From: Anthony Liguori <anthony@codemonkey.ws>
To: Paolo Bonzini <pbonzini@redhat.com>, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH 05/30] exec: do not use qemu/tls.h
Date: Fri, 28 Jun 2013 15:43:15 -0500 [thread overview]
Message-ID: <8738s29db0.fsf@codemonkey.ws> (raw)
In-Reply-To: <1372444009-11544-6-git-send-email-pbonzini@redhat.com>
Paolo Bonzini <pbonzini@redhat.com> writes:
> The next patch will change qemu/tls.h to support more platforms, but at
> some performance cost. Declare cpu_single_env directly instead of using
> the tls.h abstractions.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Anthony Liguori <aliguori@us.ibm.com>
Regards,
Anthony Liguori
> ---
> exec.c | 10 ++++++++--
> include/exec/cpu-all.h | 14 +++++++++++---
> include/qemu/tls.h | 52 --------------------------------------------------
> 3 files changed, 19 insertions(+), 57 deletions(-)
> delete mode 100644 include/qemu/tls.h
>
> diff --git a/exec.c b/exec.c
> index d28403b..a788981 100644
> --- a/exec.c
> +++ b/exec.c
> @@ -70,9 +70,15 @@ static MemoryRegion io_mem_unassigned;
> #endif
>
> CPUArchState *first_cpu;
> +
> /* current CPU in the current thread. It is only valid inside
> - cpu_exec() */
> -DEFINE_TLS(CPUArchState *,cpu_single_env);
> + * cpu_exec(). See comment in include/exec/cpu-all.h. */
> +#if defined CONFIG_KVM || (defined CONFIG_USER_ONLY && defined CONFIG_USE_NPTL)
> +__thread CPUArchState *cpu_single_env;
> +#else
> +CPUArchState *cpu_single_env;
> +#endif
> +
> /* 0 = Do not count executed instructions.
> 1 = Precise instruction counting.
> 2 = Adaptive rate instruction counting. */
> diff --git a/include/exec/cpu-all.h b/include/exec/cpu-all.h
> index e9c3717..2202ba3 100644
> --- a/include/exec/cpu-all.h
> +++ b/include/exec/cpu-all.h
> @@ -20,7 +20,6 @@
> #define CPU_ALL_H
>
> #include "qemu-common.h"
> -#include "qemu/tls.h"
> #include "exec/cpu-common.h"
> #include "qemu/thread.h"
>
> @@ -368,8 +367,17 @@ void cpu_dump_statistics(CPUArchState *env, FILE *f, fprintf_function cpu_fprint
> void QEMU_NORETURN cpu_abort(CPUArchState *env, const char *fmt, ...)
> GCC_FMT_ATTR(2, 3);
> extern CPUArchState *first_cpu;
> -DECLARE_TLS(CPUArchState *,cpu_single_env);
> -#define cpu_single_env tls_var(cpu_single_env)
> +
> +/* This is thread-local depending on __linux__ because:
> + * - the only -user mode supporting multiple VCPU threads is linux-user
> + * - TCG system mode is single-threaded regarding VCPUs
> + * - KVM system mode is multi-threaded but limited to Linux
> + */
> +#if defined CONFIG_KVM || (defined CONFIG_USER_ONLY && defined CONFIG_USE_NPTL)
> +extern __thread CPUArchState *cpu_single_env;
> +#else
> +extern CPUArchState *cpu_single_env;
> +#endif
>
> /* Flags for use in ENV->INTERRUPT_PENDING.
>
> diff --git a/include/qemu/tls.h b/include/qemu/tls.h
> deleted file mode 100644
> index b92ea9d..0000000
> --- a/include/qemu/tls.h
> +++ /dev/null
> @@ -1,52 +0,0 @@
> -/*
> - * Abstraction layer for defining and using TLS variables
> - *
> - * Copyright (c) 2011 Red Hat, Inc
> - * Copyright (c) 2011 Linaro Limited
> - *
> - * Authors:
> - * Paolo Bonzini <pbonzini@redhat.com>
> - * Peter Maydell <peter.maydell@linaro.org>
> - *
> - * This program is free software; you can redistribute it and/or
> - * modify it under the terms of the GNU General Public License as
> - * published by the Free Software Foundation; either version 2 of
> - * the License, or (at your option) any later version.
> - *
> - * This program is distributed in the hope that it will be useful,
> - * but WITHOUT ANY WARRANTY; without even the implied warranty of
> - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> - * GNU General Public License for more details.
> - *
> - * You should have received a copy of the GNU General Public License along
> - * with this program; if not, see <http://www.gnu.org/licenses/>.
> - */
> -
> -#ifndef QEMU_TLS_H
> -#define QEMU_TLS_H
> -
> -/* Per-thread variables. Note that we only have implementations
> - * which are really thread-local on Linux; the dummy implementations
> - * define plain global variables.
> - *
> - * This means that for the moment use should be restricted to
> - * per-VCPU variables, which are OK because:
> - * - the only -user mode supporting multiple VCPU threads is linux-user
> - * - TCG system mode is single-threaded regarding VCPUs
> - * - KVM system mode is multi-threaded but limited to Linux
> - *
> - * TODO: proper implementations via Win32 .tls sections and
> - * POSIX pthread_getspecific.
> - */
> -#ifdef __linux__
> -#define DECLARE_TLS(type, x) extern DEFINE_TLS(type, x)
> -#define DEFINE_TLS(type, x) __thread __typeof__(type) tls__##x
> -#define tls_var(x) tls__##x
> -#else
> -/* Dummy implementations which define plain global variables */
> -#define DECLARE_TLS(type, x) extern DEFINE_TLS(type, x)
> -#define DEFINE_TLS(type, x) __typeof__(type) tls__##x
> -#define tls_var(x) tls__##x
> -#endif
> -
> -#endif
> --
> 1.8.1.4
next prev parent reply other threads:[~2013-06-28 20:43 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-06-28 18:26 [Qemu-devel] [PATCH 00/30] Memory API changes for 1.6: RCU-protected address space dispatch Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 01/30] memory: access FlatView from a local variable Paolo Bonzini
2013-06-28 20:01 ` Anthony Liguori
2013-06-28 18:26 ` [Qemu-devel] [PATCH 02/30] memory: use a new FlatView pointer on every topology update Paolo Bonzini
2013-06-28 20:02 ` Anthony Liguori
2013-06-28 18:26 ` [Qemu-devel] [PATCH 03/30] memory: add reference counting to FlatView Paolo Bonzini
2013-06-28 20:07 ` Anthony Liguori
2013-06-28 18:26 ` [Qemu-devel] [PATCH 04/30] add a header file for atomic operations Paolo Bonzini
2013-06-28 20:41 ` Anthony Liguori
2013-07-01 10:21 ` Paolo Bonzini
2013-07-01 13:00 ` Anthony Liguori
2013-07-01 13:04 ` Paolo Bonzini
2013-07-01 13:20 ` Anthony Liguori
2013-07-04 5:24 ` liu ping fan
2013-07-01 11:08 ` Peter Maydell
2013-07-03 2:24 ` liu ping fan
2013-07-03 5:59 ` Paolo Bonzini
2013-07-03 7:07 ` liu ping fan
2013-06-28 18:26 ` [Qemu-devel] [PATCH 05/30] exec: do not use qemu/tls.h Paolo Bonzini
2013-06-28 20:43 ` Anthony Liguori [this message]
2013-06-28 23:53 ` Ed Maste
2013-07-01 10:16 ` Paolo Bonzini
2013-06-29 10:55 ` Peter Maydell
2013-07-01 10:45 ` Paolo Bonzini
2013-07-01 11:05 ` Peter Maydell
2013-07-01 16:21 ` Paolo Bonzini
2013-07-01 16:26 ` Peter Maydell
2013-07-01 20:52 ` Paolo Bonzini
2013-07-01 21:34 ` Peter Maydell
2013-07-02 13:40 ` Andreas Färber
2013-07-02 14:06 ` Alexander Graf
2013-06-28 18:26 ` [Qemu-devel] [PATCH 06/30] qemu-thread: add TLS wrappers Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 07/30] qemu-thread: add QemuEvent Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 08/30] rcu: add rcu library Paolo Bonzini
2013-07-01 9:47 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 09/30] qemu-thread: register threads with RCU Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 10/30] rcu: add call_rcu Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 11/30] rcu: add rcutorture Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 12/30] rcu: allow nested calls to rcu_thread_offline/rcu_thread_online Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 13/30] qemu-thread: report RCU quiescent states Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 14/30] event loop: " Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 15/30] cpus: " Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 16/30] block: " Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 17/30] migration: " Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 18/30] memory: protect current_map by RCU Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 19/30] memory: avoid ref/unref in memory_region_find Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 20/30] exec: change well-known physical sections to macros Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 21/30] exec: separate current memory map from the one being built Paolo Bonzini
2013-07-02 14:41 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 22/30] memory: move MemoryListener declaration earlier Paolo Bonzini
2013-07-02 14:41 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 23/30] exec: move listener from AddressSpaceDispatch to AddressSpace Paolo Bonzini
2013-07-02 14:41 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 24/30] exec: separate current radix tree from the one being built Paolo Bonzini
2013-07-02 14:41 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 25/30] exec: put memory map in AddressSpaceDispatch Paolo Bonzini
2013-07-02 14:42 ` Jan Kiszka
2013-07-02 15:08 ` Paolo Bonzini
2013-07-02 15:48 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 26/30] exec: remove cur_map Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 27/30] exec: change some APIs to take AddressSpaceDispatch Paolo Bonzini
2013-07-02 14:47 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 28/30] exec: change iotlb " Paolo Bonzini
2013-07-02 10:00 ` Jan Kiszka
2013-06-28 18:26 ` [Qemu-devel] [PATCH 29/30] exec: add a reference to the region returned by address_space_translate Paolo Bonzini
2013-06-28 18:26 ` [Qemu-devel] [PATCH 30/30] exec: put address space dispatch under RCU critical section Paolo Bonzini
2013-06-28 19:38 ` Jan Kiszka
2013-07-01 11:48 ` Paolo Bonzini
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=8738s29db0.fsf@codemonkey.ws \
--to=anthony@codemonkey.ws \
--cc=pbonzini@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).