All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eduardo Habkost <ehabkost@redhat.com>
To: Greg Kurz <groug@kaod.org>
Cc: qemu-devel@nongnu.org, qemu-ppc@nongnu.org,
	"David Gibson" <david@gibson.dropbear.id.au>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Richard Henderson" <rth@twiddle.net>,
	"Cédric Le Goater" <clg@kaod.org>
Subject: Re: [Qemu-devel] [PATCH 3/3] accel: forbid early use of kvm_enabled() and friends
Date: Fri, 29 Jun 2018 16:58:23 -0300	[thread overview]
Message-ID: <20180629195823.GX7451@localhost.localdomain> (raw)
In-Reply-To: <153018093372.336571.7266716996862582164.stgit@bahia>

On Thu, Jun 28, 2018 at 12:15:33PM +0200, Greg Kurz wrote:
> It is unsafe to rely on *_enabled() helpers before the accelerator has
> been initialized, ie, accel_init_machine() has succeeded, because they
> always return false. But it is still possible to end up calling them
> indirectly by inadvertance, and cause QEMU to misbehave.
> 
> This patch causes QEMU to abort if we try to check for an accelerator
> before it has been set up. This will help to catch bugs earlier.
> 
> Signed-off-by: Greg Kurz <groug@kaod.org>
> ---
>  accel/accel.c          |    7 +++++++
>  include/qemu-common.h  |    3 ++-
>  include/sysemu/accel.h |    1 +
>  include/sysemu/kvm.h   |    3 ++-
>  qom/cpu.c              |    1 +
>  stubs/Makefile.objs    |    1 +
>  stubs/accel.c          |   14 ++++++++++++++
>  target/i386/hax-all.c  |    2 +-
>  target/i386/whpx-all.c |    2 +-
>  9 files changed, 30 insertions(+), 4 deletions(-)
>  create mode 100644 stubs/accel.c
> 
> diff --git a/accel/accel.c b/accel/accel.c
> index 966b2d8f536c..27900aac9cc5 100644
> --- a/accel/accel.c
> +++ b/accel/accel.c
> @@ -51,6 +51,13 @@ static AccelClass *accel_find(const char *opt_name)
>      return ac;
>  }
>  
> +bool assert_accelerator_initialized(bool allowed)
> +{
> +    assert(current_machine != NULL);
> +    assert(current_machine->accelerator != NULL);
> +    return allowed;
> +}
> +
>  static int accel_init_machine(AccelClass *acc, MachineState *ms)
>  {
>      ObjectClass *oc = OBJECT_CLASS(acc);
> diff --git a/include/qemu-common.h b/include/qemu-common.h
> index 85f4749aefb7..01d5e4d97dbf 100644
> --- a/include/qemu-common.h
> +++ b/include/qemu-common.h
> @@ -82,7 +82,8 @@ int qemu_openpty_raw(int *aslave, char *pty_name);
>  extern bool tcg_allowed;
>  void tcg_exec_init(unsigned long tb_size);
>  #ifdef CONFIG_TCG
> -#define tcg_enabled() (tcg_allowed)
> +#include "sysemu/accel.h"
> +#define tcg_enabled() (assert_accelerator_initialized(tcg_allowed))
>  #else
>  #define tcg_enabled() 0

It would be nice to catch mistakes even if
the CONFIG_{TCG,KVM,HAX,XEN} is disabled.  That would require making
assert_accelerator_initialized() a macro or inline function,
though.


>  #endif
> diff --git a/include/sysemu/accel.h b/include/sysemu/accel.h
> index 637358f43014..76965cb69cc9 100644
> --- a/include/sysemu/accel.h
> +++ b/include/sysemu/accel.h
> @@ -71,5 +71,6 @@ void configure_accelerator(MachineState *ms);
>  void accel_register_compat_props(AccelState *accel);
>  /* Called just before os_setup_post (ie just before drop OS privs) */
>  void accel_setup_post(MachineState *ms);
> +bool assert_accelerator_initialized(bool allowed);
>  
>  #endif
> diff --git a/include/sysemu/kvm.h b/include/sysemu/kvm.h
> index 0b64b8e06786..ac4dbb2d6d6d 100644
> --- a/include/sysemu/kvm.h
> +++ b/include/sysemu/kvm.h
> @@ -18,6 +18,7 @@
>  #include "qom/cpu.h"
>  #include "exec/memattrs.h"
>  #include "hw/irq.h"
> +#include "sysemu/accel.h"
>  
>  #ifdef NEED_CPU_H
>  # ifdef CONFIG_KVM
> @@ -46,7 +47,7 @@ extern bool kvm_direct_msi_allowed;
>  extern bool kvm_ioeventfd_any_length_allowed;
>  extern bool kvm_msi_use_devid;
>  
> -#define kvm_enabled()           (kvm_allowed)
> +#define kvm_enabled()           (assert_accelerator_initialized(kvm_allowed))
>  /**
>   * kvm_irqchip_in_kernel:
>   *
> diff --git a/qom/cpu.c b/qom/cpu.c
> index 92599f35413b..65a8f03a66a4 100644
> --- a/qom/cpu.c
> +++ b/qom/cpu.c
> @@ -23,6 +23,7 @@
>  #include "qemu-common.h"
>  #include "qom/cpu.h"
>  #include "sysemu/hw_accel.h"
> +#include "sysemu/accel.h"
>  #include "qemu/notify.h"
>  #include "qemu/log.h"
>  #include "exec/log.h"
> diff --git a/stubs/Makefile.objs b/stubs/Makefile.objs
> index 53d3f32cb258..2d5142287525 100644
> --- a/stubs/Makefile.objs
> +++ b/stubs/Makefile.objs
> @@ -43,3 +43,4 @@ stub-obj-y += xen-common.o
>  stub-obj-y += xen-hvm.o
>  stub-obj-y += pci-host-piix.o
>  stub-obj-y += ram-block.o
> +stub-obj-y += accel.o
> diff --git a/stubs/accel.c b/stubs/accel.c
> new file mode 100644
> index 000000000000..4f480f2d3f29
> --- /dev/null
> +++ b/stubs/accel.c
> @@ -0,0 +1,14 @@
> +/*
> + * accel stubs
> + *
> + * This work is licensed under the terms of the GNU GPL, version 2 or later.
> + * See the COPYING file in the top-level directory.
> + */
> +
> +#include "qemu/osdep.h"
> +#include "sysemu/accel.h"
> +
> +bool assert_accelerator_initialized(bool allowed)
> +{
> +    return allowed;
> +}
> diff --git a/target/i386/hax-all.c b/target/i386/hax-all.c
> index d2e512856bb8..7c78bd7d094d 100644
> --- a/target/i386/hax-all.c
> +++ b/target/i386/hax-all.c
> @@ -57,7 +57,7 @@ static int hax_arch_get_registers(CPUArchState *env);
>  
>  int hax_enabled(void)
>  {
> -    return hax_allowed;
> +    return assert_accelerator_initialized(hax_allowed);
>  }
>  
>  int valid_hax_tunnel_size(uint16_t size)
> diff --git a/target/i386/whpx-all.c b/target/i386/whpx-all.c
> index 6b42096698ee..e7f6bc5958e7 100644
> --- a/target/i386/whpx-all.c
> +++ b/target/i386/whpx-all.c
> @@ -1422,7 +1422,7 @@ static int whpx_accel_init(MachineState *ms)
>  
>  int whpx_enabled(void)
>  {
> -    return whpx_allowed;
> +    return assert_accelerator_initialized(whpx_allowed);
>  }
>  
>  static void whpx_accel_class_init(ObjectClass *oc, void *data)
> 

-- 
Eduardo

  parent reply	other threads:[~2018-06-29 19:58 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-28 10:14 [Qemu-devel] [PATCH 0/3] spapr: fix regression with older machine types Greg Kurz
2018-06-28 10:14 ` [Qemu-devel] [PATCH 1/3] target/ppc/kvm: don't pass cpu to kvm_get_smmu_info() Greg Kurz
2018-06-28 11:56   ` Cédric Le Goater
2018-06-28 12:14     ` Greg Kurz
2018-06-29  5:16   ` David Gibson
2018-06-28 10:15 ` [Qemu-devel] [PATCH 2/3] spapr: compute default value of "hpt-max-page-size" later Greg Kurz
2018-06-29  5:16   ` David Gibson
2018-06-29 19:08   ` Eduardo Habkost
2018-07-02  4:06     ` David Gibson
2018-06-28 10:15 ` [Qemu-devel] [PATCH 3/3] accel: forbid early use of kvm_enabled() and friends Greg Kurz
2018-06-29  5:18   ` David Gibson
2018-06-29 10:23     ` Greg Kurz
2018-06-29 19:58   ` Eduardo Habkost [this message]
2018-06-28 19:48 ` [Qemu-devel] [Qemu-ppc] [PATCH 0/3] spapr: fix regression with older machine types Greg Kurz
2018-06-29  5:21   ` David Gibson

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=20180629195823.GX7451@localhost.localdomain \
    --to=ehabkost@redhat.com \
    --cc=clg@kaod.org \
    --cc=david@gibson.dropbear.id.au \
    --cc=groug@kaod.org \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    --cc=rth@twiddle.net \
    /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.