qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: Claudio Fontana <cfontana@suse.de>
Subject: [PULL 21/33] accel: add init_accel_cpu for adapting accel behavior to CPU type
Date: Tue, 11 May 2021 04:13:38 -0400	[thread overview]
Message-ID: <20210511081350.419428-22-pbonzini@redhat.com> (raw)
In-Reply-To: <20210511081350.419428-1-pbonzini@redhat.com>

From: Claudio Fontana <cfontana@suse.de>

while on x86 all CPU classes can use the same set of TCGCPUOps,
on ARM the right accel behavior depends on the type of the CPU.

So we need a way to specialize the accel behavior according to
the CPU. Therefore, add a second initialization, after the
accel_cpu->cpu_class_init, that allows to do this.

Signed-off-by: Claudio Fontana <cfontana@suse.de>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210322132800.7470-24-cfontana@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 accel/accel-common.c      | 13 +++++++++++++
 include/hw/core/cpu.h     |  6 ++++++
 target/i386/tcg/tcg-cpu.c |  8 +++++++-
 3 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/accel/accel-common.c b/accel/accel-common.c
index d77c09d7b5..cf07f78421 100644
--- a/accel/accel-common.c
+++ b/accel/accel-common.c
@@ -54,10 +54,23 @@ static void accel_init_cpu_int_aux(ObjectClass *klass, void *opaque)
     CPUClass *cc = CPU_CLASS(klass);
     AccelCPUClass *accel_cpu = opaque;
 
+    /*
+     * The first callback allows accel-cpu to run initializations
+     * for the CPU, customizing CPU behavior according to the accelerator.
+     *
+     * The second one allows the CPU to customize the accel-cpu
+     * behavior according to the CPU.
+     *
+     * The second is currently only used by TCG, to specialize the
+     * TCGCPUOps depending on the CPU type.
+     */
     cc->accel_cpu = accel_cpu;
     if (accel_cpu->cpu_class_init) {
         accel_cpu->cpu_class_init(cc);
     }
+    if (cc->init_accel_cpu) {
+        cc->init_accel_cpu(accel_cpu, cc);
+    }
 }
 
 /* initialize the arch-specific accel CpuClass interfaces */
diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index c68bc3ba8a..d45f78290e 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -192,6 +192,12 @@ struct CPUClass {
 
     /* when TCG is not available, this pointer is NULL */
     struct TCGCPUOps *tcg_ops;
+
+    /*
+     * if not NULL, this is called in order for the CPUClass to initialize
+     * class data that depends on the accelerator, see accel/accel-common.c.
+     */
+    void (*init_accel_cpu)(struct AccelCPUClass *accel_cpu, CPUClass *cc);
 };
 
 /*
diff --git a/target/i386/tcg/tcg-cpu.c b/target/i386/tcg/tcg-cpu.c
index e311f52855..ba39531aa5 100644
--- a/target/i386/tcg/tcg-cpu.c
+++ b/target/i386/tcg/tcg-cpu.c
@@ -69,11 +69,17 @@ static struct TCGCPUOps x86_tcg_ops = {
 #endif /* !CONFIG_USER_ONLY */
 };
 
-static void tcg_cpu_class_init(CPUClass *cc)
+static void tcg_cpu_init_ops(AccelCPUClass *accel_cpu, CPUClass *cc)
 {
+    /* for x86, all cpus use the same set of operations */
     cc->tcg_ops = &x86_tcg_ops;
 }
 
+static void tcg_cpu_class_init(CPUClass *cc)
+{
+    cc->init_accel_cpu = tcg_cpu_init_ops;
+}
+
 /*
  * TCG-specific defaults that override all CPU models when using TCG
  */
-- 
2.26.2




  parent reply	other threads:[~2021-05-11  8:36 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-05-11  8:13 [PULL 00/33] Misc (mostly i386) patches for 2021-05-11 Paolo Bonzini
2021-05-11  8:13 ` [PULL 01/33] target/i386: Rename helper_fldt, helper_fstt Paolo Bonzini
2021-05-11  8:13 ` [PULL 02/33] target/i386: Split out do_fsave, do_frstor, do_fxsave, do_fxrstor Paolo Bonzini
2021-05-11  8:13 ` [PULL 03/33] i386: split cpu accelerators from cpu.c, using AccelCPUClass Paolo Bonzini
2021-08-20 16:55   ` Peter Maydell
2021-08-20 17:02     ` Philippe Mathieu-Daudé
2021-08-20 17:10       ` Peter Maydell
2021-05-11  8:13 ` [PULL 04/33] cpu: call AccelCPUClass::cpu_realizefn in cpu_exec_realizefn Paolo Bonzini
2021-05-11  8:13 ` [PULL 05/33] accel: introduce new accessor functions Paolo Bonzini
2021-05-11  8:13 ` [PULL 06/33] target/i386: fix host_cpu_adjust_phys_bits error handling Paolo Bonzini
2021-05-11  8:13 ` [PULL 07/33] accel-cpu: make cpu_realizefn return a bool Paolo Bonzini
2021-05-11  8:13 ` [PULL 08/33] i386: split off sysemu-only functionality in tcg-cpu Paolo Bonzini
2021-05-11  8:13 ` [PULL 09/33] i386: split smm helper (sysemu) Paolo Bonzini
2021-05-11  8:13 ` [PULL 10/33] i386: split tcg excp_helper into sysemu and user parts Paolo Bonzini
2021-05-11  8:13 ` [PULL 11/33] i386: move TCG bpt_helper into sysemu/ Paolo Bonzini
2021-05-11  8:13 ` [PULL 12/33] i386: split misc helper user stubs and sysemu part Paolo Bonzini
2021-05-11  8:13 ` [PULL 13/33] i386: separate fpu_helper sysemu-only parts Paolo Bonzini
2021-05-11  8:13 ` [PULL 14/33] i386: split svm_helper into sysemu and stub-only user Paolo Bonzini
2021-05-11  8:13 ` [PULL 15/33] i386: split seg_helper into user-only and sysemu parts Paolo Bonzini
2021-05-11  8:13 ` [PULL 16/33] i386: split off sysemu part of cpu.c Paolo Bonzini
2021-05-11  8:13 ` [PULL 17/33] target/i386: gdbstub: introduce aux functions to read/write CS64 regs Paolo Bonzini
2021-05-11  8:13 ` [PULL 18/33] target/i386: gdbstub: only write CR0/CR2/CR3/EFER for sysemu Paolo Bonzini
2021-05-11  8:13 ` [PULL 19/33] i386: make cpu_load_efer sysemu-only Paolo Bonzini
2021-05-11  8:13 ` [PULL 20/33] accel: move call to accel_init_interfaces Paolo Bonzini
2021-05-11  8:13 ` Paolo Bonzini [this message]
2021-05-11  8:13 ` [PULL 22/33] target/i386: merge SVM_NPTEXIT_* with PF_ERROR_* constants Paolo Bonzini
2021-05-11  8:13 ` [PULL 23/33] target/i386: move paging mode constants from SVM to cpu.h Paolo Bonzini
2021-05-11  8:13 ` [PULL 24/33] target/i386: extract mmu_translate Paolo Bonzini
2021-05-11  8:13 ` [PULL 25/33] target/i386: pass cr3 to mmu_translate Paolo Bonzini
2021-05-11  8:13 ` [PULL 26/33] target/i386: extend pg_mode to more CR0 and CR4 bits Paolo Bonzini
2021-05-11  8:13 ` [PULL 27/33] target/i386: allow customizing the next phase of the translation Paolo Bonzini
2021-05-11  8:13 ` [PULL 28/33] target/i386: use mmu_translate for NPT walk Paolo Bonzini
2021-05-11  8:13 ` [PULL 29/33] main-loop: remove dead code Paolo Bonzini
2021-05-11  8:13 ` [PULL 30/33] migration: do not restart VM after successful snapshot-load Paolo Bonzini
2021-05-11  8:56   ` Dr. David Alan Gilbert
2021-05-12  8:05     ` Paolo Bonzini
2021-05-11  9:09   ` Daniel P. Berrangé
2021-05-11 17:06   ` Daniel P. Berrangé
2021-05-12  7:45     ` Kevin Wolf
2021-05-12 18:11       ` Paolo Bonzini
2021-05-11  8:13 ` [PULL 31/33] qemu-option: support accept-any QemuOptsList in qemu_opts_absorb_qdict Paolo Bonzini
2021-05-11  8:13 ` [PULL 32/33] configure: fix detection of gdbus-codegen Paolo Bonzini
2021-05-11  8:13 ` [PULL 33/33] coverity-scan: list components, move model to scripts/coverity-scan Paolo Bonzini
2021-05-11  8:48 ` [PULL 00/33] Misc (mostly i386) patches for 2021-05-11 no-reply
2021-05-12 16:29 ` Peter Maydell

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=20210511081350.419428-22-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=cfontana@suse.de \
    --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).