19:21 ` [PULL 09/11] bsd-user: Eliminate unused regs arg in load_elf_binary Warner Losh
2026-09-22 19:21 ` [PULL 10/11] bsd-user: #ifdef sockopt that aren't in 14 Warner Losh
2026-09-22 19:21 ` [PULL 11/11] bsd-user: Remove host-os.h Warner Losh
2026-09-23 19:44 ` [PULL 00/11] Bsd user 2026q3 patches Richard Henderson

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=20260922192109.98467-4-imp@bsdimp.com \
    --to=imp@bsdimp.com \
    --cc=kevans@freebsd.org \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.