[7/7,v3] KVM: PPC: Add userspace debug stub support

Message ID	1365417138-8202-8-git-send-email-Bharat.Bhushan@freescale.com (mailing list archive)
State	New, archived
Headers	show Return-Path: <kvm-owner@vger.kernel.org> From: Bharat Bhushan <r65777@freescale.com> To: <kvm-ppc@vger.kernel.org>, <kvm@vger.kernel.org>, <agraf@suse.de>, <scottwood@freescale.com> CC: Bharat Bhushan <bharat.bhushan@freescale.com> Subject: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support Date: Mon, 8 Apr 2013 16:02:18 +0530 Message-ID: <1365417138-8202-8-git-send-email-Bharat.Bhushan@freescale.com> In-Reply-To: <1365417138-8202-7-git-send-email-Bharat.Bhushan@freescale.com> References: <1365417138-8202-1-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-2-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-3-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-4-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-5-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-6-git-send-email-Bharat.Bhushan@freescale.com> <1365417138-8202-7-git-send-email-Bharat.Bhushan@freescale.com> MIME-Version: 1.0 Content-Type: text/plain Sender: kvm-owner@vger.kernel.org Precedence: bulk

Bharat Bhushan April 8, 2013, 10:32 a.m. UTC

From: Bharat Bhushan <bharat.bhushan@freescale.com>

This patch adds the debug stub support on booke/bookehv.
Now QEMU debug stub can use hw breakpoint, watchpoint and software
breakpoint to debug guest.

Debug registers are saved/restored on vcpu_put()/vcpu_get().
Also the debug registers are saved restored only if guest
is using debug resources.

Currently we do not support debug resource emulation to guest,
so always exit to user space irrespective of user space is expecting
the debug exception or not. This is unexpected event and let us
leave the action on user space. This is similar to what it was before,
only thing is that now we have proper exit state available to user space.

Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
---
 arch/powerpc/include/asm/kvm_host.h |    8 +
 arch/powerpc/include/uapi/asm/kvm.h |   22 +++-
 arch/powerpc/kvm/booke.c            |  242 ++++++++++++++++++++++++++++++++---
 arch/powerpc/kvm/booke.h            |    5 +
 4 files changed, 255 insertions(+), 22 deletions(-)

Alexander Graf April 26, 2013, 11:15 a.m. UTC | #1

On 08.04.2013, at 12:32, Bharat Bhushan wrote:

> From: Bharat Bhushan <bharat.bhushan@freescale.com>
> 
> This patch adds the debug stub support on booke/bookehv.
> Now QEMU debug stub can use hw breakpoint, watchpoint and software
> breakpoint to debug guest.
> 
> Debug registers are saved/restored on vcpu_put()/vcpu_get().
> Also the debug registers are saved restored only if guest
> is using debug resources.
> 
> Currently we do not support debug resource emulation to guest,
> so always exit to user space irrespective of user space is expecting
> the debug exception or not. This is unexpected event and let us
> leave the action on user space. This is similar to what it was before,
> only thing is that now we have proper exit state available to user space.
> 
> Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
> ---
> arch/powerpc/include/asm/kvm_host.h |    8 +
> arch/powerpc/include/uapi/asm/kvm.h |   22 +++-
> arch/powerpc/kvm/booke.c            |  242 ++++++++++++++++++++++++++++++++---
> arch/powerpc/kvm/booke.h            |    5 +
> 4 files changed, 255 insertions(+), 22 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h
> index e34f8fe..b9ad20f 100644
> --- a/arch/powerpc/include/asm/kvm_host.h
> +++ b/arch/powerpc/include/asm/kvm_host.h
> @@ -505,7 +505,15 @@ struct kvm_vcpu_arch {
> 	u32 mmucfg;
> 	u32 epr;
> 	u32 crit_save;
> +
> +	/* Flag indicating that debug registers are used by guest */
> +	bool debug_active;
> +	/* for save/restore thread->dbcr0 on vcpu run/heavyweight_exit */
> +	u32 saved_dbcr0;
> +	/* guest debug registers*/
> 	struct kvmppc_booke_debug_reg dbg_reg;
> +	/* shadow debug registers */
> +	struct kvmppc_booke_debug_reg shadow_dbg_reg;
> #endif
> 	gpa_t paddr_accessed;
> 	gva_t vaddr_accessed;
> diff --git a/arch/powerpc/include/uapi/asm/kvm.h b/arch/powerpc/include/uapi/asm/kvm.h
> index c0c38ed..d7ce449 100644
> --- a/arch/powerpc/include/uapi/asm/kvm.h
> +++ b/arch/powerpc/include/uapi/asm/kvm.h
> @@ -25,6 +25,7 @@
> /* Select powerpc specific features in <linux/kvm.h> */
> #define __KVM_HAVE_SPAPR_TCE
> #define __KVM_HAVE_PPC_SMT
> +#define __KVM_HAVE_GUEST_DEBUG
> 
> struct kvm_regs {
> 	__u64 pc;
> @@ -267,7 +268,24 @@ struct kvm_fpu {
> 	__u64 fpr[32];
> };
> 
> +/*
> + * Defines for h/w breakpoint, watchpoint (read, write or both) and
> + * software breakpoint.
> + * These are used as "type" in KVM_SET_GUEST_DEBUG ioctl and "status"
> + * for KVM_DEBUG_EXIT.
> + */
> +#define KVMPPC_DEBUG_NONE		0x0
> +#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> +#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> +#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> struct kvm_debug_exit_arch {
> +	__u64 address;
> +	/*
> +	 * exiting to userspace because of h/w breakpoint, watchpoint
> +	 * (read, write or both) and software breakpoint.
> +	 */
> +	__u32 status;
> +	__u32 reserved;
> };
> 
> /* for KVM_SET_GUEST_DEBUG */
> @@ -279,10 +297,6 @@ struct kvm_guest_debug_arch {
> 		 * Type denotes h/w breakpoint, read watchpoint, write
> 		 * watchpoint or watchpoint (both read and write).
> 		 */
> -#define KVMPPC_DEBUG_NONE		0x0
> -#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> -#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> -#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> 		__u32 type;
> 		__u32 reserved;
> 	} bp[16];
> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c
> index 97ae158..0e93416 100644
> --- a/arch/powerpc/kvm/booke.c
> +++ b/arch/powerpc/kvm/booke.c
> @@ -133,6 +133,29 @@ static void kvmppc_vcpu_sync_fpu(struct kvm_vcpu *vcpu)
> #endif
> }
> 
> +static void kvmppc_vcpu_sync_debug(struct kvm_vcpu *vcpu)
> +{
> +	/* Synchronize guest's desire to get debug interrupts into shadow MSR */
> +#ifndef CONFIG_KVM_BOOKE_HV
> +	vcpu->arch.shadow_msr &= ~MSR_DE;
> +	vcpu->arch.shadow_msr |= vcpu->arch.shared->msr & MSR_DE;
> +#endif
> +
> +	/* Force enable debug interrupts when user space wants to debug */
> +	if (vcpu->guest_debug) {
> +#ifdef CONFIG_KVM_BOOKE_HV
> +		/*
> +		 * Since there is no shadow MSR, sync MSR_DE into the guest
> +		 * visible MSR.
> +		 */
> +		vcpu->arch.shared->msr |= MSR_DE;
> +#else
> +		vcpu->arch.shadow_msr |= MSR_DE;
> +		vcpu->arch.shared->msr &= ~MSR_DE;
> +#endif
> +	}
> +}
> +
> /*
>  * Helper function for "full" MSR writes.  No need to call this if only
>  * EE/CE/ME/DE/RI are changing.
> @@ -150,6 +173,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u32 new_msr)
> 	kvmppc_mmu_msr_notify(vcpu, old_msr);
> 	kvmppc_vcpu_sync_spe(vcpu);
> 	kvmppc_vcpu_sync_fpu(vcpu);
> +	kvmppc_vcpu_sync_debug(vcpu);
> }
> 
> static void kvmppc_booke_queue_irqprio(struct kvm_vcpu *vcpu,
> @@ -646,6 +670,46 @@ int kvmppc_core_check_requests(struct kvm_vcpu *vcpu)
> 	return r;
> }
> 
> +static void kvmppc_load_usespace_gebug(void)

...

> +{
> +	switch_booke_debug_regs(&current->thread);
> +}
> +
> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu *vcpu)
> +{
> +	if (!vcpu->arch.debug_active)
> +		return;
> +
> +	/* Disable all debug events and clead pending debug events */
> +	mtspr(SPRN_DBCR0, 0x0);
> +	kvmppc_clear_dbsr();
> +
> +	/*
> +	 * Check whether guest still need debug resource, if not then there
> +	 * is no need to restore guest context.
> +	 */
> +	if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> +		return;
> +
> +	/* Load Guest Context */
> +	mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> +	mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2);
> +#ifdef CONFIG_KVM_E500MC
> +	mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);

You need to make sure DBCR4 is 0 when you leave things back to normal user space. Otherwise guest debug can interfere with host debug.

> +#endif
> +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> +#endif
> +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> +
> +	/* Enable debug events after other debug registers restored */
> +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0);
> +}

All of the code above looks suspiciously similar to prime_debug_regs();. Can't we somehow reuse that?

> +
> int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
> {
> 	int ret, s;
> @@ -693,11 +757,25 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
> 	kvmppc_load_guest_fp(vcpu);
> #endif
> 
> +	/*
> +	 * Clear current->thread.dbcr0 so that kernel does not
> +	 * restore h/w registers on context switch in vcpu running state.
> +	 */
> +	vcpu->arch.debug_active = 1;

= true;

> +	vcpu->arch.saved_dbcr0 = current->thread.dbcr0;
> +	current->thread.dbcr0 = 0;
> +	kvmppc_booke_vcpu_load_debug_regs(vcpu);

static void switch_booke_debug_regs(struct thread_struct *new_thread)
{
        if ((current->thread.dbcr0 & DBCR0_IDM)
                || (new_thread->dbcr0 & DBCR0_IDM))
                        prime_debug_regs(new_thread);
}

The kernel will also restore debug state if the process we come from has debugging enabled. Please adjust the comment accordingly.

> +
> 	ret = __kvmppc_vcpu_run(kvm_run, vcpu);
> 
> 	/* No need for kvm_guest_exit. It's done in handle_exit.
> 	   We also get here with interrupts enabled. */
> 
> +	/* Restore thread->dbcr0 */
> +	vcpu->arch.debug_active = 0;
> +	current->thread.dbcr0 = vcpu->arch.saved_dbcr0;
> +	kvmppc_load_usespace_gebug();
> +
> #ifdef CONFIG_PPC_FPU
> 	kvmppc_save_guest_fp(vcpu);
> 
> @@ -753,6 +831,36 @@ static int emulation_exit(struct kvm_run *run, struct kvm_vcpu *vcpu)
> 	}
> }
> 
> +/*
> + * Currently we do not support debug resource emulation to guest,
> + * so always exit to user space irrespective of user space is
> + * expecting the debug exception or not. This is unexpected event
> + * and let us leave the action on user space.
> + */
> +static int kvmppc_handle_debug(struct kvm_run *run, struct kvm_vcpu *vcpu)
> +{
> +	u32 dbsr = mfspr(SPRN_DBSR);
> +
> +	kvmppc_clear_dbsr();
> +	run->debug.arch.status = 0;
> +	run->debug.arch.address = vcpu->arch.pc;
> +
> +	if (dbsr & (DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4)) {
> +		run->debug.arch.status |= KVMPPC_DEBUG_BREAKPOINT;
> +	} else {
> +		if (dbsr & (DBSR_DAC1W | DBSR_DAC2W))
> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_WRITE;
> +		else if (dbsr & (DBSR_DAC1R | DBSR_DAC2R))
> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_READ;
> +		if (dbsr & (DBSR_DAC1R | DBSR_DAC1W))
> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[0];
> +		else if (dbsr & (DBSR_DAC2R | DBSR_DAC2W))
> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[1];
> +	}
> +
> +	return RESUME_HOST;
> +}
> +
> static void kvmppc_fill_pt_regs(struct pt_regs *regs)
> {
> 	ulong r1, ip, msr, lr;
> @@ -1112,18 +1220,10 @@ int kvmppc_handle_exit(struct kvm_run *run, struct kvm_vcpu *vcpu,
> 	}
> 
> 	case BOOKE_INTERRUPT_DEBUG: {
> -		u32 dbsr;
> -
> -		vcpu->arch.pc = mfspr(SPRN_CSRR0);
> -
> -		/* clear IAC events in DBSR register */
> -		dbsr = mfspr(SPRN_DBSR);
> -		dbsr &= DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4;
> -		mtspr(SPRN_DBSR, dbsr);
> -
> -		run->exit_reason = KVM_EXIT_DEBUG;
> +		r = kvmppc_handle_debug(run, vcpu);
> +		if (r == RESUME_HOST)
> +			run->exit_reason = KVM_EXIT_DEBUG;
> 		kvmppc_account_exit(vcpu, DEBUG_EXITS);
> -		r = RESUME_HOST;
> 		break;
> 	}
> 
> @@ -1174,7 +1274,7 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
> 	kvmppc_set_msr(vcpu, 0);
> 
> #ifndef CONFIG_KVM_BOOKE_HV
> -	vcpu->arch.shadow_msr = MSR_USER | MSR_DE | MSR_IS | MSR_DS;
> +	vcpu->arch.shadow_msr = MSR_USER | MSR_IS | MSR_DS;
> 	vcpu->arch.shadow_pid = 1;
> 	vcpu->arch.shared->msr = 0;
> #endif
> @@ -1529,12 +1629,6 @@ int kvm_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu, struct kvm_one_reg *reg)
> 	return r;
> }
> 
> -int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> -					 struct kvm_guest_debug *dbg)
> -{
> -	return -EINVAL;
> -}
> -
> int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
> {
> 	return -ENOTSUPP;
> @@ -1640,16 +1734,128 @@ void kvmppc_decrementer_func(unsigned long data)
> 	kvmppc_set_tsr_bits(vcpu, TSR_DIS);
> }
> 
> +static void kvmppc_booke_vcpu_put_debug_regs(struct kvm_vcpu *vcpu)
> +{
> +	/* Disable all debug events First */

first

> +	mtspr(SPRN_DBCR0, 0x0);
> +	/* Disable pending debug event by clearing DBSR */
> +	kvmppc_clear_dbsr();

kvmppc_handle_debug() happens with preemption enabled, no? So we can have a debug event that gets cleared on preempt by this.

> +}
> +
> +int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> +					 struct kvm_guest_debug *dbg)
> +{
> +	struct kvmppc_booke_debug_reg *dbg_reg;
> +	int n, b = 0, w = 0;
> +	const u32 bp_code[] = {
> +		DBCR0_IAC1 | DBCR0_IDM,
> +		DBCR0_IAC2 | DBCR0_IDM,
> +		DBCR0_IAC3 | DBCR0_IDM,
> +		DBCR0_IAC4 | DBCR0_IDM
> +	};
> +	const u32 wp_code[] = {
> +		DBCR0_DAC1W | DBCR0_IDM,
> +		DBCR0_DAC2W | DBCR0_IDM,
> +		DBCR0_DAC1R | DBCR0_IDM,
> +		DBCR0_DAC2R | DBCR0_IDM
> +	};
> +
> +	if (!(dbg->control & KVM_GUESTDBG_ENABLE)) {
> +		/* Clear All debug events */
> +		vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> +		vcpu->guest_debug = 0;
> +#ifdef CONFIG_KVM_BOOKE_HV
> +		/*
> +		 * When user space is not using the debug resources
> +		 * then allow guest to change the MSR.DE.
> +		 */
> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
> +#endif
> +		return 0;
> +	}
> +
> +#ifdef CONFIG_KVM_BOOKE_HV
> +		/*
> +		 * When user space is using the debug resource then
> +		 * do not allow guest to change the MSR.DE.
> +		 */
> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;

This is supposed to be |= right?

> +#endif
> +	vcpu->guest_debug = dbg->control;
> +	vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> +	/* Set DBCR0_EDM in guest visible DBCR0 register. */
> +	vcpu->arch.dbg_reg.dbcr0 = DBCR0_EDM;
> +
> +	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
> +		vcpu->arch.shadow_dbg_reg.dbcr0 |= DBCR0_IDM | DBCR0_IC;
> +
> +	if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
> +		/* Code below handles only HW breakpoints */

Please move the comment out of the one-lined branch. Just put it below the return.

> +		return 0;
> +
> +	dbg_reg = &(vcpu->arch.shadow_dbg_reg);
> +
> +	/*
> +	 * On BOOKE (e500v2); Set DBCR1 and DBCR2 to allow debug events
> +	 * to occur when MSR.PR is set.
> +	 * On BOOKE-HV (e500mc+); MSR.PR = 0 when guest is running. So we
> +	 * should clear DBCR1 and DBCR2. And EPCR.DUVD is used to control
> +	 * that debug events will not come in hypervisor (GS = 0).
> +	 */
> +#ifdef CONFIG_KVM_BOOKE_HV
> +	dbg_reg->dbcr1 = 0;
> +	dbg_reg->dbcr2 = 0;
> +#else
> +	dbg_reg->dbcr1 = DBCR1_IAC1US | DBCR1_IAC2US | DBCR1_IAC3US |
> +			  DBCR1_IAC4US;
> +	dbg_reg->dbcr2 = DBCR2_DAC1US | DBCR2_DAC2US;
> +#endif
> +
> +	for (n = 0; n < (KVMPPC_BOOKE_IAC_NUM + KVMPPC_BOOKE_DAC_NUM); n++) {
> +		u32 type = dbg->arch.bp[n].type;
> +
> +		if (!type)
> +			continue;

Scott's comment on why a zero-type is different from any other invalid type is still outstanding I think.


Alex

> +
> +		if (type & (KVMPPC_DEBUG_WATCH_READ |
> +			    KVMPPC_DEBUG_WATCH_WRITE)) {
> +			if (w >= KVMPPC_BOOKE_DAC_NUM)
> +				return -EINVAL;
> +
> +			if (type & KVMPPC_DEBUG_WATCH_READ)
> +				dbg_reg->dbcr0 |= wp_code[w + 2];
> +			if (type & KVMPPC_DEBUG_WATCH_WRITE)
> +				dbg_reg->dbcr0 |= wp_code[w];
> +
> +			dbg_reg->dac[w] = dbg->arch.bp[n].addr;
> +			w++;
> +		} else if (type & KVMPPC_DEBUG_BREAKPOINT) {
> +			if (b >= KVMPPC_BOOKE_IAC_NUM)
> +				return -EINVAL;
> +
> +			dbg_reg->dbcr0 |= bp_code[b];
> +			dbg_reg->iac[b] = dbg->arch.bp[n].addr;
> +			b++;
> +		}
> +	}
> +
> +	return 0;
> +}
> +
> void kvmppc_booke_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> {
> 	vcpu->cpu = smp_processor_id();
> 	current->thread.kvm_vcpu = vcpu;
> +
> +	kvmppc_booke_vcpu_load_debug_regs(vcpu);
> }
> 
> void kvmppc_booke_vcpu_put(struct kvm_vcpu *vcpu)
> {
> 	current->thread.kvm_vcpu = NULL;
> 	vcpu->cpu = -1;
> +
> +	kvmppc_booke_vcpu_put_debug_regs(vcpu);
> }
> 
> int __init kvmppc_booke_init(void)
> diff --git a/arch/powerpc/kvm/booke.h b/arch/powerpc/kvm/booke.h
> index 5fd1ba6..a1ff67d 100644
> --- a/arch/powerpc/kvm/booke.h
> +++ b/arch/powerpc/kvm/booke.h
> @@ -129,4 +129,9 @@ static inline void kvmppc_save_guest_fp(struct kvm_vcpu *vcpu)
> 		giveup_fpu(current);
> #endif
> }
> +
> +static inline void kvmppc_clear_dbsr(void)
> +{
> +	mtspr(SPRN_DBSR, mfspr(SPRN_DBSR));
> +}
> #endif /* __KVM_BOOKE_H__ */
> -- 
> 1.7.0.4
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Bharat Bhushan May 2, 2013, 9:46 a.m. UTC | #2

> -----Original Message-----
> From: Alexander Graf [mailto:agraf@suse.de]
> Sent: Friday, April 26, 2013 4:46 PM
> To: Bhushan Bharat-R65777
> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421; Bhushan
> Bharat-R65777
> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
> 
> 
> On 08.04.2013, at 12:32, Bharat Bhushan wrote:
> 
> > From: Bharat Bhushan <bharat.bhushan@freescale.com>
> >
> > This patch adds the debug stub support on booke/bookehv.
> > Now QEMU debug stub can use hw breakpoint, watchpoint and software
> > breakpoint to debug guest.
> >
> > Debug registers are saved/restored on vcpu_put()/vcpu_get().
> > Also the debug registers are saved restored only if guest
> > is using debug resources.
> >
> > Currently we do not support debug resource emulation to guest,
> > so always exit to user space irrespective of user space is expecting
> > the debug exception or not. This is unexpected event and let us
> > leave the action on user space. This is similar to what it was before,
> > only thing is that now we have proper exit state available to user space.
> >
> > Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
> > ---
> > arch/powerpc/include/asm/kvm_host.h |    8 +
> > arch/powerpc/include/uapi/asm/kvm.h |   22 +++-
> > arch/powerpc/kvm/booke.c            |  242 ++++++++++++++++++++++++++++++++---
> > arch/powerpc/kvm/booke.h            |    5 +
> > 4 files changed, 255 insertions(+), 22 deletions(-)
> >
> > diff --git a/arch/powerpc/include/asm/kvm_host.h
> b/arch/powerpc/include/asm/kvm_host.h
> > index e34f8fe..b9ad20f 100644
> > --- a/arch/powerpc/include/asm/kvm_host.h
> > +++ b/arch/powerpc/include/asm/kvm_host.h
> > @@ -505,7 +505,15 @@ struct kvm_vcpu_arch {
> > 	u32 mmucfg;
> > 	u32 epr;
> > 	u32 crit_save;
> > +
> > +	/* Flag indicating that debug registers are used by guest */
> > +	bool debug_active;
> > +	/* for save/restore thread->dbcr0 on vcpu run/heavyweight_exit */
> > +	u32 saved_dbcr0;
> > +	/* guest debug registers*/
> > 	struct kvmppc_booke_debug_reg dbg_reg;
> > +	/* shadow debug registers */
> > +	struct kvmppc_booke_debug_reg shadow_dbg_reg;
> > #endif
> > 	gpa_t paddr_accessed;
> > 	gva_t vaddr_accessed;
> > diff --git a/arch/powerpc/include/uapi/asm/kvm.h
> b/arch/powerpc/include/uapi/asm/kvm.h
> > index c0c38ed..d7ce449 100644
> > --- a/arch/powerpc/include/uapi/asm/kvm.h
> > +++ b/arch/powerpc/include/uapi/asm/kvm.h
> > @@ -25,6 +25,7 @@
> > /* Select powerpc specific features in <linux/kvm.h> */
> > #define __KVM_HAVE_SPAPR_TCE
> > #define __KVM_HAVE_PPC_SMT
> > +#define __KVM_HAVE_GUEST_DEBUG
> >
> > struct kvm_regs {
> > 	__u64 pc;
> > @@ -267,7 +268,24 @@ struct kvm_fpu {
> > 	__u64 fpr[32];
> > };
> >
> > +/*
> > + * Defines for h/w breakpoint, watchpoint (read, write or both) and
> > + * software breakpoint.
> > + * These are used as "type" in KVM_SET_GUEST_DEBUG ioctl and "status"
> > + * for KVM_DEBUG_EXIT.
> > + */
> > +#define KVMPPC_DEBUG_NONE		0x0
> > +#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> > +#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> > +#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> > struct kvm_debug_exit_arch {
> > +	__u64 address;
> > +	/*
> > +	 * exiting to userspace because of h/w breakpoint, watchpoint
> > +	 * (read, write or both) and software breakpoint.
> > +	 */
> > +	__u32 status;
> > +	__u32 reserved;
> > };
> >
> > /* for KVM_SET_GUEST_DEBUG */
> > @@ -279,10 +297,6 @@ struct kvm_guest_debug_arch {
> > 		 * Type denotes h/w breakpoint, read watchpoint, write
> > 		 * watchpoint or watchpoint (both read and write).
> > 		 */
> > -#define KVMPPC_DEBUG_NONE		0x0
> > -#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> > -#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> > -#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> > 		__u32 type;
> > 		__u32 reserved;
> > 	} bp[16];
> > diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c
> > index 97ae158..0e93416 100644
> > --- a/arch/powerpc/kvm/booke.c
> > +++ b/arch/powerpc/kvm/booke.c
> > @@ -133,6 +133,29 @@ static void kvmppc_vcpu_sync_fpu(struct kvm_vcpu *vcpu)
> > #endif
> > }
> >
> > +static void kvmppc_vcpu_sync_debug(struct kvm_vcpu *vcpu)
> > +{
> > +	/* Synchronize guest's desire to get debug interrupts into shadow MSR */
> > +#ifndef CONFIG_KVM_BOOKE_HV
> > +	vcpu->arch.shadow_msr &= ~MSR_DE;
> > +	vcpu->arch.shadow_msr |= vcpu->arch.shared->msr & MSR_DE;
> > +#endif
> > +
> > +	/* Force enable debug interrupts when user space wants to debug */
> > +	if (vcpu->guest_debug) {
> > +#ifdef CONFIG_KVM_BOOKE_HV
> > +		/*
> > +		 * Since there is no shadow MSR, sync MSR_DE into the guest
> > +		 * visible MSR.
> > +		 */
> > +		vcpu->arch.shared->msr |= MSR_DE;
> > +#else
> > +		vcpu->arch.shadow_msr |= MSR_DE;
> > +		vcpu->arch.shared->msr &= ~MSR_DE;
> > +#endif
> > +	}
> > +}
> > +
> > /*
> >  * Helper function for "full" MSR writes.  No need to call this if only
> >  * EE/CE/ME/DE/RI are changing.
> > @@ -150,6 +173,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u32 new_msr)
> > 	kvmppc_mmu_msr_notify(vcpu, old_msr);
> > 	kvmppc_vcpu_sync_spe(vcpu);
> > 	kvmppc_vcpu_sync_fpu(vcpu);
> > +	kvmppc_vcpu_sync_debug(vcpu);
> > }
> >
> > static void kvmppc_booke_queue_irqprio(struct kvm_vcpu *vcpu,
> > @@ -646,6 +670,46 @@ int kvmppc_core_check_requests(struct kvm_vcpu *vcpu)
> > 	return r;
> > }
> >
> > +static void kvmppc_load_usespace_gebug(void)
> 
> ...

Please tell What does "..."  mean. Does that mean next comment is about the above function?

> 
> > +{
> > +	switch_booke_debug_regs(&current->thread);
> > +}
> > +
> > +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu *vcpu)
> > +{
> > +	if (!vcpu->arch.debug_active)
> > +		return;
> > +
> > +	/* Disable all debug events and clead pending debug events */
> > +	mtspr(SPRN_DBCR0, 0x0);
> > +	kvmppc_clear_dbsr();
> > +
> > +	/*
> > +	 * Check whether guest still need debug resource, if not then there
> > +	 * is no need to restore guest context.
> > +	 */
> > +	if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> > +		return;
> > +
> > +	/* Load Guest Context */
> > +	mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> > +	mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2);
> > +#ifdef CONFIG_KVM_E500MC
> > +	mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
> 
> You need to make sure DBCR4 is 0 when you leave things back to normal user
> space. Otherwise guest debug can interfere with host debug.


ok

> 
> > +#endif
> > +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> > +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> > +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> > +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> > +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> > +#endif
> > +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> > +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> > +
> > +	/* Enable debug events after other debug registers restored */
> > +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0);
> > +}
> 
> All of the code above looks suspiciously similar to prime_debug_regs();. Can't
> we somehow reuse that?

I think we can if
 - Save thread->debug_regs in local data structure 
 - Load vcpu->arch->debug_regs in thread->debug_regs
 - Call prime_debug_regs();
 - Restore thread->debug_regs from local save values in first step

> 
> > +
> > int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
> > {
> > 	int ret, s;
> > @@ -693,11 +757,25 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct
> kvm_vcpu *vcpu)
> > 	kvmppc_load_guest_fp(vcpu);
> > #endif
> >
> > +	/*
> > +	 * Clear current->thread.dbcr0 so that kernel does not
> > +	 * restore h/w registers on context switch in vcpu running state.
> > +	 */
> > +	vcpu->arch.debug_active = 1;
> 
> = true;

Ok

> 
> > +	vcpu->arch.saved_dbcr0 = current->thread.dbcr0;
> > +	current->thread.dbcr0 = 0;
> > +	kvmppc_booke_vcpu_load_debug_regs(vcpu);
> 
> static void switch_booke_debug_regs(struct thread_struct *new_thread)
> {
>         if ((current->thread.dbcr0 & DBCR0_IDM)
>                 || (new_thread->dbcr0 & DBCR0_IDM))
>                         prime_debug_regs(new_thread);
> }
> 
> The kernel will also restore debug state if the process we come from has
> debugging enabled. Please adjust the comment accordingly.

kvmppc_booke_vcpu_load_debug_regs(vcpu); cleares DBSR and DBCR0, even if previous process have used debug registers. Is not that sufficient? I do not think we should load all other registers.

> 
> > +
> > 	ret = __kvmppc_vcpu_run(kvm_run, vcpu);
> >
> > 	/* No need for kvm_guest_exit. It's done in handle_exit.
> > 	   We also get here with interrupts enabled. */
> >
> > +	/* Restore thread->dbcr0 */
> > +	vcpu->arch.debug_active = 0;
> > +	current->thread.dbcr0 = vcpu->arch.saved_dbcr0;
> > +	kvmppc_load_usespace_gebug();
> > +
> > #ifdef CONFIG_PPC_FPU
> > 	kvmppc_save_guest_fp(vcpu);
> >
> > @@ -753,6 +831,36 @@ static int emulation_exit(struct kvm_run *run, struct
> kvm_vcpu *vcpu)
> > 	}
> > }
> >
> > +/*
> > + * Currently we do not support debug resource emulation to guest,
> > + * so always exit to user space irrespective of user space is
> > + * expecting the debug exception or not. This is unexpected event
> > + * and let us leave the action on user space.
> > + */
> > +static int kvmppc_handle_debug(struct kvm_run *run, struct kvm_vcpu *vcpu)
> > +{
> > +	u32 dbsr = mfspr(SPRN_DBSR);
> > +
> > +	kvmppc_clear_dbsr();
> > +	run->debug.arch.status = 0;
> > +	run->debug.arch.address = vcpu->arch.pc;
> > +
> > +	if (dbsr & (DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4)) {
> > +		run->debug.arch.status |= KVMPPC_DEBUG_BREAKPOINT;
> > +	} else {
> > +		if (dbsr & (DBSR_DAC1W | DBSR_DAC2W))
> > +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_WRITE;
> > +		else if (dbsr & (DBSR_DAC1R | DBSR_DAC2R))
> > +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_READ;
> > +		if (dbsr & (DBSR_DAC1R | DBSR_DAC1W))
> > +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[0];
> > +		else if (dbsr & (DBSR_DAC2R | DBSR_DAC2W))
> > +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[1];
> > +	}
> > +
> > +	return RESUME_HOST;
> > +}
> > +
> > static void kvmppc_fill_pt_regs(struct pt_regs *regs)
> > {
> > 	ulong r1, ip, msr, lr;
> > @@ -1112,18 +1220,10 @@ int kvmppc_handle_exit(struct kvm_run *run, struct
> kvm_vcpu *vcpu,
> > 	}
> >
> > 	case BOOKE_INTERRUPT_DEBUG: {
> > -		u32 dbsr;
> > -
> > -		vcpu->arch.pc = mfspr(SPRN_CSRR0);
> > -
> > -		/* clear IAC events in DBSR register */
> > -		dbsr = mfspr(SPRN_DBSR);
> > -		dbsr &= DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4;
> > -		mtspr(SPRN_DBSR, dbsr);
> > -
> > -		run->exit_reason = KVM_EXIT_DEBUG;
> > +		r = kvmppc_handle_debug(run, vcpu);
> > +		if (r == RESUME_HOST)
> > +			run->exit_reason = KVM_EXIT_DEBUG;
> > 		kvmppc_account_exit(vcpu, DEBUG_EXITS);
> > -		r = RESUME_HOST;
> > 		break;
> > 	}
> >
> > @@ -1174,7 +1274,7 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
> > 	kvmppc_set_msr(vcpu, 0);
> >
> > #ifndef CONFIG_KVM_BOOKE_HV
> > -	vcpu->arch.shadow_msr = MSR_USER | MSR_DE | MSR_IS | MSR_DS;
> > +	vcpu->arch.shadow_msr = MSR_USER | MSR_IS | MSR_DS;
> > 	vcpu->arch.shadow_pid = 1;
> > 	vcpu->arch.shared->msr = 0;
> > #endif
> > @@ -1529,12 +1629,6 @@ int kvm_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
> struct kvm_one_reg *reg)
> > 	return r;
> > }
> >
> > -int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> > -					 struct kvm_guest_debug *dbg)
> > -{
> > -	return -EINVAL;
> > -}
> > -
> > int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
> > {
> > 	return -ENOTSUPP;
> > @@ -1640,16 +1734,128 @@ void kvmppc_decrementer_func(unsigned long data)
> > 	kvmppc_set_tsr_bits(vcpu, TSR_DIS);
> > }
> >
> > +static void kvmppc_booke_vcpu_put_debug_regs(struct kvm_vcpu *vcpu)
> > +{
> > +	/* Disable all debug events First */
> 
> first
> 
> > +	mtspr(SPRN_DBCR0, 0x0);
> > +	/* Disable pending debug event by clearing DBSR */
> > +	kvmppc_clear_dbsr();
> 
> kvmppc_handle_debug() happens with preemption enabled, no?

Want to clarify,  preemption will be enabled on calling local_irq_enable(); in kvmppc_handle_exit()?

> So we can have a
> debug event that gets cleared on preempt by this.

Should we read the DBSR in before local_irq_enable() in kvmppc_handle_exit()?

> 
> > +}
> > +
> > +int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> > +					 struct kvm_guest_debug *dbg)
> > +{
> > +	struct kvmppc_booke_debug_reg *dbg_reg;
> > +	int n, b = 0, w = 0;
> > +	const u32 bp_code[] = {
> > +		DBCR0_IAC1 | DBCR0_IDM,
> > +		DBCR0_IAC2 | DBCR0_IDM,
> > +		DBCR0_IAC3 | DBCR0_IDM,
> > +		DBCR0_IAC4 | DBCR0_IDM
> > +	};
> > +	const u32 wp_code[] = {
> > +		DBCR0_DAC1W | DBCR0_IDM,
> > +		DBCR0_DAC2W | DBCR0_IDM,
> > +		DBCR0_DAC1R | DBCR0_IDM,
> > +		DBCR0_DAC2R | DBCR0_IDM
> > +	};
> > +
> > +	if (!(dbg->control & KVM_GUESTDBG_ENABLE)) {
> > +		/* Clear All debug events */
> > +		vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> > +		vcpu->guest_debug = 0;
> > +#ifdef CONFIG_KVM_BOOKE_HV
> > +		/*
> > +		 * When user space is not using the debug resources
> > +		 * then allow guest to change the MSR.DE.
> > +		 */
> > +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
> > +#endif
> > +		return 0;
> > +	}
> > +
> > +#ifdef CONFIG_KVM_BOOKE_HV
> > +		/*
> > +		 * When user space is using the debug resource then
> > +		 * do not allow guest to change the MSR.DE.
> > +		 */
> > +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
> 
> This is supposed to be |= right?

Yes :-)

> 
> > +#endif
> > +	vcpu->guest_debug = dbg->control;
> > +	vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> > +	/* Set DBCR0_EDM in guest visible DBCR0 register. */
> > +	vcpu->arch.dbg_reg.dbcr0 = DBCR0_EDM;
> > +
> > +	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
> > +		vcpu->arch.shadow_dbg_reg.dbcr0 |= DBCR0_IDM | DBCR0_IC;
> > +
> > +	if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
> > +		/* Code below handles only HW breakpoints */
> 
> Please move the comment out of the one-lined branch. Just put it below the
> return.

Ok

> 
> > +		return 0;
> > +
> > +	dbg_reg = &(vcpu->arch.shadow_dbg_reg);
> > +
> > +	/*
> > +	 * On BOOKE (e500v2); Set DBCR1 and DBCR2 to allow debug events
> > +	 * to occur when MSR.PR is set.
> > +	 * On BOOKE-HV (e500mc+); MSR.PR = 0 when guest is running. So we
> > +	 * should clear DBCR1 and DBCR2. And EPCR.DUVD is used to control
> > +	 * that debug events will not come in hypervisor (GS = 0).
> > +	 */
> > +#ifdef CONFIG_KVM_BOOKE_HV
> > +	dbg_reg->dbcr1 = 0;
> > +	dbg_reg->dbcr2 = 0;
> > +#else
> > +	dbg_reg->dbcr1 = DBCR1_IAC1US | DBCR1_IAC2US | DBCR1_IAC3US |
> > +			  DBCR1_IAC4US;
> > +	dbg_reg->dbcr2 = DBCR2_DAC1US | DBCR2_DAC2US;
> > +#endif
> > +
> > +	for (n = 0; n < (KVMPPC_BOOKE_IAC_NUM + KVMPPC_BOOKE_DAC_NUM); n++) {
> > +		u32 type = dbg->arch.bp[n].type;
> > +
> > +		if (!type)
> > +			continue;
> 
> Scott's comment on why a zero-type is different from any other invalid type is
> still outstanding I think.

Userspce does following 
 - dbg->arch.bp[] is first zero initialized. So dbg->arch.bp[n].type is 0 (KVMPPC_DEBUG_NONE)
 - Then set following for valid breakpoints/watchpoints:
   -- dbg->arch.bp[n].type (KVMPPC_DEBUG_BREAKPOINT or KVMPPC_DEBUG_WATCH_WRITE or KVMPPC_DEBUG_WATCH_READ)
   -- dbg->arch.bp[n].addr

I tried to avoid loop when type is 0 (probably saying type == KVMPPC_DEBUG_NONE should be more clear).
if type is invalid then should we return -EINVAL.

-Bharat

> 
> 
> Alex
> 
> > +
> > +		if (type & (KVMPPC_DEBUG_WATCH_READ |
> > +			    KVMPPC_DEBUG_WATCH_WRITE)) {
> > +			if (w >= KVMPPC_BOOKE_DAC_NUM)
> > +				return -EINVAL;
> > +
> > +			if (type & KVMPPC_DEBUG_WATCH_READ)
> > +				dbg_reg->dbcr0 |= wp_code[w + 2];
> > +			if (type & KVMPPC_DEBUG_WATCH_WRITE)
> > +				dbg_reg->dbcr0 |= wp_code[w];
> > +
> > +			dbg_reg->dac[w] = dbg->arch.bp[n].addr;
> > +			w++;
> > +		} else if (type & KVMPPC_DEBUG_BREAKPOINT) {
> > +			if (b >= KVMPPC_BOOKE_IAC_NUM)
> > +				return -EINVAL;
> > +
> > +			dbg_reg->dbcr0 |= bp_code[b];
> > +			dbg_reg->iac[b] = dbg->arch.bp[n].addr;
> > +			b++;
> > +		}
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> > void kvmppc_booke_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> > {
> > 	vcpu->cpu = smp_processor_id();
> > 	current->thread.kvm_vcpu = vcpu;
> > +
> > +	kvmppc_booke_vcpu_load_debug_regs(vcpu);
> > }
> >
> > void kvmppc_booke_vcpu_put(struct kvm_vcpu *vcpu)
> > {
> > 	current->thread.kvm_vcpu = NULL;
> > 	vcpu->cpu = -1;
> > +
> > +	kvmppc_booke_vcpu_put_debug_regs(vcpu);
> > }
> >
> > int __init kvmppc_booke_init(void)
> > diff --git a/arch/powerpc/kvm/booke.h b/arch/powerpc/kvm/booke.h
> > index 5fd1ba6..a1ff67d 100644
> > --- a/arch/powerpc/kvm/booke.h
> > +++ b/arch/powerpc/kvm/booke.h
> > @@ -129,4 +129,9 @@ static inline void kvmppc_save_guest_fp(struct kvm_vcpu
> *vcpu)
> > 		giveup_fpu(current);
> > #endif
> > }
> > +
> > +static inline void kvmppc_clear_dbsr(void)
> > +{
> > +	mtspr(SPRN_DBSR, mfspr(SPRN_DBSR));
> > +}
> > #endif /* __KVM_BOOKE_H__ */
> > --
> > 1.7.0.4
> >
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 


--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Alexander Graf May 2, 2013, 11:05 a.m. UTC | #3

On 02.05.2013, at 11:46, Bhushan Bharat-R65777 wrote:

> 
> 
>> -----Original Message-----
>> From: Alexander Graf [mailto:agraf@suse.de]
>> Sent: Friday, April 26, 2013 4:46 PM
>> To: Bhushan Bharat-R65777
>> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421; Bhushan
>> Bharat-R65777
>> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
>> 
>> 
>> On 08.04.2013, at 12:32, Bharat Bhushan wrote:
>> 
>>> From: Bharat Bhushan <bharat.bhushan@freescale.com>
>>> 
>>> This patch adds the debug stub support on booke/bookehv.
>>> Now QEMU debug stub can use hw breakpoint, watchpoint and software
>>> breakpoint to debug guest.
>>> 
>>> Debug registers are saved/restored on vcpu_put()/vcpu_get().
>>> Also the debug registers are saved restored only if guest
>>> is using debug resources.
>>> 
>>> Currently we do not support debug resource emulation to guest,
>>> so always exit to user space irrespective of user space is expecting
>>> the debug exception or not. This is unexpected event and let us
>>> leave the action on user space. This is similar to what it was before,
>>> only thing is that now we have proper exit state available to user space.
>>> 
>>> Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
>>> ---
>>> arch/powerpc/include/asm/kvm_host.h |    8 +
>>> arch/powerpc/include/uapi/asm/kvm.h |   22 +++-
>>> arch/powerpc/kvm/booke.c            |  242 ++++++++++++++++++++++++++++++++---
>>> arch/powerpc/kvm/booke.h            |    5 +
>>> 4 files changed, 255 insertions(+), 22 deletions(-)
>>> 
>>> diff --git a/arch/powerpc/include/asm/kvm_host.h
>> b/arch/powerpc/include/asm/kvm_host.h
>>> index e34f8fe..b9ad20f 100644
>>> --- a/arch/powerpc/include/asm/kvm_host.h
>>> +++ b/arch/powerpc/include/asm/kvm_host.h
>>> @@ -505,7 +505,15 @@ struct kvm_vcpu_arch {
>>> 	u32 mmucfg;
>>> 	u32 epr;
>>> 	u32 crit_save;
>>> +
>>> +	/* Flag indicating that debug registers are used by guest */
>>> +	bool debug_active;
>>> +	/* for save/restore thread->dbcr0 on vcpu run/heavyweight_exit */
>>> +	u32 saved_dbcr0;
>>> +	/* guest debug registers*/
>>> 	struct kvmppc_booke_debug_reg dbg_reg;
>>> +	/* shadow debug registers */
>>> +	struct kvmppc_booke_debug_reg shadow_dbg_reg;
>>> #endif
>>> 	gpa_t paddr_accessed;
>>> 	gva_t vaddr_accessed;
>>> diff --git a/arch/powerpc/include/uapi/asm/kvm.h
>> b/arch/powerpc/include/uapi/asm/kvm.h
>>> index c0c38ed..d7ce449 100644
>>> --- a/arch/powerpc/include/uapi/asm/kvm.h
>>> +++ b/arch/powerpc/include/uapi/asm/kvm.h
>>> @@ -25,6 +25,7 @@
>>> /* Select powerpc specific features in <linux/kvm.h> */
>>> #define __KVM_HAVE_SPAPR_TCE
>>> #define __KVM_HAVE_PPC_SMT
>>> +#define __KVM_HAVE_GUEST_DEBUG
>>> 
>>> struct kvm_regs {
>>> 	__u64 pc;
>>> @@ -267,7 +268,24 @@ struct kvm_fpu {
>>> 	__u64 fpr[32];
>>> };
>>> 
>>> +/*
>>> + * Defines for h/w breakpoint, watchpoint (read, write or both) and
>>> + * software breakpoint.
>>> + * These are used as "type" in KVM_SET_GUEST_DEBUG ioctl and "status"
>>> + * for KVM_DEBUG_EXIT.
>>> + */
>>> +#define KVMPPC_DEBUG_NONE		0x0
>>> +#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
>>> +#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
>>> +#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
>>> struct kvm_debug_exit_arch {
>>> +	__u64 address;
>>> +	/*
>>> +	 * exiting to userspace because of h/w breakpoint, watchpoint
>>> +	 * (read, write or both) and software breakpoint.
>>> +	 */
>>> +	__u32 status;
>>> +	__u32 reserved;
>>> };
>>> 
>>> /* for KVM_SET_GUEST_DEBUG */
>>> @@ -279,10 +297,6 @@ struct kvm_guest_debug_arch {
>>> 		 * Type denotes h/w breakpoint, read watchpoint, write
>>> 		 * watchpoint or watchpoint (both read and write).
>>> 		 */
>>> -#define KVMPPC_DEBUG_NONE		0x0
>>> -#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
>>> -#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
>>> -#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
>>> 		__u32 type;
>>> 		__u32 reserved;
>>> 	} bp[16];
>>> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c
>>> index 97ae158..0e93416 100644
>>> --- a/arch/powerpc/kvm/booke.c
>>> +++ b/arch/powerpc/kvm/booke.c
>>> @@ -133,6 +133,29 @@ static void kvmppc_vcpu_sync_fpu(struct kvm_vcpu *vcpu)
>>> #endif
>>> }
>>> 
>>> +static void kvmppc_vcpu_sync_debug(struct kvm_vcpu *vcpu)
>>> +{
>>> +	/* Synchronize guest's desire to get debug interrupts into shadow MSR */
>>> +#ifndef CONFIG_KVM_BOOKE_HV
>>> +	vcpu->arch.shadow_msr &= ~MSR_DE;
>>> +	vcpu->arch.shadow_msr |= vcpu->arch.shared->msr & MSR_DE;
>>> +#endif
>>> +
>>> +	/* Force enable debug interrupts when user space wants to debug */
>>> +	if (vcpu->guest_debug) {
>>> +#ifdef CONFIG_KVM_BOOKE_HV
>>> +		/*
>>> +		 * Since there is no shadow MSR, sync MSR_DE into the guest
>>> +		 * visible MSR.
>>> +		 */
>>> +		vcpu->arch.shared->msr |= MSR_DE;
>>> +#else
>>> +		vcpu->arch.shadow_msr |= MSR_DE;
>>> +		vcpu->arch.shared->msr &= ~MSR_DE;
>>> +#endif
>>> +	}
>>> +}
>>> +
>>> /*
>>> * Helper function for "full" MSR writes.  No need to call this if only
>>> * EE/CE/ME/DE/RI are changing.
>>> @@ -150,6 +173,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u32 new_msr)
>>> 	kvmppc_mmu_msr_notify(vcpu, old_msr);
>>> 	kvmppc_vcpu_sync_spe(vcpu);
>>> 	kvmppc_vcpu_sync_fpu(vcpu);
>>> +	kvmppc_vcpu_sync_debug(vcpu);
>>> }
>>> 
>>> static void kvmppc_booke_queue_irqprio(struct kvm_vcpu *vcpu,
>>> @@ -646,6 +670,46 @@ int kvmppc_core_check_requests(struct kvm_vcpu *vcpu)
>>> 	return r;
>>> }
>>> 
>>> +static void kvmppc_load_usespace_gebug(void)
>> 
>> ...
> 
> Please tell What does "..."  mean. Does that mean next comment is about the above function?

It means "this one is so obvious I don't even have to write anything here". Usespace? Gebug? Seriously? :)

> 
>> 
>>> +{
>>> +	switch_booke_debug_regs(&current->thread);
>>> +}
>>> +
>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu *vcpu)
>>> +{
>>> +	if (!vcpu->arch.debug_active)
>>> +		return;
>>> +
>>> +	/* Disable all debug events and clead pending debug events */
>>> +	mtspr(SPRN_DBCR0, 0x0);
>>> +	kvmppc_clear_dbsr();
>>> +
>>> +	/*
>>> +	 * Check whether guest still need debug resource, if not then there
>>> +	 * is no need to restore guest context.
>>> +	 */
>>> +	if (!vcpu->arch.shadow_dbg_reg.dbcr0)
>>> +		return;
>>> +
>>> +	/* Load Guest Context */
>>> +	mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
>>> +	mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2);
>>> +#ifdef CONFIG_KVM_E500MC
>>> +	mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
>> 
>> You need to make sure DBCR4 is 0 when you leave things back to normal user
>> space. Otherwise guest debug can interfere with host debug.
> 
> 
> ok
> 
>> 
>>> +#endif
>>> +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
>>> +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
>>> +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
>>> +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
>>> +#endif
>>> +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
>>> +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
>>> +
>>> +	/* Enable debug events after other debug registers restored */
>>> +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0);
>>> +}
>> 
>> All of the code above looks suspiciously similar to prime_debug_regs();. Can't
>> we somehow reuse that?
> 
> I think we can if
> - Save thread->debug_regs in local data structure 

Yes, it can even be on the stack.

> - Load vcpu->arch->debug_regs in thread->debug_regs
> - Call prime_debug_regs();
> - Restore thread->debug_regs from local save values in first step

On heavyweight exit, based on the values on stack, yes.

> 
>> 
>>> +
>>> int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
>>> {
>>> 	int ret, s;
>>> @@ -693,11 +757,25 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct
>> kvm_vcpu *vcpu)
>>> 	kvmppc_load_guest_fp(vcpu);
>>> #endif
>>> 
>>> +	/*
>>> +	 * Clear current->thread.dbcr0 so that kernel does not
>>> +	 * restore h/w registers on context switch in vcpu running state.
>>> +	 */
>>> +	vcpu->arch.debug_active = 1;
>> 
>> = true;
> 
> Ok
> 
>> 
>>> +	vcpu->arch.saved_dbcr0 = current->thread.dbcr0;
>>> +	current->thread.dbcr0 = 0;
>>> +	kvmppc_booke_vcpu_load_debug_regs(vcpu);
>> 
>> static void switch_booke_debug_regs(struct thread_struct *new_thread)
>> {
>>        if ((current->thread.dbcr0 & DBCR0_IDM)
>>                || (new_thread->dbcr0 & DBCR0_IDM))
>>                        prime_debug_regs(new_thread);
>> }
>> 
>> The kernel will also restore debug state if the process we come from has
>> debugging enabled. Please adjust the comment accordingly.
> 
> kvmppc_booke_vcpu_load_debug_regs(vcpu); cleares DBSR and DBCR0, even if previous process have used debug registers. Is not that sufficient? I do not think we should load all other registers.

It's sufficient, but the comment is wrong.

> 
>> 
>>> +
>>> 	ret = __kvmppc_vcpu_run(kvm_run, vcpu);
>>> 
>>> 	/* No need for kvm_guest_exit. It's done in handle_exit.
>>> 	   We also get here with interrupts enabled. */
>>> 
>>> +	/* Restore thread->dbcr0 */
>>> +	vcpu->arch.debug_active = 0;
>>> +	current->thread.dbcr0 = vcpu->arch.saved_dbcr0;
>>> +	kvmppc_load_usespace_gebug();
>>> +
>>> #ifdef CONFIG_PPC_FPU
>>> 	kvmppc_save_guest_fp(vcpu);
>>> 
>>> @@ -753,6 +831,36 @@ static int emulation_exit(struct kvm_run *run, struct
>> kvm_vcpu *vcpu)
>>> 	}
>>> }
>>> 
>>> +/*
>>> + * Currently we do not support debug resource emulation to guest,
>>> + * so always exit to user space irrespective of user space is
>>> + * expecting the debug exception or not. This is unexpected event
>>> + * and let us leave the action on user space.
>>> + */
>>> +static int kvmppc_handle_debug(struct kvm_run *run, struct kvm_vcpu *vcpu)
>>> +{
>>> +	u32 dbsr = mfspr(SPRN_DBSR);
>>> +
>>> +	kvmppc_clear_dbsr();
>>> +	run->debug.arch.status = 0;
>>> +	run->debug.arch.address = vcpu->arch.pc;
>>> +
>>> +	if (dbsr & (DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4)) {
>>> +		run->debug.arch.status |= KVMPPC_DEBUG_BREAKPOINT;
>>> +	} else {
>>> +		if (dbsr & (DBSR_DAC1W | DBSR_DAC2W))
>>> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_WRITE;
>>> +		else if (dbsr & (DBSR_DAC1R | DBSR_DAC2R))
>>> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_READ;
>>> +		if (dbsr & (DBSR_DAC1R | DBSR_DAC1W))
>>> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[0];
>>> +		else if (dbsr & (DBSR_DAC2R | DBSR_DAC2W))
>>> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[1];
>>> +	}
>>> +
>>> +	return RESUME_HOST;
>>> +}
>>> +
>>> static void kvmppc_fill_pt_regs(struct pt_regs *regs)
>>> {
>>> 	ulong r1, ip, msr, lr;
>>> @@ -1112,18 +1220,10 @@ int kvmppc_handle_exit(struct kvm_run *run, struct
>> kvm_vcpu *vcpu,
>>> 	}
>>> 
>>> 	case BOOKE_INTERRUPT_DEBUG: {
>>> -		u32 dbsr;
>>> -
>>> -		vcpu->arch.pc = mfspr(SPRN_CSRR0);
>>> -
>>> -		/* clear IAC events in DBSR register */
>>> -		dbsr = mfspr(SPRN_DBSR);
>>> -		dbsr &= DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4;
>>> -		mtspr(SPRN_DBSR, dbsr);
>>> -
>>> -		run->exit_reason = KVM_EXIT_DEBUG;
>>> +		r = kvmppc_handle_debug(run, vcpu);
>>> +		if (r == RESUME_HOST)
>>> +			run->exit_reason = KVM_EXIT_DEBUG;
>>> 		kvmppc_account_exit(vcpu, DEBUG_EXITS);
>>> -		r = RESUME_HOST;
>>> 		break;
>>> 	}
>>> 
>>> @@ -1174,7 +1274,7 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
>>> 	kvmppc_set_msr(vcpu, 0);
>>> 
>>> #ifndef CONFIG_KVM_BOOKE_HV
>>> -	vcpu->arch.shadow_msr = MSR_USER | MSR_DE | MSR_IS | MSR_DS;
>>> +	vcpu->arch.shadow_msr = MSR_USER | MSR_IS | MSR_DS;
>>> 	vcpu->arch.shadow_pid = 1;
>>> 	vcpu->arch.shared->msr = 0;
>>> #endif
>>> @@ -1529,12 +1629,6 @@ int kvm_vcpu_ioctl_set_one_reg(struct kvm_vcpu *vcpu,
>> struct kvm_one_reg *reg)
>>> 	return r;
>>> }
>>> 
>>> -int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
>>> -					 struct kvm_guest_debug *dbg)
>>> -{
>>> -	return -EINVAL;
>>> -}
>>> -
>>> int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
>>> {
>>> 	return -ENOTSUPP;
>>> @@ -1640,16 +1734,128 @@ void kvmppc_decrementer_func(unsigned long data)
>>> 	kvmppc_set_tsr_bits(vcpu, TSR_DIS);
>>> }
>>> 
>>> +static void kvmppc_booke_vcpu_put_debug_regs(struct kvm_vcpu *vcpu)
>>> +{
>>> +	/* Disable all debug events First */
>> 
>> first
>> 
>>> +	mtspr(SPRN_DBCR0, 0x0);
>>> +	/* Disable pending debug event by clearing DBSR */
>>> +	kvmppc_clear_dbsr();
>> 
>> kvmppc_handle_debug() happens with preemption enabled, no?
> 
> Want to clarify,  preemption will be enabled on calling local_irq_enable(); in kvmppc_handle_exit()?

Yes. Implicitly :).

> 
>> So we can have a
>> debug event that gets cleared on preempt by this.
> 
> Should we read the DBSR in before local_irq_enable() in kvmppc_handle_exit()?

We have to, yes. Otherwise we could get preempted in between and get bogus data I presume.

> 
>> 
>>> +}
>>> +
>>> +int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
>>> +					 struct kvm_guest_debug *dbg)
>>> +{
>>> +	struct kvmppc_booke_debug_reg *dbg_reg;
>>> +	int n, b = 0, w = 0;
>>> +	const u32 bp_code[] = {
>>> +		DBCR0_IAC1 | DBCR0_IDM,
>>> +		DBCR0_IAC2 | DBCR0_IDM,
>>> +		DBCR0_IAC3 | DBCR0_IDM,
>>> +		DBCR0_IAC4 | DBCR0_IDM
>>> +	};
>>> +	const u32 wp_code[] = {
>>> +		DBCR0_DAC1W | DBCR0_IDM,
>>> +		DBCR0_DAC2W | DBCR0_IDM,
>>> +		DBCR0_DAC1R | DBCR0_IDM,
>>> +		DBCR0_DAC2R | DBCR0_IDM
>>> +	};
>>> +
>>> +	if (!(dbg->control & KVM_GUESTDBG_ENABLE)) {
>>> +		/* Clear All debug events */
>>> +		vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
>>> +		vcpu->guest_debug = 0;
>>> +#ifdef CONFIG_KVM_BOOKE_HV
>>> +		/*
>>> +		 * When user space is not using the debug resources
>>> +		 * then allow guest to change the MSR.DE.
>>> +		 */
>>> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
>>> +#endif
>>> +		return 0;
>>> +	}
>>> +
>>> +#ifdef CONFIG_KVM_BOOKE_HV
>>> +		/*
>>> +		 * When user space is using the debug resource then
>>> +		 * do not allow guest to change the MSR.DE.
>>> +		 */
>>> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
>> 
>> This is supposed to be |= right?
> 
> Yes :-)
> 
>> 
>>> +#endif
>>> +	vcpu->guest_debug = dbg->control;
>>> +	vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
>>> +	/* Set DBCR0_EDM in guest visible DBCR0 register. */
>>> +	vcpu->arch.dbg_reg.dbcr0 = DBCR0_EDM;
>>> +
>>> +	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
>>> +		vcpu->arch.shadow_dbg_reg.dbcr0 |= DBCR0_IDM | DBCR0_IC;
>>> +
>>> +	if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
>>> +		/* Code below handles only HW breakpoints */
>> 
>> Please move the comment out of the one-lined branch. Just put it below the
>> return.
> 
> Ok
> 
>> 
>>> +		return 0;
>>> +
>>> +	dbg_reg = &(vcpu->arch.shadow_dbg_reg);
>>> +
>>> +	/*
>>> +	 * On BOOKE (e500v2); Set DBCR1 and DBCR2 to allow debug events
>>> +	 * to occur when MSR.PR is set.
>>> +	 * On BOOKE-HV (e500mc+); MSR.PR = 0 when guest is running. So we
>>> +	 * should clear DBCR1 and DBCR2. And EPCR.DUVD is used to control
>>> +	 * that debug events will not come in hypervisor (GS = 0).
>>> +	 */
>>> +#ifdef CONFIG_KVM_BOOKE_HV
>>> +	dbg_reg->dbcr1 = 0;
>>> +	dbg_reg->dbcr2 = 0;
>>> +#else
>>> +	dbg_reg->dbcr1 = DBCR1_IAC1US | DBCR1_IAC2US | DBCR1_IAC3US |
>>> +			  DBCR1_IAC4US;
>>> +	dbg_reg->dbcr2 = DBCR2_DAC1US | DBCR2_DAC2US;
>>> +#endif
>>> +
>>> +	for (n = 0; n < (KVMPPC_BOOKE_IAC_NUM + KVMPPC_BOOKE_DAC_NUM); n++) {
>>> +		u32 type = dbg->arch.bp[n].type;
>>> +
>>> +		if (!type)
>>> +			continue;
>> 
>> Scott's comment on why a zero-type is different from any other invalid type is
>> still outstanding I think.
> 
> Userspce does following 
> - dbg->arch.bp[] is first zero initialized. So dbg->arch.bp[n].type is 0 (KVMPPC_DEBUG_NONE)
> - Then set following for valid breakpoints/watchpoints:
>   -- dbg->arch.bp[n].type (KVMPPC_DEBUG_BREAKPOINT or KVMPPC_DEBUG_WATCH_WRITE or KVMPPC_DEBUG_WATCH_READ)
>   -- dbg->arch.bp[n].addr
> 
> I tried to avoid loop when type is 0 (probably saying type == KVMPPC_DEBUG_NONE should be more clear).

That works, yes.

> if type is invalid then should we return -EINVAL.

Yes.


Alex

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Bharat Bhushan May 2, 2013, 2 p.m. UTC | #4

> -----Original Message-----
> From: Alexander Graf [mailto:agraf@suse.de]
> Sent: Thursday, May 02, 2013 4:35 PM
> To: Bhushan Bharat-R65777
> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
> 
> 
> On 02.05.2013, at 11:46, Bhushan Bharat-R65777 wrote:
> 
> >
> >
> >> -----Original Message-----
> >> From: Alexander Graf [mailto:agraf@suse.de]
> >> Sent: Friday, April 26, 2013 4:46 PM
> >> To: Bhushan Bharat-R65777
> >> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421;
> >> Bhushan
> >> Bharat-R65777
> >> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub
> >> support
> >>
> >>
> >> On 08.04.2013, at 12:32, Bharat Bhushan wrote:
> >>
> >>> From: Bharat Bhushan <bharat.bhushan@freescale.com>
> >>>
> >>> This patch adds the debug stub support on booke/bookehv.
> >>> Now QEMU debug stub can use hw breakpoint, watchpoint and software
> >>> breakpoint to debug guest.
> >>>
> >>> Debug registers are saved/restored on vcpu_put()/vcpu_get().
> >>> Also the debug registers are saved restored only if guest is using
> >>> debug resources.
> >>>
> >>> Currently we do not support debug resource emulation to guest, so
> >>> always exit to user space irrespective of user space is expecting
> >>> the debug exception or not. This is unexpected event and let us
> >>> leave the action on user space. This is similar to what it was
> >>> before, only thing is that now we have proper exit state available to user
> space.
> >>>
> >>> Signed-off-by: Bharat Bhushan <bharat.bhushan@freescale.com>
> >>> ---
> >>> arch/powerpc/include/asm/kvm_host.h |    8 +
> >>> arch/powerpc/include/uapi/asm/kvm.h |   22 +++-
> >>> arch/powerpc/kvm/booke.c            |  242 ++++++++++++++++++++++++++++++++-
> --
> >>> arch/powerpc/kvm/booke.h            |    5 +
> >>> 4 files changed, 255 insertions(+), 22 deletions(-)
> >>>
> >>> diff --git a/arch/powerpc/include/asm/kvm_host.h
> >> b/arch/powerpc/include/asm/kvm_host.h
> >>> index e34f8fe..b9ad20f 100644
> >>> --- a/arch/powerpc/include/asm/kvm_host.h
> >>> +++ b/arch/powerpc/include/asm/kvm_host.h
> >>> @@ -505,7 +505,15 @@ struct kvm_vcpu_arch {
> >>> 	u32 mmucfg;
> >>> 	u32 epr;
> >>> 	u32 crit_save;
> >>> +
> >>> +	/* Flag indicating that debug registers are used by guest */
> >>> +	bool debug_active;
> >>> +	/* for save/restore thread->dbcr0 on vcpu run/heavyweight_exit */
> >>> +	u32 saved_dbcr0;
> >>> +	/* guest debug registers*/
> >>> 	struct kvmppc_booke_debug_reg dbg_reg;
> >>> +	/* shadow debug registers */
> >>> +	struct kvmppc_booke_debug_reg shadow_dbg_reg;
> >>> #endif
> >>> 	gpa_t paddr_accessed;
> >>> 	gva_t vaddr_accessed;
> >>> diff --git a/arch/powerpc/include/uapi/asm/kvm.h
> >> b/arch/powerpc/include/uapi/asm/kvm.h
> >>> index c0c38ed..d7ce449 100644
> >>> --- a/arch/powerpc/include/uapi/asm/kvm.h
> >>> +++ b/arch/powerpc/include/uapi/asm/kvm.h
> >>> @@ -25,6 +25,7 @@
> >>> /* Select powerpc specific features in <linux/kvm.h> */ #define
> >>> __KVM_HAVE_SPAPR_TCE #define __KVM_HAVE_PPC_SMT
> >>> +#define __KVM_HAVE_GUEST_DEBUG
> >>>
> >>> struct kvm_regs {
> >>> 	__u64 pc;
> >>> @@ -267,7 +268,24 @@ struct kvm_fpu {
> >>> 	__u64 fpr[32];
> >>> };
> >>>
> >>> +/*
> >>> + * Defines for h/w breakpoint, watchpoint (read, write or both) and
> >>> + * software breakpoint.
> >>> + * These are used as "type" in KVM_SET_GUEST_DEBUG ioctl and "status"
> >>> + * for KVM_DEBUG_EXIT.
> >>> + */
> >>> +#define KVMPPC_DEBUG_NONE		0x0
> >>> +#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> >>> +#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> >>> +#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> >>> struct kvm_debug_exit_arch {
> >>> +	__u64 address;
> >>> +	/*
> >>> +	 * exiting to userspace because of h/w breakpoint, watchpoint
> >>> +	 * (read, write or both) and software breakpoint.
> >>> +	 */
> >>> +	__u32 status;
> >>> +	__u32 reserved;
> >>> };
> >>>
> >>> /* for KVM_SET_GUEST_DEBUG */
> >>> @@ -279,10 +297,6 @@ struct kvm_guest_debug_arch {
> >>> 		 * Type denotes h/w breakpoint, read watchpoint, write
> >>> 		 * watchpoint or watchpoint (both read and write).
> >>> 		 */
> >>> -#define KVMPPC_DEBUG_NONE		0x0
> >>> -#define KVMPPC_DEBUG_BREAKPOINT		(1UL << 1)
> >>> -#define KVMPPC_DEBUG_WATCH_WRITE	(1UL << 2)
> >>> -#define KVMPPC_DEBUG_WATCH_READ		(1UL << 3)
> >>> 		__u32 type;
> >>> 		__u32 reserved;
> >>> 	} bp[16];
> >>> diff --git a/arch/powerpc/kvm/booke.c b/arch/powerpc/kvm/booke.c
> >>> index 97ae158..0e93416 100644
> >>> --- a/arch/powerpc/kvm/booke.c
> >>> +++ b/arch/powerpc/kvm/booke.c
> >>> @@ -133,6 +133,29 @@ static void kvmppc_vcpu_sync_fpu(struct
> >>> kvm_vcpu *vcpu) #endif }
> >>>
> >>> +static void kvmppc_vcpu_sync_debug(struct kvm_vcpu *vcpu) {
> >>> +	/* Synchronize guest's desire to get debug interrupts into shadow
> >>> +MSR */ #ifndef CONFIG_KVM_BOOKE_HV
> >>> +	vcpu->arch.shadow_msr &= ~MSR_DE;
> >>> +	vcpu->arch.shadow_msr |= vcpu->arch.shared->msr & MSR_DE; #endif
> >>> +
> >>> +	/* Force enable debug interrupts when user space wants to debug */
> >>> +	if (vcpu->guest_debug) {
> >>> +#ifdef CONFIG_KVM_BOOKE_HV
> >>> +		/*
> >>> +		 * Since there is no shadow MSR, sync MSR_DE into the guest
> >>> +		 * visible MSR.
> >>> +		 */
> >>> +		vcpu->arch.shared->msr |= MSR_DE; #else
> >>> +		vcpu->arch.shadow_msr |= MSR_DE;
> >>> +		vcpu->arch.shared->msr &= ~MSR_DE; #endif
> >>> +	}
> >>> +}
> >>> +
> >>> /*
> >>> * Helper function for "full" MSR writes.  No need to call this if
> >>> only
> >>> * EE/CE/ME/DE/RI are changing.
> >>> @@ -150,6 +173,7 @@ void kvmppc_set_msr(struct kvm_vcpu *vcpu, u32 new_msr)
> >>> 	kvmppc_mmu_msr_notify(vcpu, old_msr);
> >>> 	kvmppc_vcpu_sync_spe(vcpu);
> >>> 	kvmppc_vcpu_sync_fpu(vcpu);
> >>> +	kvmppc_vcpu_sync_debug(vcpu);
> >>> }
> >>>
> >>> static void kvmppc_booke_queue_irqprio(struct kvm_vcpu *vcpu, @@
> >>> -646,6 +670,46 @@ int kvmppc_core_check_requests(struct kvm_vcpu *vcpu)
> >>> 	return r;
> >>> }
> >>>
> >>> +static void kvmppc_load_usespace_gebug(void)
> >>
> >> ...
> >
> > Please tell What does "..."  mean. Does that mean next comment is about the
> above function?
> 
> It means "this one is so obvious I don't even have to write anything here".
> Usespace? Gebug? Seriously? :)

ehhh, kicking myself 

> 
> >
> >>
> >>> +{
> >>> +	switch_booke_debug_regs(&current->thread);
> >>> +}
> >>> +
> >>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
> >>> +*vcpu) {
> >>> +	if (!vcpu->arch.debug_active)
> >>> +		return;
> >>> +
> >>> +	/* Disable all debug events and clead pending debug events */
> >>> +	mtspr(SPRN_DBCR0, 0x0);
> >>> +	kvmppc_clear_dbsr();
> >>> +
> >>> +	/*
> >>> +	 * Check whether guest still need debug resource, if not then there
> >>> +	 * is no need to restore guest context.
> >>> +	 */
> >>> +	if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> >>> +		return;
> >>> +
> >>> +	/* Load Guest Context */
> >>> +	mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> >>> +	mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
> >>> +CONFIG_KVM_E500MC
> >>> +	mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
> >>
> >> You need to make sure DBCR4 is 0 when you leave things back to normal
> >> user space. Otherwise guest debug can interfere with host debug.
> >
> >
> > ok
> >
> >>
> >>> +#endif
> >>> +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> >>> +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> >>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> >>> +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> >>> +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> >>> +#endif
> >>> +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> >>> +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> >>> +
> >>> +	/* Enable debug events after other debug registers restored */
> >>> +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
> >>
> >> All of the code above looks suspiciously similar to
> >> prime_debug_regs();. Can't we somehow reuse that?
> >
> > I think we can if
> > - Save thread->debug_regs in local data structure
> 
> Yes, it can even be on the stack.

Thread->denug_regs is not struct , so memcpy() will not work, it is assigning all registers one by one.

> 
> > - Load vcpu->arch->debug_regs in thread->debug_regs
> > - Call prime_debug_regs();
> > - Restore thread->debug_regs from local save values in first step
> 
> On heavyweight exit, based on the values on stack, yes.
> 
> >
> >>
> >>> +
> >>> int kvmppc_vcpu_run(struct kvm_run *kvm_run, struct kvm_vcpu *vcpu)
> >>> {
> >>> 	int ret, s;
> >>> @@ -693,11 +757,25 @@ int kvmppc_vcpu_run(struct kvm_run *kvm_run,
> >>> struct
> >> kvm_vcpu *vcpu)
> >>> 	kvmppc_load_guest_fp(vcpu);
> >>> #endif
> >>>
> >>> +	/*
> >>> +	 * Clear current->thread.dbcr0 so that kernel does not
> >>> +	 * restore h/w registers on context switch in vcpu running state.
> >>> +	 */
> >>> +	vcpu->arch.debug_active = 1;
> >>
> >> = true;
> >
> > Ok
> >
> >>
> >>> +	vcpu->arch.saved_dbcr0 = current->thread.dbcr0;
> >>> +	current->thread.dbcr0 = 0;
> >>> +	kvmppc_booke_vcpu_load_debug_regs(vcpu);
> >>
> >> static void switch_booke_debug_regs(struct thread_struct *new_thread)
> >> {
> >>        if ((current->thread.dbcr0 & DBCR0_IDM)
> >>                || (new_thread->dbcr0 & DBCR0_IDM))
> >>                        prime_debug_regs(new_thread); }
> >>
> >> The kernel will also restore debug state if the process we come from
> >> has debugging enabled. Please adjust the comment accordingly.
> >
> > kvmppc_booke_vcpu_load_debug_regs(vcpu); cleares DBSR and DBCR0, even if
> previous process have used debug registers. Is not that sufficient? I do not
> think we should load all other registers.
> 
> It's sufficient, but the comment is wrong.

ok

> 
> >
> >>
> >>> +
> >>> 	ret = __kvmppc_vcpu_run(kvm_run, vcpu);
> >>>
> >>> 	/* No need for kvm_guest_exit. It's done in handle_exit.
> >>> 	   We also get here with interrupts enabled. */
> >>>
> >>> +	/* Restore thread->dbcr0 */
> >>> +	vcpu->arch.debug_active = 0;
> >>> +	current->thread.dbcr0 = vcpu->arch.saved_dbcr0;
> >>> +	kvmppc_load_usespace_gebug();
> >>> +
> >>> #ifdef CONFIG_PPC_FPU
> >>> 	kvmppc_save_guest_fp(vcpu);
> >>>
> >>> @@ -753,6 +831,36 @@ static int emulation_exit(struct kvm_run *run,
> >>> struct
> >> kvm_vcpu *vcpu)
> >>> 	}
> >>> }
> >>>
> >>> +/*
> >>> + * Currently we do not support debug resource emulation to guest,
> >>> + * so always exit to user space irrespective of user space is
> >>> + * expecting the debug exception or not. This is unexpected event
> >>> + * and let us leave the action on user space.
> >>> + */
> >>> +static int kvmppc_handle_debug(struct kvm_run *run, struct kvm_vcpu
> >>> +*vcpu) {
> >>> +	u32 dbsr = mfspr(SPRN_DBSR);
> >>> +
> >>> +	kvmppc_clear_dbsr();
> >>> +	run->debug.arch.status = 0;
> >>> +	run->debug.arch.address = vcpu->arch.pc;
> >>> +
> >>> +	if (dbsr & (DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4)) {
> >>> +		run->debug.arch.status |= KVMPPC_DEBUG_BREAKPOINT;
> >>> +	} else {
> >>> +		if (dbsr & (DBSR_DAC1W | DBSR_DAC2W))
> >>> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_WRITE;
> >>> +		else if (dbsr & (DBSR_DAC1R | DBSR_DAC2R))
> >>> +			run->debug.arch.status |= KVMPPC_DEBUG_WATCH_READ;
> >>> +		if (dbsr & (DBSR_DAC1R | DBSR_DAC1W))
> >>> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[0];
> >>> +		else if (dbsr & (DBSR_DAC2R | DBSR_DAC2W))
> >>> +			run->debug.arch.address = vcpu->arch.shadow_dbg_reg.dac[1];
> >>> +	}
> >>> +
> >>> +	return RESUME_HOST;
> >>> +}
> >>> +
> >>> static void kvmppc_fill_pt_regs(struct pt_regs *regs) {
> >>> 	ulong r1, ip, msr, lr;
> >>> @@ -1112,18 +1220,10 @@ int kvmppc_handle_exit(struct kvm_run *run,
> >>> struct
> >> kvm_vcpu *vcpu,
> >>> 	}
> >>>
> >>> 	case BOOKE_INTERRUPT_DEBUG: {
> >>> -		u32 dbsr;
> >>> -
> >>> -		vcpu->arch.pc = mfspr(SPRN_CSRR0);
> >>> -
> >>> -		/* clear IAC events in DBSR register */
> >>> -		dbsr = mfspr(SPRN_DBSR);
> >>> -		dbsr &= DBSR_IAC1 | DBSR_IAC2 | DBSR_IAC3 | DBSR_IAC4;
> >>> -		mtspr(SPRN_DBSR, dbsr);
> >>> -
> >>> -		run->exit_reason = KVM_EXIT_DEBUG;
> >>> +		r = kvmppc_handle_debug(run, vcpu);
> >>> +		if (r == RESUME_HOST)
> >>> +			run->exit_reason = KVM_EXIT_DEBUG;
> >>> 		kvmppc_account_exit(vcpu, DEBUG_EXITS);
> >>> -		r = RESUME_HOST;
> >>> 		break;
> >>> 	}
> >>>
> >>> @@ -1174,7 +1274,7 @@ int kvm_arch_vcpu_setup(struct kvm_vcpu *vcpu)
> >>> 	kvmppc_set_msr(vcpu, 0);
> >>>
> >>> #ifndef CONFIG_KVM_BOOKE_HV
> >>> -	vcpu->arch.shadow_msr = MSR_USER | MSR_DE | MSR_IS | MSR_DS;
> >>> +	vcpu->arch.shadow_msr = MSR_USER | MSR_IS | MSR_DS;
> >>> 	vcpu->arch.shadow_pid = 1;
> >>> 	vcpu->arch.shared->msr = 0;
> >>> #endif
> >>> @@ -1529,12 +1629,6 @@ int kvm_vcpu_ioctl_set_one_reg(struct
> >>> kvm_vcpu *vcpu,
> >> struct kvm_one_reg *reg)
> >>> 	return r;
> >>> }
> >>>
> >>> -int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> >>> -					 struct kvm_guest_debug *dbg)
> >>> -{
> >>> -	return -EINVAL;
> >>> -}
> >>> -
> >>> int kvm_arch_vcpu_ioctl_get_fpu(struct kvm_vcpu *vcpu, struct
> >>> kvm_fpu *fpu) {
> >>> 	return -ENOTSUPP;
> >>> @@ -1640,16 +1734,128 @@ void kvmppc_decrementer_func(unsigned long data)
> >>> 	kvmppc_set_tsr_bits(vcpu, TSR_DIS); }
> >>>
> >>> +static void kvmppc_booke_vcpu_put_debug_regs(struct kvm_vcpu *vcpu)
> >>> +{
> >>> +	/* Disable all debug events First */
> >>
> >> first
> >>
> >>> +	mtspr(SPRN_DBCR0, 0x0);
> >>> +	/* Disable pending debug event by clearing DBSR */
> >>> +	kvmppc_clear_dbsr();
> >>
> >> kvmppc_handle_debug() happens with preemption enabled, no?
> >
> > Want to clarify,  preemption will be enabled on calling local_irq_enable(); in
> kvmppc_handle_exit()?
> 
> Yes. Implicitly :).
> 
> >
> >> So we can have a
> >> debug event that gets cleared on preempt by this.
> >
> > Should we read the DBSR in before local_irq_enable() in kvmppc_handle_exit()?
> 
> We have to, yes. Otherwise we could get preempted in between and get bogus data
> I presume.

Ok

Thanks
-Bharat

> 
> >
> >>
> >>> +}
> >>> +
> >>> +int kvm_arch_vcpu_ioctl_set_guest_debug(struct kvm_vcpu *vcpu,
> >>> +					 struct kvm_guest_debug *dbg)
> >>> +{
> >>> +	struct kvmppc_booke_debug_reg *dbg_reg;
> >>> +	int n, b = 0, w = 0;
> >>> +	const u32 bp_code[] = {
> >>> +		DBCR0_IAC1 | DBCR0_IDM,
> >>> +		DBCR0_IAC2 | DBCR0_IDM,
> >>> +		DBCR0_IAC3 | DBCR0_IDM,
> >>> +		DBCR0_IAC4 | DBCR0_IDM
> >>> +	};
> >>> +	const u32 wp_code[] = {
> >>> +		DBCR0_DAC1W | DBCR0_IDM,
> >>> +		DBCR0_DAC2W | DBCR0_IDM,
> >>> +		DBCR0_DAC1R | DBCR0_IDM,
> >>> +		DBCR0_DAC2R | DBCR0_IDM
> >>> +	};
> >>> +
> >>> +	if (!(dbg->control & KVM_GUESTDBG_ENABLE)) {
> >>> +		/* Clear All debug events */
> >>> +		vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> >>> +		vcpu->guest_debug = 0;
> >>> +#ifdef CONFIG_KVM_BOOKE_HV
> >>> +		/*
> >>> +		 * When user space is not using the debug resources
> >>> +		 * then allow guest to change the MSR.DE.
> >>> +		 */
> >>> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP; #endif
> >>> +		return 0;
> >>> +	}
> >>> +
> >>> +#ifdef CONFIG_KVM_BOOKE_HV
> >>> +		/*
> >>> +		 * When user space is using the debug resource then
> >>> +		 * do not allow guest to change the MSR.DE.
> >>> +		 */
> >>> +		vcpu->arch.shadow_msrp &= ~MSRP_DEP;
> >>
> >> This is supposed to be |= right?
> >
> > Yes :-)
> >
> >>
> >>> +#endif
> >>> +	vcpu->guest_debug = dbg->control;
> >>> +	vcpu->arch.shadow_dbg_reg.dbcr0 = 0;
> >>> +	/* Set DBCR0_EDM in guest visible DBCR0 register. */
> >>> +	vcpu->arch.dbg_reg.dbcr0 = DBCR0_EDM;
> >>> +
> >>> +	if (vcpu->guest_debug & KVM_GUESTDBG_SINGLESTEP)
> >>> +		vcpu->arch.shadow_dbg_reg.dbcr0 |= DBCR0_IDM | DBCR0_IC;
> >>> +
> >>> +	if (!(vcpu->guest_debug & KVM_GUESTDBG_USE_HW_BP))
> >>> +		/* Code below handles only HW breakpoints */
> >>
> >> Please move the comment out of the one-lined branch. Just put it
> >> below the return.
> >
> > Ok
> >
> >>
> >>> +		return 0;
> >>> +
> >>> +	dbg_reg = &(vcpu->arch.shadow_dbg_reg);
> >>> +
> >>> +	/*
> >>> +	 * On BOOKE (e500v2); Set DBCR1 and DBCR2 to allow debug events
> >>> +	 * to occur when MSR.PR is set.
> >>> +	 * On BOOKE-HV (e500mc+); MSR.PR = 0 when guest is running. So we
> >>> +	 * should clear DBCR1 and DBCR2. And EPCR.DUVD is used to control
> >>> +	 * that debug events will not come in hypervisor (GS = 0).
> >>> +	 */
> >>> +#ifdef CONFIG_KVM_BOOKE_HV
> >>> +	dbg_reg->dbcr1 = 0;
> >>> +	dbg_reg->dbcr2 = 0;
> >>> +#else
> >>> +	dbg_reg->dbcr1 = DBCR1_IAC1US | DBCR1_IAC2US | DBCR1_IAC3US |
> >>> +			  DBCR1_IAC4US;
> >>> +	dbg_reg->dbcr2 = DBCR2_DAC1US | DBCR2_DAC2US; #endif
> >>> +
> >>> +	for (n = 0; n < (KVMPPC_BOOKE_IAC_NUM + KVMPPC_BOOKE_DAC_NUM); n++) {
> >>> +		u32 type = dbg->arch.bp[n].type;
> >>> +
> >>> +		if (!type)
> >>> +			continue;
> >>
> >> Scott's comment on why a zero-type is different from any other
> >> invalid type is still outstanding I think.
> >
> > Userspce does following
> > - dbg->arch.bp[] is first zero initialized. So dbg->arch.bp[n].type is
> > 0 (KVMPPC_DEBUG_NONE)
> > - Then set following for valid breakpoints/watchpoints:
> >   -- dbg->arch.bp[n].type (KVMPPC_DEBUG_BREAKPOINT or KVMPPC_DEBUG_WATCH_WRITE
> or KVMPPC_DEBUG_WATCH_READ)
> >   -- dbg->arch.bp[n].addr
> >
> > I tried to avoid loop when type is 0 (probably saying type ==
> KVMPPC_DEBUG_NONE should be more clear).
> 
> That works, yes.
> 
> > if type is invalid then should we return -EINVAL.
> 
> Yes.
> 
> 
> Alex
> 


--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Alexander Graf May 2, 2013, 2:04 p.m. UTC | #5

On 02.05.2013, at 16:00, Bhushan Bharat-R65777 wrote:

> 
> 
>> -----Original Message-----
>> From: Alexander Graf [mailto:agraf@suse.de]
>> Sent: Thursday, May 02, 2013 4:35 PM
>> To: Bhushan Bharat-R65777
>> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
>> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
>> 
>> 
>> On 02.05.2013, at 11:46, Bhushan Bharat-R65777 wrote:

[...]

>>> 
>>>> 
>>>>> +#endif
>>>>> +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
>>>>> +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
>>>>> +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
>>>>> +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
>>>>> +#endif
>>>>> +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
>>>>> +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
>>>>> +
>>>>> +	/* Enable debug events after other debug registers restored */
>>>>> +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
>>>> 
>>>> All of the code above looks suspiciously similar to
>>>> prime_debug_regs();. Can't we somehow reuse that?
>>> 
>>> I think we can if
>>> - Save thread->debug_regs in local data structure
>> 
>> Yes, it can even be on the stack.
> 
> Thread->denug_regs is not struct , so memcpy() will not work, it is assigning all registers one by one.

We could make it a struct, no? Then it's a matter of a = b; :)


Alex

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Bharat Bhushan May 3, 2013, 10:48 a.m. UTC | #6

> >>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
> >>> +*vcpu) {
> >>> +	if (!vcpu->arch.debug_active)
> >>> +		return;
> >>> +
> >>> +	/* Disable all debug events and clead pending debug events */
> >>> +	mtspr(SPRN_DBCR0, 0x0);
> >>> +	kvmppc_clear_dbsr();
> >>> +
> >>> +	/*
> >>> +	 * Check whether guest still need debug resource, if not then there
> >>> +	 * is no need to restore guest context.
> >>> +	 */
> >>> +	if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> >>> +		return;
> >>> +
> >>> +	/* Load Guest Context */
> >>> +	mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> >>> +	mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
> >>> +CONFIG_KVM_E500MC
> >>> +	mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
> >>
> >> You need to make sure DBCR4 is 0 when you leave things back to normal
> >> user space. Otherwise guest debug can interfere with host debug.
> >
> >
> > ok
> >
> >>
> >>> +#endif
> >>> +	mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> >>> +	mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> >>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> >>> +	mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> >>> +	mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> >>> +#endif
> >>> +	mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> >>> +	mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> >>> +
> >>> +	/* Enable debug events after other debug registers restored */
> >>> +	mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
> >>
> >> All of the code above looks suspiciously similar to
> >> prime_debug_regs();. Can't we somehow reuse that?
> >
> > I think we can if
> > - Save thread->debug_regs in local data structure
> 
> Yes, it can even be on the stack.
> 
> > - Load vcpu->arch->debug_regs in thread->debug_regs
> > - Call prime_debug_regs();
> > - Restore thread->debug_regs from local save values in first step
> 
> On heavyweight exit, based on the values on stack, yes.

This is how I think we can save/restore debug context. Please correct if I am missing something.

1) When QEMU is running

-> thread->debug_reg == QEMU debug register context.
-> Kernel will handle switching the debug register on context switch.
-> no vcpu_load() called

2) QEMU makes ioctls (except RUN) 
 -> This will call vcpu_load()
 -> should not change context.
 -> Some ioctls can change vcpu debug register, context saved in vcpu->debug_regs

3) QEMU Makes RUN ioctl
 -> Save thread->debug_reg on STACK
 -> Store thread->debug_reg == vcpu->debug_reg
 -> load thread->debug_reg 
 -> RUN VCPU ( So thread points to vcpu context )

4) Context switch happens When VCPU running
 -> makes vcpu_load() should not load any context
 -> kernel loads the vcpu context as thread->debug_regs points to vcpu context.

5) On heavyweight_exit
 -> Load the context saved on stack in thread->debug_reg

Thanks
-Bharat



--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Alexander Graf May 3, 2013, 11:08 a.m. UTC | #7

Am 03.05.2013 um 12:48 schrieb Bhushan Bharat-R65777 <R65777@freescale.com>:

>>>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
>>>>> +*vcpu) {
>>>>> +    if (!vcpu->arch.debug_active)
>>>>> +        return;
>>>>> +
>>>>> +    /* Disable all debug events and clead pending debug events */
>>>>> +    mtspr(SPRN_DBCR0, 0x0);
>>>>> +    kvmppc_clear_dbsr();
>>>>> +
>>>>> +    /*
>>>>> +     * Check whether guest still need debug resource, if not then there
>>>>> +     * is no need to restore guest context.
>>>>> +     */
>>>>> +    if (!vcpu->arch.shadow_dbg_reg.dbcr0)
>>>>> +        return;
>>>>> +
>>>>> +    /* Load Guest Context */
>>>>> +    mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
>>>>> +    mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
>>>>> +CONFIG_KVM_E500MC
>>>>> +    mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
>>>> 
>>>> You need to make sure DBCR4 is 0 when you leave things back to normal
>>>> user space. Otherwise guest debug can interfere with host debug.
>>> 
>>> 
>>> ok
>>> 
>>>> 
>>>>> +#endif
>>>>> +    mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
>>>>> +    mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
>>>>> +    mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
>>>>> +    mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
>>>>> +#endif
>>>>> +    mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
>>>>> +    mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
>>>>> +
>>>>> +    /* Enable debug events after other debug registers restored */
>>>>> +    mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
>>>> 
>>>> All of the code above looks suspiciously similar to
>>>> prime_debug_regs();. Can't we somehow reuse that?
>>> 
>>> I think we can if
>>> - Save thread->debug_regs in local data structure
>> 
>> Yes, it can even be on the stack.
>> 
>>> - Load vcpu->arch->debug_regs in thread->debug_regs
>>> - Call prime_debug_regs();
>>> - Restore thread->debug_regs from local save values in first step
>> 
>> On heavyweight exit, based on the values on stack, yes.
> 
> This is how I think we can save/restore debug context. Please correct if I am missing something.

Sounds about right :)

Alex

> 
> 1) When QEMU is running
> 
> -> thread->debug_reg == QEMU debug register context.
> -> Kernel will handle switching the debug register on context switch.
> -> no vcpu_load() called
> 
> 2) QEMU makes ioctls (except RUN) 
> -> This will call vcpu_load()
> -> should not change context.
> -> Some ioctls can change vcpu debug register, context saved in vcpu->debug_regs
> 
> 3) QEMU Makes RUN ioctl
> -> Save thread->debug_reg on STACK
> -> Store thread->debug_reg == vcpu->debug_reg
> -> load thread->debug_reg 
> -> RUN VCPU ( So thread points to vcpu context )
> 
> 4) Context switch happens When VCPU running
> -> makes vcpu_load() should not load any context
> -> kernel loads the vcpu context as thread->debug_regs points to vcpu context.
> 
> 5) On heavyweight_exit
> -> Load the context saved on stack in thread->debug_reg
> 
> Thanks
> -Bharat
> 
> 
> 
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Alexander Graf May 3, 2013, 12:29 p.m. UTC | #8

On 03.05.2013, at 13:08, Alexander Graf wrote:

> 
> 
> Am 03.05.2013 um 12:48 schrieb Bhushan Bharat-R65777 <R65777@freescale.com>:
> 
>>>>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
>>>>>> +*vcpu) {
>>>>>> +    if (!vcpu->arch.debug_active)
>>>>>> +        return;
>>>>>> +
>>>>>> +    /* Disable all debug events and clead pending debug events */
>>>>>> +    mtspr(SPRN_DBCR0, 0x0);
>>>>>> +    kvmppc_clear_dbsr();
>>>>>> +
>>>>>> +    /*
>>>>>> +     * Check whether guest still need debug resource, if not then there
>>>>>> +     * is no need to restore guest context.
>>>>>> +     */
>>>>>> +    if (!vcpu->arch.shadow_dbg_reg.dbcr0)
>>>>>> +        return;
>>>>>> +
>>>>>> +    /* Load Guest Context */
>>>>>> +    mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
>>>>>> +    mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
>>>>>> +CONFIG_KVM_E500MC
>>>>>> +    mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
>>>>> 
>>>>> You need to make sure DBCR4 is 0 when you leave things back to normal
>>>>> user space. Otherwise guest debug can interfere with host debug.
>>>> 
>>>> 
>>>> ok
>>>> 
>>>>> 
>>>>>> +#endif
>>>>>> +    mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
>>>>>> +    mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
>>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
>>>>>> +    mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
>>>>>> +    mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
>>>>>> +#endif
>>>>>> +    mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
>>>>>> +    mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
>>>>>> +
>>>>>> +    /* Enable debug events after other debug registers restored */
>>>>>> +    mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
>>>>> 
>>>>> All of the code above looks suspiciously similar to
>>>>> prime_debug_regs();. Can't we somehow reuse that?
>>>> 
>>>> I think we can if
>>>> - Save thread->debug_regs in local data structure
>>> 
>>> Yes, it can even be on the stack.
>>> 
>>>> - Load vcpu->arch->debug_regs in thread->debug_regs
>>>> - Call prime_debug_regs();
>>>> - Restore thread->debug_regs from local save values in first step
>>> 
>>> On heavyweight exit, based on the values on stack, yes.
>> 
>> This is how I think we can save/restore debug context. Please correct if I am missing something.
> 
> Sounds about right :)

Actually, what happens if a guest breakpoint is set to a kernel address that happens to be within the scope of kvm code? We do accept debug events between vcpu_run and the assembly code, right?


Alex

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Bharat Bhushan May 3, 2013, 1:11 p.m. UTC | #9

> -----Original Message-----
> From: Alexander Graf [mailto:agraf@suse.de]
> Sent: Friday, May 03, 2013 6:00 PM
> To: Bhushan Bharat-R65777
> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
> 
> 
> On 03.05.2013, at 13:08, Alexander Graf wrote:
> 
> >
> >
> > Am 03.05.2013 um 12:48 schrieb Bhushan Bharat-R65777 <R65777@freescale.com>:
> >
> >>>>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
> >>>>>> +*vcpu) {
> >>>>>> +    if (!vcpu->arch.debug_active)
> >>>>>> +        return;
> >>>>>> +
> >>>>>> +    /* Disable all debug events and clead pending debug events */
> >>>>>> +    mtspr(SPRN_DBCR0, 0x0);
> >>>>>> +    kvmppc_clear_dbsr();
> >>>>>> +
> >>>>>> +    /*
> >>>>>> +     * Check whether guest still need debug resource, if not then there
> >>>>>> +     * is no need to restore guest context.
> >>>>>> +     */
> >>>>>> +    if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> >>>>>> +        return;
> >>>>>> +
> >>>>>> +    /* Load Guest Context */
> >>>>>> +    mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> >>>>>> +    mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
> >>>>>> +CONFIG_KVM_E500MC
> >>>>>> +    mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
> >>>>>
> >>>>> You need to make sure DBCR4 is 0 when you leave things back to
> >>>>> normal user space. Otherwise guest debug can interfere with host debug.
> >>>>
> >>>>
> >>>> ok
> >>>>
> >>>>>
> >>>>>> +#endif
> >>>>>> +    mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> >>>>>> +    mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> >>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> >>>>>> +    mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> >>>>>> +    mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> >>>>>> +#endif
> >>>>>> +    mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> >>>>>> +    mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> >>>>>> +
> >>>>>> +    /* Enable debug events after other debug registers restored */
> >>>>>> +    mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
> >>>>>
> >>>>> All of the code above looks suspiciously similar to
> >>>>> prime_debug_regs();. Can't we somehow reuse that?
> >>>>
> >>>> I think we can if
> >>>> - Save thread->debug_regs in local data structure
> >>>
> >>> Yes, it can even be on the stack.
> >>>
> >>>> - Load vcpu->arch->debug_regs in thread->debug_regs
> >>>> - Call prime_debug_regs();
> >>>> - Restore thread->debug_regs from local save values in first step
> >>>
> >>> On heavyweight exit, based on the values on stack, yes.
> >>
> >> This is how I think we can save/restore debug context. Please correct if I am
> missing something.
> >
> > Sounds about right :)
> 
> Actually, what happens if a guest breakpoint is set to a kernel address that
> happens to be within the scope of kvm code?

You mean address of kvm code in guest or host?

If host, we already mentioned that we do not support that. Right?

-Bharat

> We do accept debug events between
> vcpu_run and the assembly code, right?
> 
> 
> Alex
> 


--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Alexander Graf May 3, 2013, 1:18 p.m. UTC | #10

On 03.05.2013, at 15:11, Bhushan Bharat-R65777 wrote:

> 
> 
>> -----Original Message-----
>> From: Alexander Graf [mailto:agraf@suse.de]
>> Sent: Friday, May 03, 2013 6:00 PM
>> To: Bhushan Bharat-R65777
>> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
>> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
>> 
>> 
>> On 03.05.2013, at 13:08, Alexander Graf wrote:
>> 
>>> 
>>> 
>>> Am 03.05.2013 um 12:48 schrieb Bhushan Bharat-R65777 <R65777@freescale.com>:
>>> 
>>>>>>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
>>>>>>>> +*vcpu) {
>>>>>>>> +    if (!vcpu->arch.debug_active)
>>>>>>>> +        return;
>>>>>>>> +
>>>>>>>> +    /* Disable all debug events and clead pending debug events */
>>>>>>>> +    mtspr(SPRN_DBCR0, 0x0);
>>>>>>>> +    kvmppc_clear_dbsr();
>>>>>>>> +
>>>>>>>> +    /*
>>>>>>>> +     * Check whether guest still need debug resource, if not then there
>>>>>>>> +     * is no need to restore guest context.
>>>>>>>> +     */
>>>>>>>> +    if (!vcpu->arch.shadow_dbg_reg.dbcr0)
>>>>>>>> +        return;
>>>>>>>> +
>>>>>>>> +    /* Load Guest Context */
>>>>>>>> +    mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
>>>>>>>> +    mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
>>>>>>>> +CONFIG_KVM_E500MC
>>>>>>>> +    mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
>>>>>>> 
>>>>>>> You need to make sure DBCR4 is 0 when you leave things back to
>>>>>>> normal user space. Otherwise guest debug can interfere with host debug.
>>>>>> 
>>>>>> 
>>>>>> ok
>>>>>> 
>>>>>>> 
>>>>>>>> +#endif
>>>>>>>> +    mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
>>>>>>>> +    mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
>>>>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
>>>>>>>> +    mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
>>>>>>>> +    mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
>>>>>>>> +#endif
>>>>>>>> +    mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
>>>>>>>> +    mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
>>>>>>>> +
>>>>>>>> +    /* Enable debug events after other debug registers restored */
>>>>>>>> +    mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
>>>>>>> 
>>>>>>> All of the code above looks suspiciously similar to
>>>>>>> prime_debug_regs();. Can't we somehow reuse that?
>>>>>> 
>>>>>> I think we can if
>>>>>> - Save thread->debug_regs in local data structure
>>>>> 
>>>>> Yes, it can even be on the stack.
>>>>> 
>>>>>> - Load vcpu->arch->debug_regs in thread->debug_regs
>>>>>> - Call prime_debug_regs();
>>>>>> - Restore thread->debug_regs from local save values in first step
>>>>> 
>>>>> On heavyweight exit, based on the values on stack, yes.
>>>> 
>>>> This is how I think we can save/restore debug context. Please correct if I am
>> missing something.
>>> 
>>> Sounds about right :)
>> 
>> Actually, what happens if a guest breakpoint is set to a kernel address that
>> happens to be within the scope of kvm code?
> 
> You mean address of kvm code in guest or host?
> 
> If host, we already mentioned that we do not support that. Right?

QEMU wants to debug the guest at address 0xc0000123. kvm_run happens to be at that address. We switch the debug registers through prime_debug_regs. Will the host kernel receive a debug interrupt when it runs kvm_run()?


Alex

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Bharat Bhushan May 6, 2013, 5:51 a.m. UTC | #11

> -----Original Message-----
> From: kvm-owner@vger.kernel.org [mailto:kvm-owner@vger.kernel.org] On Behalf Of
> Alexander Graf
> Sent: Friday, May 03, 2013 6:48 PM
> To: Bhushan Bharat-R65777
> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub support
> 
> 
> On 03.05.2013, at 15:11, Bhushan Bharat-R65777 wrote:
> 
> >
> >
> >> -----Original Message-----
> >> From: Alexander Graf [mailto:agraf@suse.de]
> >> Sent: Friday, May 03, 2013 6:00 PM
> >> To: Bhushan Bharat-R65777
> >> Cc: kvm-ppc@vger.kernel.org; kvm@vger.kernel.org; Wood Scott-B07421
> >> Subject: Re: [PATCH 7/7 v3] KVM: PPC: Add userspace debug stub
> >> support
> >>
> >>
> >> On 03.05.2013, at 13:08, Alexander Graf wrote:
> >>
> >>>
> >>>
> >>> Am 03.05.2013 um 12:48 schrieb Bhushan Bharat-R65777 <R65777@freescale.com>:
> >>>
> >>>>>>>> +static void kvmppc_booke_vcpu_load_debug_regs(struct kvm_vcpu
> >>>>>>>> +*vcpu) {
> >>>>>>>> +    if (!vcpu->arch.debug_active)
> >>>>>>>> +        return;
> >>>>>>>> +
> >>>>>>>> +    /* Disable all debug events and clead pending debug events */
> >>>>>>>> +    mtspr(SPRN_DBCR0, 0x0);
> >>>>>>>> +    kvmppc_clear_dbsr();
> >>>>>>>> +
> >>>>>>>> +    /*
> >>>>>>>> +     * Check whether guest still need debug resource, if not then
> there
> >>>>>>>> +     * is no need to restore guest context.
> >>>>>>>> +     */
> >>>>>>>> +    if (!vcpu->arch.shadow_dbg_reg.dbcr0)
> >>>>>>>> +        return;
> >>>>>>>> +
> >>>>>>>> +    /* Load Guest Context */
> >>>>>>>> +    mtspr(SPRN_DBCR1, vcpu->arch.shadow_dbg_reg.dbcr1);
> >>>>>>>> +    mtspr(SPRN_DBCR2, vcpu->arch.shadow_dbg_reg.dbcr2); #ifdef
> >>>>>>>> +CONFIG_KVM_E500MC
> >>>>>>>> +    mtspr(SPRN_DBCR4, vcpu->arch.shadow_dbg_reg.dbcr4);
> >>>>>>>
> >>>>>>> You need to make sure DBCR4 is 0 when you leave things back to
> >>>>>>> normal user space. Otherwise guest debug can interfere with host debug.
> >>>>>>
> >>>>>>
> >>>>>> ok
> >>>>>>
> >>>>>>>
> >>>>>>>> +#endif
> >>>>>>>> +    mtspr(SPRN_IAC1, vcpu->arch.shadow_dbg_reg.iac[0]);
> >>>>>>>> +    mtspr(SPRN_IAC2, vcpu->arch.shadow_dbg_reg.iac[1]);
> >>>>>>>> +#if CONFIG_PPC_ADV_DEBUG_IACS > 2
> >>>>>>>> +    mtspr(SPRN_IAC3, vcpu->arch.shadow_dbg_reg.iac[2]);
> >>>>>>>> +    mtspr(SPRN_IAC4, vcpu->arch.shadow_dbg_reg.iac[3]);
> >>>>>>>> +#endif
> >>>>>>>> +    mtspr(SPRN_DAC1, vcpu->arch.shadow_dbg_reg.dac[0]);
> >>>>>>>> +    mtspr(SPRN_DAC2, vcpu->arch.shadow_dbg_reg.dac[1]);
> >>>>>>>> +
> >>>>>>>> +    /* Enable debug events after other debug registers restored */
> >>>>>>>> +    mtspr(SPRN_DBCR0, vcpu->arch.shadow_dbg_reg.dbcr0); }
> >>>>>>>
> >>>>>>> All of the code above looks suspiciously similar to
> >>>>>>> prime_debug_regs();. Can't we somehow reuse that?
> >>>>>>
> >>>>>> I think we can if
> >>>>>> - Save thread->debug_regs in local data structure
> >>>>>
> >>>>> Yes, it can even be on the stack.
> >>>>>
> >>>>>> - Load vcpu->arch->debug_regs in thread->debug_regs
> >>>>>> - Call prime_debug_regs();
> >>>>>> - Restore thread->debug_regs from local save values in first step
> >>>>>
> >>>>> On heavyweight exit, based on the values on stack, yes.
> >>>>
> >>>> This is how I think we can save/restore debug context. Please
> >>>> correct if I am
> >> missing something.
> >>>
> >>> Sounds about right :)
> >>
> >> Actually, what happens if a guest breakpoint is set to a kernel
> >> address that happens to be within the scope of kvm code?
> >
> > You mean address of kvm code in guest or host?
> >
> > If host, we already mentioned that we do not support that. Right?
> 
> QEMU wants to debug the guest at address 0xc0000123. kvm_run happens to be at
> that address. We switch the debug registers through prime_debug_regs. Will the
> host kernel receive a debug interrupt when it runs kvm_run()?

No,
On e500v2, we uses DBCR1 and DBCR2 to not allow debug events when MSR.PR = 0
On e500mc+, we uses EPCR.DUVD to not allow debug events when in hypervisor mode.

-Bharat

> 
> 
> Alex
> 
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a
> message to majordomo@vger.kernel.org More majordomo info at
> http://vger.kernel.org/majordomo-info.html


--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[7/7,v3] KVM: PPC: Add userspace debug stub support

Commit Message

Comments

Patch