All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sergey Fedorov <serge.fdrv@gmail.com>
To: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org
Cc: qemu-arm@nongnu.org, patches@linaro.org
Subject: Re: [Qemu-arm] [PATCH 03/11] target-arm: Raw CPSR writes should skip checks and bank switching
Date: Thu, 18 Feb 2016 20:42:29 +0300	[thread overview]
Message-ID: <56C60285.2080007@gmail.com> (raw)
In-Reply-To: <1455556977-3644-4-git-send-email-peter.maydell@linaro.org>

On 15.02.2016 20:22, Peter Maydell wrote:
> Raw CPSR writes should skip the architectural checks for whether
> we're allowed to set the A or F bits and should also not do
> the switching of register banks if the mode changes. Handle
> this inside cpsr_write(), which allows us to drop the "manually
> set the mode bits to avoid the bank switch" code from all the
> callsites which are using CPSRWriteRaw.
>
> This fixes a bug in 32-bit KVM handling where we had forgotten
> the "manually set the mode bits" part and could thus potentially
> trash the register state if the mode from the last exit to userspace
> differed from the mode on this exit.
>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Reviewed-by: Sergey Fedorov <serge.fdrv@gmail.com>

> ---
>  target-arm/helper.c    | 5 +++--
>  target-arm/kvm64.c     | 1 -
>  target-arm/machine.c   | 2 --
>  target-arm/op_helper.c | 5 ++++-
>  4 files changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/target-arm/helper.c b/target-arm/helper.c
> index 828822b..d1919bb 100644
> --- a/target-arm/helper.c
> +++ b/target-arm/helper.c
> @@ -5234,7 +5234,7 @@ void cpsr_write(CPUARMState *env, uint32_t val, uint32_t mask,
>       * In a V8 implementation, it is permitted for privileged software to
>       * change the CPSR A/F bits regardless of the SCR.AW/FW bits.
>       */
> -    if (!arm_feature(env, ARM_FEATURE_V8) &&
> +    if (write_type != CPSRWriteRaw && !arm_feature(env, ARM_FEATURE_V8) &&
>          arm_feature(env, ARM_FEATURE_EL3) &&
>          !arm_feature(env, ARM_FEATURE_EL2) &&
>          !arm_is_secure(env)) {
> @@ -5281,7 +5281,8 @@ void cpsr_write(CPUARMState *env, uint32_t val, uint32_t mask,
>      env->daif &= ~(CPSR_AIF & mask);
>      env->daif |= val & CPSR_AIF & mask;
>  
> -    if ((env->uncached_cpsr ^ val) & mask & CPSR_M) {
> +    if (write_type != CPSRWriteRaw &&
> +        ((env->uncached_cpsr ^ val) & mask & CPSR_M)) {
>          if (bad_mode_switch(env, val & CPSR_M)) {
>              /* Attempt to switch to an invalid mode: this is UNPREDICTABLE.
>               * We choose to ignore the attempt and leave the CPSR M field
> diff --git a/target-arm/kvm64.c b/target-arm/kvm64.c
> index 08c2c81..e8527bf 100644
> --- a/target-arm/kvm64.c
> +++ b/target-arm/kvm64.c
> @@ -722,7 +722,6 @@ int kvm_arch_get_registers(CPUState *cs)
>      if (is_a64(env)) {
>          pstate_write(env, val);
>      } else {
> -        env->uncached_cpsr = val & CPSR_M;
>          cpsr_write(env, val, 0xffffffff, CPSRWriteRaw);
>      }
>  
> diff --git a/target-arm/machine.c b/target-arm/machine.c
> index 0fc7df0..03a73d9 100644
> --- a/target-arm/machine.c
> +++ b/target-arm/machine.c
> @@ -173,8 +173,6 @@ static int get_cpsr(QEMUFile *f, void *opaque, size_t size)
>          return 0;
>      }
>  
> -    /* Avoid mode switch when restoring CPSR */
> -    env->uncached_cpsr = val & CPSR_M;
>      cpsr_write(env, val, 0xffffffff, CPSRWriteRaw);
>      return 0;
>  }
> diff --git a/target-arm/op_helper.c b/target-arm/op_helper.c
> index 543d33a..4881e34 100644
> --- a/target-arm/op_helper.c
> +++ b/target-arm/op_helper.c
> @@ -779,7 +779,10 @@ void HELPER(exception_return)(CPUARMState *env)
>  
>      if (!return_to_aa64) {
>          env->aarch64 = 0;
> -        env->uncached_cpsr = spsr & CPSR_M;
> +        /* We do a raw CPSR write because aarch64_sync_64_to_32()
> +         * will sort the register banks out for us, and we've already
> +         * caught all the bad-mode cases in el_from_spsr().
> +         */
>          cpsr_write(env, spsr, ~0, CPSRWriteRaw);
>          if (!arm_singlestep_active(env)) {
>              env->uncached_cpsr &= ~PSTATE_SS;


WARNING: multiple messages have this Message-ID (diff)
From: Sergey Fedorov <serge.fdrv@gmail.com>
To: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org
Cc: "Edgar E. Iglesias" <edgar.iglesias@gmail.com>,
	qemu-arm@nongnu.org, patches@linaro.org
Subject: Re: [Qemu-devel] [PATCH 03/11] target-arm: Raw CPSR writes should skip checks and bank switching
Date: Thu, 18 Feb 2016 20:42:29 +0300	[thread overview]
Message-ID: <56C60285.2080007@gmail.com> (raw)
In-Reply-To: <1455556977-3644-4-git-send-email-peter.maydell@linaro.org>

On 15.02.2016 20:22, Peter Maydell wrote:
> Raw CPSR writes should skip the architectural checks for whether
> we're allowed to set the A or F bits and should also not do
> the switching of register banks if the mode changes. Handle
> this inside cpsr_write(), which allows us to drop the "manually
> set the mode bits to avoid the bank switch" code from all the
> callsites which are using CPSRWriteRaw.
>
> This fixes a bug in 32-bit KVM handling where we had forgotten
> the "manually set the mode bits" part and could thus potentially
> trash the register state if the mode from the last exit to userspace
> differed from the mode on this exit.
>
> Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Reviewed-by: Sergey Fedorov <serge.fdrv@gmail.com>

> ---
>  target-arm/helper.c    | 5 +++--
>  target-arm/kvm64.c     | 1 -
>  target-arm/machine.c   | 2 --
>  target-arm/op_helper.c | 5 ++++-
>  4 files changed, 7 insertions(+), 6 deletions(-)
>
> diff --git a/target-arm/helper.c b/target-arm/helper.c
> index 828822b..d1919bb 100644
> --- a/target-arm/helper.c
> +++ b/target-arm/helper.c
> @@ -5234,7 +5234,7 @@ void cpsr_write(CPUARMState *env, uint32_t val, uint32_t mask,
>       * In a V8 implementation, it is permitted for privileged software to
>       * change the CPSR A/F bits regardless of the SCR.AW/FW bits.
>       */
> -    if (!arm_feature(env, ARM_FEATURE_V8) &&
> +    if (write_type != CPSRWriteRaw && !arm_feature(env, ARM_FEATURE_V8) &&
>          arm_feature(env, ARM_FEATURE_EL3) &&
>          !arm_feature(env, ARM_FEATURE_EL2) &&
>          !arm_is_secure(env)) {
> @@ -5281,7 +5281,8 @@ void cpsr_write(CPUARMState *env, uint32_t val, uint32_t mask,
>      env->daif &= ~(CPSR_AIF & mask);
>      env->daif |= val & CPSR_AIF & mask;
>  
> -    if ((env->uncached_cpsr ^ val) & mask & CPSR_M) {
> +    if (write_type != CPSRWriteRaw &&
> +        ((env->uncached_cpsr ^ val) & mask & CPSR_M)) {
>          if (bad_mode_switch(env, val & CPSR_M)) {
>              /* Attempt to switch to an invalid mode: this is UNPREDICTABLE.
>               * We choose to ignore the attempt and leave the CPSR M field
> diff --git a/target-arm/kvm64.c b/target-arm/kvm64.c
> index 08c2c81..e8527bf 100644
> --- a/target-arm/kvm64.c
> +++ b/target-arm/kvm64.c
> @@ -722,7 +722,6 @@ int kvm_arch_get_registers(CPUState *cs)
>      if (is_a64(env)) {
>          pstate_write(env, val);
>      } else {
> -        env->uncached_cpsr = val & CPSR_M;
>          cpsr_write(env, val, 0xffffffff, CPSRWriteRaw);
>      }
>  
> diff --git a/target-arm/machine.c b/target-arm/machine.c
> index 0fc7df0..03a73d9 100644
> --- a/target-arm/machine.c
> +++ b/target-arm/machine.c
> @@ -173,8 +173,6 @@ static int get_cpsr(QEMUFile *f, void *opaque, size_t size)
>          return 0;
>      }
>  
> -    /* Avoid mode switch when restoring CPSR */
> -    env->uncached_cpsr = val & CPSR_M;
>      cpsr_write(env, val, 0xffffffff, CPSRWriteRaw);
>      return 0;
>  }
> diff --git a/target-arm/op_helper.c b/target-arm/op_helper.c
> index 543d33a..4881e34 100644
> --- a/target-arm/op_helper.c
> +++ b/target-arm/op_helper.c
> @@ -779,7 +779,10 @@ void HELPER(exception_return)(CPUARMState *env)
>  
>      if (!return_to_aa64) {
>          env->aarch64 = 0;
> -        env->uncached_cpsr = spsr & CPSR_M;
> +        /* We do a raw CPSR write because aarch64_sync_64_to_32()
> +         * will sort the register banks out for us, and we've already
> +         * caught all the bad-mode cases in el_from_spsr().
> +         */
>          cpsr_write(env, spsr, ~0, CPSRWriteRaw);
>          if (!arm_singlestep_active(env)) {
>              env->uncached_cpsr &= ~PSTATE_SS;

  reply	other threads:[~2016-02-18 17:43 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-15 17:22 [Qemu-devel] [PATCH 00/11] target-arm: clean up cpsr_write mode changing Peter Maydell
2016-02-15 17:22 ` [Qemu-arm] [PATCH 01/11] target-arm: Give CPSR setting on 32-bit exception return its own helper Peter Maydell
2016-02-15 17:22   ` [Qemu-devel] " Peter Maydell
2016-02-18 17:41   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:41     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-arm] [PATCH 02/11] target-arm: Add write_type argument to cpsr_write() Peter Maydell
2016-02-15 17:22   ` [Qemu-devel] " Peter Maydell
2016-02-18 17:42   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:42     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 03/11] target-arm: Raw CPSR writes should skip checks and bank switching Peter Maydell
2016-02-18 17:42   ` Sergey Fedorov [this message]
2016-02-18 17:42     ` Sergey Fedorov
2016-02-15 17:22 ` [Qemu-arm] [PATCH 04/11] linux-user: Use restrictive mask when calling cpsr_write() Peter Maydell
2016-02-15 17:22   ` [Qemu-devel] " Peter Maydell
2016-02-18 17:42   ` Sergey Fedorov
2016-02-15 17:22 ` [Qemu-arm] [PATCH 05/11] target-arm: In cpsr_write() ignore mode switches from User mode Peter Maydell
2016-02-15 17:22   ` [Qemu-devel] " Peter Maydell
2016-02-18 17:43   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:43     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 06/11] target-arm: Add comment about not implementing NSACR.RFR Peter Maydell
2016-02-18 17:43   ` Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 07/11] target-arm: Add Hyp mode checks to bad_mode_switch() Peter Maydell
2016-02-18 17:43   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:43     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 08/11] target-arm: Forbid mode switch to Mon from Secure EL1 Peter Maydell
2016-02-18 17:43   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:43     ` Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 09/11] target-arm: In v8, make illegal AArch32 mode changes set PSTATE.IL Peter Maydell
2016-02-18 17:44   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:44     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 10/11] target-arm: Make mode switches from Hyp via CPS and MRS illegal Peter Maydell
2016-02-18 17:44   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:44     ` [Qemu-devel] " Sergey Fedorov
2016-02-15 17:22 ` [Qemu-devel] [PATCH 11/11] target-arm: Make Monitor->NS PL1 mode changes illegal if HCR.TGE is 1 Peter Maydell
2016-02-18 17:44   ` [Qemu-arm] " Sergey Fedorov
2016-02-18 17:44     ` [Qemu-devel] " Sergey Fedorov

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=56C60285.2080007@gmail.com \
    --to=serge.fdrv@gmail.com \
    --cc=patches@linaro.org \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --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 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.