linux-kernel - Re: [PATCH v2 1/1] powerpc/kvm: Save Timebase Offset to fix sched

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <10b5e7e55f18c14567d4076e76c204c8805b33c8.camel@gmail.com>
Date:   Fri, 05 Feb 2021 17:35:12 -0300
From:   Leonardo Bras <leobras.c@...il.com>
To:     Fabiano Rosas <farosas@...ux.ibm.com>,
        Paul Mackerras <paulus@...abs.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        Benjamin Herrenschmidt <benh@...nel.crashing.org>,
        Christophe Leroy <christophe.leroy@...roup.eu>,
        Athira Rajeev <atrajeev@...ux.vnet.ibm.com>,
        "Aneesh Kumar K.V" <aneesh.kumar@...ux.ibm.com>,
        Jordan Niethe <jniethe5@...il.com>,
        Nicholas Piggin <npiggin@...il.com>,
        Frederic Weisbecker <frederic@...nel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Geert Uytterhoeven <geert+renesas@...der.be>
Cc:     kvm-ppc@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/1] powerpc/kvm: Save Timebase Offset to fix
 sched_clock() while running guest code.

Hello Fabiano, 
Thanks for reviewing! 
(answers inline)

On Fri, 2021-02-05 at 10:09 -0300, Fabiano Rosas wrote:
> Leonardo Bras <leobras.c@...il.com> writes:
> 
> > Before guest entry, TBU40 register is changed to reflect guest timebase.
> > After exitting guest, the register is reverted to it's original value.
> > 
> > If one tries to get the timestamp from host between those changes, it
> > will present an incorrect value.
> > 
> > An example would be trying to add a tracepoint in
> > kvmppc_guest_entry_inject_int(), which depending on last tracepoint
> > acquired could actually cause the host to crash.
> > 
> > Save the Timebase Offset to PACA and use it on sched_clock() to always
> > get the correct timestamp.
> > 
> > Signed-off-by: Leonardo Bras <leobras.c@...il.com>
> > Suggested-by: Paul Mackerras <paulus@...abs.org>
> > ---
> > Changes since v1:
> > - Subtracts offset only when CONFIG_KVM_BOOK3S_HANDLER and
> >   CONFIG_PPC_BOOK3S_64 are defined.
> > ---
> >  arch/powerpc/include/asm/kvm_book3s_asm.h | 1 +
> >  arch/powerpc/kernel/asm-offsets.c         | 1 +
> >  arch/powerpc/kernel/time.c                | 8 +++++++-
> >  arch/powerpc/kvm/book3s_hv.c              | 2 ++
> >  arch/powerpc/kvm/book3s_hv_rmhandlers.S   | 2 ++
> >  5 files changed, 13 insertions(+), 1 deletion(-)
> > 
> > diff --git a/arch/powerpc/include/asm/kvm_book3s_asm.h b/arch/powerpc/include/asm/kvm_book3s_asm.h
> > index 078f4648ea27..e2c12a10eed2 100644
> > --- a/arch/powerpc/include/asm/kvm_book3s_asm.h
> > +++ b/arch/powerpc/include/asm/kvm_book3s_asm.h
> > @@ -131,6 +131,7 @@ struct kvmppc_host_state {
> >  	u64 cfar;
> >  	u64 ppr;
> >  	u64 host_fscr;
> > +	u64 tb_offset;		/* Timebase offset: keeps correct
> > timebase while on guest */
> 
> Couldn't you use the vc->tb_offset_applied for this? We have a reference
> for the vcore in the hstate already.

But it's a pointer, which means we would have to keep checking for NULL
every time we need sched_clock(). 
Potentially it would cost a cache miss for PACA memory region that
contain vc, another for getting the part of *vc that contains the
tb_offset_applied, instead of only one for PACA struct region that
contains tb_offset.

On the other hand, it got me thinking: If the offset is applied per
cpu, why don't we get this info only in PACA, instead of in vc?
It could be a general way to get an offset applied for any purpose and
still get the sched_clock() right. 
(Not that I have any idea of any other purpose we could use it) 

Best regards!
Leonardo Bras

> 
> >  #endif
> >  };
> > 
> > diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
> > index b12d7c049bfe..0beb8fdc6352 100644
> > --- a/arch/powerpc/kernel/asm-offsets.c
> > +++ b/arch/powerpc/kernel/asm-offsets.c
> > @@ -706,6 +706,7 @@ int main(void)
> >  	HSTATE_FIELD(HSTATE_CFAR, cfar);
> >  	HSTATE_FIELD(HSTATE_PPR, ppr);
> >  	HSTATE_FIELD(HSTATE_HOST_FSCR, host_fscr);
> > +	HSTATE_FIELD(HSTATE_TB_OFFSET, tb_offset);
> >  #endif /* CONFIG_PPC_BOOK3S_64 */
> > 
> >  #else /* CONFIG_PPC_BOOK3S */
> > diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
> > index 67feb3524460..f27f0163792b 100644
> > --- a/arch/powerpc/kernel/time.c
> > +++ b/arch/powerpc/kernel/time.c
> > @@ -699,7 +699,13 @@ EXPORT_SYMBOL_GPL(tb_to_ns);
> >   */
> >  notrace unsigned long long sched_clock(void)
> >  {
> > -	return mulhdu(get_tb() - boot_tb, tb_to_ns_scale) << tb_to_ns_shift;
> > +	u64 tb = get_tb() - boot_tb;
> > +
> > +#if defined(CONFIG_PPC_BOOK3S_64) && defined(CONFIG_KVM_BOOK3S_HANDLER)
> > +	tb -= local_paca->kvm_hstate.tb_offset;
> > +#endif
> > +
> > +	return mulhdu(tb, tb_to_ns_scale) << tb_to_ns_shift;
> >  }
> > 
> > 
> > diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
> > index b3731572295e..c08593c63353 100644
> > --- a/arch/powerpc/kvm/book3s_hv.c
> > +++ b/arch/powerpc/kvm/book3s_hv.c
> > @@ -3491,6 +3491,7 @@ static int kvmhv_load_hv_regs_and_go(struct kvm_vcpu *vcpu, u64 time_limit,
> >  		if ((tb & 0xffffff) < (new_tb & 0xffffff))
> >  			mtspr(SPRN_TBU40, new_tb + 0x1000000);
> >  		vc->tb_offset_applied = vc->tb_offset;
> > +		local_paca->kvm_hstate.tb_offset = vc->tb_offset;
> >  	}
> > 
> >  	if (vc->pcr)
> > @@ -3594,6 +3595,7 @@ static int kvmhv_load_hv_regs_and_go(struct kvm_vcpu *vcpu, u64 time_limit,
> >  		if ((tb & 0xffffff) < (new_tb & 0xffffff))
> >  			mtspr(SPRN_TBU40, new_tb + 0x1000000);
> >  		vc->tb_offset_applied = 0;
> > +		local_paca->kvm_hstate.tb_offset = 0;
> >  	}
> > 
> >  	mtspr(SPRN_HDEC, 0x7fffffff);
> > diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> > index b73140607875..8f7a9f7f4ee6 100644
> > --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> > +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
> > @@ -632,6 +632,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300)
> >  	cmpdi	r8,0
> >  	beq	37f
> >  	std	r8, VCORE_TB_OFFSET_APPL(r5)
> > +	std	r8, HSTATE_TB_OFFSET(r13)
> >  	mftb	r6		/* current host timebase */
> >  	add	r8,r8,r6
> >  	mtspr	SPRN_TBU40,r8	/* update upper 40 bits */
> > @@ -1907,6 +1908,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
> >  	beq	17f
> >  	li	r0, 0
> >  	std	r0, VCORE_TB_OFFSET_APPL(r5)
> > +	std	r0, HSTATE_TB_OFFSET(r13)
> >  	mftb	r6			/* current guest timebase */
> >  	subf	r8,r8,r6
> >  	mtspr	SPRN_TBU40,r8		/* update upper 40 bits */