From: Julien Grall <julien.grall@arm.com> To: Andrii Anisov <andrii.anisov@gmail.com>, xen-devel@lists.xen.org Cc: "Stefano Stabellini" <sstabellini@kernel.org>, "Andrii Anisov" <andrii_anisov@epam.com>, "Konrad Rzeszutek Wilk" <konrad.wilk@oracle.com>, "George Dunlap" <George.Dunlap@eu.citrix.com>, "Andrew Cooper" <andrew.cooper3@citrix.com>, "Ian Jackson" <ian.jackson@eu.citrix.com>, "Tim Deegan" <tim@xen.org>, "Jan Beulich" <JBeulich@suse.com>, xen-devel@lists.xenproject.org, "Wei Liu" <wei.liu2@citrix.com>, "Roger Pau Monné" <roger.pau@citrix.com> Subject: Re: [PATCH v2 2/2] xen: implement VCPUOP_register_runstate_phys_memory_area Date: Tue, 14 May 2019 10:58:24 +0100 [thread overview] Message-ID: <24f3e2b1-4889-1780-0e48-0c909f555997@arm.com> (raw) In-Reply-To: <959c8975-ec44-a788-25b3-84ee17520abc@gmail.com> On 13/05/2019 13:30, Andrii Anisov wrote: > > > On 08.05.19 18:40, Julien Grall wrote: >>> diff --git a/xen/arch/arm/domain.c b/xen/arch/arm/domain.c >>> index 6dc633e..8e24e63 100644 >> >>> { >>> - void __user *guest_handle = NULL; >>> + if ( !guest_handle_is_null(runstate_guest(v)) ) >>> + { >>> + void __user *guest_handle = NULL; >>> + if ( VM_ASSIST(v->domain, runstate_update_flag) ) >>> + { >>> + guest_handle = &v->runstate_guest.p->state_entry_time + 1; >>> + guest_handle--; >>> + v->runstate.state_entry_time |= XEN_RUNSTATE_UPDATE; >>> + __raw_copy_to_guest(guest_handle, >>> + (void *)(&v->runstate.state_entry_time + 1) >>> - 1, >>> + 1); >>> + smp_wmb(); >>> + } >>> - if ( guest_handle_is_null(runstate_guest(v)) ) >>> - return; >>> + __copy_to_guest(runstate_guest(v), &v->runstate, 1); >>> - if ( VM_ASSIST(v->domain, runstate_update_flag) ) >>> - { >>> - guest_handle = &v->runstate_guest.p->state_entry_time + 1; >>> - guest_handle--; >>> - v->runstate.state_entry_time |= XEN_RUNSTATE_UPDATE; >>> - __raw_copy_to_guest(guest_handle, >>> - (void *)(&v->runstate.state_entry_time + 1) - 1, >>> 1); >>> - smp_wmb(); >>> + if ( guest_handle ) >>> + { >>> + v->runstate.state_entry_time &= ~XEN_RUNSTATE_UPDATE; >>> + smp_wmb(); >>> + __raw_copy_to_guest(guest_handle, >>> + (void *)(&v->runstate.state_entry_time + 1) >>> - 1, >>> + 1); >>> + } >>> } >>> - __copy_to_guest(runstate_guest(v), &v->runstate, 1); >>> - >>> - if ( guest_handle ) >>> + spin_lock(&v->mapped_runstate_lock); >>> + if ( v->mapped_runstate ) >> >> The code looks a bit odd to me, you seem to allow a guest to provide 2 >> runstate areas: one using guest virtual address the other using guest physical >> address. >> >> It would be best if we prevent a guest to mix match them. > > Firstly I turned to implementing in that way, but the locking and decissions > code become really ugly and complex while trying to cover 'guest's misbehavior' > scenarios. I think it is possible to have a simple version taking the decision on which method to use. You can either use the spin_lock to protect everything or use something like: update_runstate_area(): if ( xchg(&v->runstate_in_use, 1) ) return; switch ( v->runstate_type ) { GVADDR: update_runstate_by_gvaddr(); GPADDR: update_runstate_by_gpaddr(); } xchg(&v->runstate_in_use, 0); registering an area while ( xchg(&v->runstate_in_use, 1) ); /* Clean-up and registering the area */ > >> IOW, if the guest provide a physical address first, then *all* the call should >> be physical address. Alternatively this could be a per vCPU decision. > > I guess we should agree what to implement first. I think there are an agreement that the two methods should not be used together. [..] >>> diff --git a/xen/include/asm-arm/domain.h b/xen/include/asm-arm/domain.h >>> index 312fec8..3fb6ea2 100644 >>> --- a/xen/include/asm-arm/domain.h >>> +++ b/xen/include/asm-arm/domain.h >>> @@ -217,6 +217,8 @@ void vcpu_show_execution_state(struct vcpu *); >>> void vcpu_show_registers(const struct vcpu *); >>> void vcpu_switch_to_aarch64_mode(struct vcpu *); >>> +void update_runstate_area(struct vcpu *); >>> + >>> /* >>> * Due to the restriction of GICv3, the number of vCPUs in AFF0 is >>> * limited to 16, thus only the first 4 bits of AFF0 are legal. We will >>> diff --git a/xen/include/xen/domain.h b/xen/include/xen/domain.h >>> index d1bfc82..ecddcfe 100644 >>> --- a/xen/include/xen/domain.h >>> +++ b/xen/include/xen/domain.h >>> @@ -118,4 +118,6 @@ struct vnuma_info { >>> void vnuma_destroy(struct vnuma_info *vnuma); >>> +struct vcpu_register_runstate_memory_area; >>> + >>> #endif /* __XEN_DOMAIN_H__ */ >>> diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h >>> index 748bb0f..2afe31c 100644 >>> --- a/xen/include/xen/sched.h >>> +++ b/xen/include/xen/sched.h >>> @@ -163,15 +163,23 @@ struct vcpu >>> void *sched_priv; /* scheduler-specific data */ >>> struct vcpu_runstate_info runstate; >>> + >>> + spinlock_t mapped_runstate_lock; >>> + >>> #ifndef CONFIG_COMPAT >>> # define runstate_guest(v) ((v)->runstate_guest) >>> XEN_GUEST_HANDLE(vcpu_runstate_info_t) runstate_guest; /* guest address */ >>> + vcpu_runstate_info_t *mapped_runstate; >>> #else >>> # define runstate_guest(v) ((v)->runstate_guest.native) >>> union { >>> XEN_GUEST_HANDLE(vcpu_runstate_info_t) native; >>> XEN_GUEST_HANDLE(vcpu_runstate_info_compat_t) compat; >>> } runstate_guest; /* guest address */ >>> + union { >>> + vcpu_runstate_info_t* native; >>> + vcpu_runstate_info_compat_t* compat; >>> + } mapped_runstate; /* guest address */ >> > The combination of mapped_runstate and runstate_guest is a bit confusing. I >> think you want to rework the interface to show that only one is possible at >> the time and make clear which one is used by who. Maybe: > > As I said before, IMO coupling those interfaces makes the code complicated and > ugly. Well, I can't see how it can be ugly (see my example above). Cheers, -- Julien Grall _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
WARNING: multiple messages have this Message-ID (diff)
From: Julien Grall <julien.grall@arm.com> To: Andrii Anisov <andrii.anisov@gmail.com>, xen-devel@lists.xen.org Cc: "Stefano Stabellini" <sstabellini@kernel.org>, "Andrii Anisov" <andrii_anisov@epam.com>, "Konrad Rzeszutek Wilk" <konrad.wilk@oracle.com>, "George Dunlap" <George.Dunlap@eu.citrix.com>, "Andrew Cooper" <andrew.cooper3@citrix.com>, "Ian Jackson" <ian.jackson@eu.citrix.com>, "Tim Deegan" <tim@xen.org>, "Jan Beulich" <JBeulich@suse.com>, xen-devel@lists.xenproject.org, "Wei Liu" <wei.liu2@citrix.com>, "Roger Pau Monné" <roger.pau@citrix.com> Subject: Re: [Xen-devel] [PATCH v2 2/2] xen: implement VCPUOP_register_runstate_phys_memory_area Date: Tue, 14 May 2019 10:58:24 +0100 [thread overview] Message-ID: <24f3e2b1-4889-1780-0e48-0c909f555997@arm.com> (raw) Message-ID: <20190514095824.qQsXlQHCK_z4e5QFtp-4CQ7Ya4iwL4WGkAM0xTrxyzU@z> (raw) In-Reply-To: <959c8975-ec44-a788-25b3-84ee17520abc@gmail.com> On 13/05/2019 13:30, Andrii Anisov wrote: > > > On 08.05.19 18:40, Julien Grall wrote: >>> diff --git a/xen/arch/arm/domain.c b/xen/arch/arm/domain.c >>> index 6dc633e..8e24e63 100644 >> >>> { >>> - void __user *guest_handle = NULL; >>> + if ( !guest_handle_is_null(runstate_guest(v)) ) >>> + { >>> + void __user *guest_handle = NULL; >>> + if ( VM_ASSIST(v->domain, runstate_update_flag) ) >>> + { >>> + guest_handle = &v->runstate_guest.p->state_entry_time + 1; >>> + guest_handle--; >>> + v->runstate.state_entry_time |= XEN_RUNSTATE_UPDATE; >>> + __raw_copy_to_guest(guest_handle, >>> + (void *)(&v->runstate.state_entry_time + 1) >>> - 1, >>> + 1); >>> + smp_wmb(); >>> + } >>> - if ( guest_handle_is_null(runstate_guest(v)) ) >>> - return; >>> + __copy_to_guest(runstate_guest(v), &v->runstate, 1); >>> - if ( VM_ASSIST(v->domain, runstate_update_flag) ) >>> - { >>> - guest_handle = &v->runstate_guest.p->state_entry_time + 1; >>> - guest_handle--; >>> - v->runstate.state_entry_time |= XEN_RUNSTATE_UPDATE; >>> - __raw_copy_to_guest(guest_handle, >>> - (void *)(&v->runstate.state_entry_time + 1) - 1, >>> 1); >>> - smp_wmb(); >>> + if ( guest_handle ) >>> + { >>> + v->runstate.state_entry_time &= ~XEN_RUNSTATE_UPDATE; >>> + smp_wmb(); >>> + __raw_copy_to_guest(guest_handle, >>> + (void *)(&v->runstate.state_entry_time + 1) >>> - 1, >>> + 1); >>> + } >>> } >>> - __copy_to_guest(runstate_guest(v), &v->runstate, 1); >>> - >>> - if ( guest_handle ) >>> + spin_lock(&v->mapped_runstate_lock); >>> + if ( v->mapped_runstate ) >> >> The code looks a bit odd to me, you seem to allow a guest to provide 2 >> runstate areas: one using guest virtual address the other using guest physical >> address. >> >> It would be best if we prevent a guest to mix match them. > > Firstly I turned to implementing in that way, but the locking and decissions > code become really ugly and complex while trying to cover 'guest's misbehavior' > scenarios. I think it is possible to have a simple version taking the decision on which method to use. You can either use the spin_lock to protect everything or use something like: update_runstate_area(): if ( xchg(&v->runstate_in_use, 1) ) return; switch ( v->runstate_type ) { GVADDR: update_runstate_by_gvaddr(); GPADDR: update_runstate_by_gpaddr(); } xchg(&v->runstate_in_use, 0); registering an area while ( xchg(&v->runstate_in_use, 1) ); /* Clean-up and registering the area */ > >> IOW, if the guest provide a physical address first, then *all* the call should >> be physical address. Alternatively this could be a per vCPU decision. > > I guess we should agree what to implement first. I think there are an agreement that the two methods should not be used together. [..] >>> diff --git a/xen/include/asm-arm/domain.h b/xen/include/asm-arm/domain.h >>> index 312fec8..3fb6ea2 100644 >>> --- a/xen/include/asm-arm/domain.h >>> +++ b/xen/include/asm-arm/domain.h >>> @@ -217,6 +217,8 @@ void vcpu_show_execution_state(struct vcpu *); >>> void vcpu_show_registers(const struct vcpu *); >>> void vcpu_switch_to_aarch64_mode(struct vcpu *); >>> +void update_runstate_area(struct vcpu *); >>> + >>> /* >>> * Due to the restriction of GICv3, the number of vCPUs in AFF0 is >>> * limited to 16, thus only the first 4 bits of AFF0 are legal. We will >>> diff --git a/xen/include/xen/domain.h b/xen/include/xen/domain.h >>> index d1bfc82..ecddcfe 100644 >>> --- a/xen/include/xen/domain.h >>> +++ b/xen/include/xen/domain.h >>> @@ -118,4 +118,6 @@ struct vnuma_info { >>> void vnuma_destroy(struct vnuma_info *vnuma); >>> +struct vcpu_register_runstate_memory_area; >>> + >>> #endif /* __XEN_DOMAIN_H__ */ >>> diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h >>> index 748bb0f..2afe31c 100644 >>> --- a/xen/include/xen/sched.h >>> +++ b/xen/include/xen/sched.h >>> @@ -163,15 +163,23 @@ struct vcpu >>> void *sched_priv; /* scheduler-specific data */ >>> struct vcpu_runstate_info runstate; >>> + >>> + spinlock_t mapped_runstate_lock; >>> + >>> #ifndef CONFIG_COMPAT >>> # define runstate_guest(v) ((v)->runstate_guest) >>> XEN_GUEST_HANDLE(vcpu_runstate_info_t) runstate_guest; /* guest address */ >>> + vcpu_runstate_info_t *mapped_runstate; >>> #else >>> # define runstate_guest(v) ((v)->runstate_guest.native) >>> union { >>> XEN_GUEST_HANDLE(vcpu_runstate_info_t) native; >>> XEN_GUEST_HANDLE(vcpu_runstate_info_compat_t) compat; >>> } runstate_guest; /* guest address */ >>> + union { >>> + vcpu_runstate_info_t* native; >>> + vcpu_runstate_info_compat_t* compat; >>> + } mapped_runstate; /* guest address */ >> > The combination of mapped_runstate and runstate_guest is a bit confusing. I >> think you want to rework the interface to show that only one is possible at >> the time and make clear which one is used by who. Maybe: > > As I said before, IMO coupling those interfaces makes the code complicated and > ugly. Well, I can't see how it can be ugly (see my example above). Cheers, -- Julien Grall _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
next prev parent reply other threads:[~2019-05-14 9:58 UTC|newest] Thread overview: 83+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-04-23 8:10 [PATCH v2 0/2] Introduce runstate area registration with phys address Andrii Anisov 2019-04-23 8:10 ` [PATCH v2 1/2] xen: introduce VCPUOP_register_runstate_phys_memory_area hypercall Andrii Anisov 2019-04-23 8:10 ` [Xen-devel] " Andrii Anisov 2019-05-08 10:10 ` George Dunlap 2019-05-08 10:10 ` [Xen-devel] " George Dunlap 2019-04-23 8:10 ` [PATCH v2 2/2] xen: implement VCPUOP_register_runstate_phys_memory_area Andrii Anisov 2019-04-23 8:10 ` [Xen-devel] " Andrii Anisov 2019-05-08 15:40 ` Julien Grall 2019-05-08 15:40 ` [Xen-devel] " Julien Grall 2019-05-09 9:27 ` Jan Beulich 2019-05-09 9:27 ` [Xen-devel] " Jan Beulich 2019-05-14 9:35 ` Julien Grall 2019-05-14 9:35 ` [Xen-devel] " Julien Grall 2019-05-14 9:48 ` Jan Beulich 2019-05-14 9:48 ` [Xen-devel] " Jan Beulich 2019-05-14 11:23 ` Julien Grall 2019-05-14 11:23 ` [Xen-devel] " Julien Grall 2019-05-14 11:29 ` Jan Beulich 2019-05-14 11:29 ` [Xen-devel] " Jan Beulich 2019-05-13 12:30 ` Andrii Anisov 2019-05-13 12:30 ` [Xen-devel] " Andrii Anisov 2019-05-14 9:58 ` Julien Grall [this message] 2019-05-14 9:58 ` Julien Grall 2019-05-14 10:08 ` Andrii Anisov 2019-05-14 10:08 ` [Xen-devel] " Andrii Anisov 2019-05-14 11:24 ` Julien Grall 2019-05-14 11:24 ` [Xen-devel] " Julien Grall 2019-05-14 11:45 ` Andrii Anisov 2019-05-14 11:45 ` [Xen-devel] " Andrii Anisov 2019-05-14 12:02 ` Jan Beulich 2019-05-14 12:02 ` [Xen-devel] " Jan Beulich 2019-05-14 13:05 ` Andrii Anisov 2019-05-14 13:05 ` [Xen-devel] " Andrii Anisov 2019-05-14 13:49 ` Julien Grall 2019-05-14 13:49 ` [Xen-devel] " Julien Grall 2019-05-15 9:04 ` Andrii Anisov 2019-05-15 9:04 ` [Xen-devel] " Andrii Anisov 2019-05-15 10:31 ` Julien Grall 2019-05-15 10:31 ` [Xen-devel] " Julien Grall 2019-05-14 13:49 ` Jan Beulich 2019-05-14 13:49 ` [Xen-devel] " Jan Beulich 2019-05-15 8:44 ` Andrii Anisov 2019-05-15 8:44 ` [Xen-devel] " Andrii Anisov 2019-05-15 11:59 ` Jan Beulich 2019-05-15 11:59 ` [Xen-devel] " Jan Beulich 2019-05-16 12:09 ` Jan Beulich 2019-05-16 12:09 ` [Xen-devel] " Jan Beulich 2019-05-16 13:30 ` Andrii Anisov 2019-05-16 13:30 ` [Xen-devel] " Andrii Anisov 2019-05-16 13:30 ` Andrii Anisov 2019-05-16 13:30 ` [Xen-devel] " Andrii Anisov 2019-05-16 13:48 ` Julien Grall 2019-05-16 13:48 ` [Xen-devel] " Julien Grall 2019-05-16 14:25 ` Andrii Anisov 2019-05-16 14:25 ` [Xen-devel] " Andrii Anisov 2019-05-16 14:28 ` Julien Grall 2019-05-16 14:28 ` [Xen-devel] " Julien Grall 2019-05-16 14:29 ` Andrii Anisov 2019-05-16 14:29 ` [Xen-devel] " Andrii Anisov [not found] ` <fa126315-31af-854e-817a-8640b431c82b@arm.com> [not found] ` <CAC1WxdiMzAq5hRC-mhRQuFDs7z_Hj5w7VAy52ec87SJQOGmp3w@mail.gmail.com> [not found] ` <a28f95a1-d9da-2caf-f4b4-013100176b02@arm.com> [not found] ` <090ce8cc-f329-fe54-4894-b7f12e3cd5a6@gmail.com> 2019-05-08 13:39 ` [PATCH v2 0/2] Introduce runstate area registration with phys address Julien Grall 2019-05-08 13:39 ` [Xen-devel] " Julien Grall 2019-05-08 13:54 ` Andrii Anisov 2019-05-08 13:54 ` [Xen-devel] " Andrii Anisov 2019-05-08 14:31 ` Julien Grall 2019-05-08 14:31 ` [Xen-devel] " Julien Grall 2019-05-08 16:01 ` Andrii Anisov 2019-05-08 16:01 ` [Xen-devel] " Andrii Anisov 2019-05-13 10:50 ` Julien Grall 2019-05-13 10:50 ` [Xen-devel] " Julien Grall 2019-05-13 14:34 ` Andrii Anisov 2019-05-13 14:34 ` [Xen-devel] " Andrii Anisov 2019-05-08 13:59 ` Julien Grall 2019-05-13 10:15 ` Andrii Anisov 2019-05-13 11:16 ` Julien Grall 2019-05-13 14:14 ` Andrii Anisov 2019-05-13 14:34 ` Julien Grall 2019-05-13 15:29 ` Andrii Anisov 2019-05-13 15:31 ` Julien Grall 2019-05-13 15:38 ` Andrii Anisov 2019-05-13 15:40 ` Julien Grall 2019-05-13 15:42 ` Andrii Anisov 2019-05-13 15:45 ` Julien Grall 2019-05-13 16:05 ` Andrii Anisov
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=24f3e2b1-4889-1780-0e48-0c909f555997@arm.com \ --to=julien.grall@arm.com \ --cc=George.Dunlap@eu.citrix.com \ --cc=JBeulich@suse.com \ --cc=andrew.cooper3@citrix.com \ --cc=andrii.anisov@gmail.com \ --cc=andrii_anisov@epam.com \ --cc=ian.jackson@eu.citrix.com \ --cc=konrad.wilk@oracle.com \ --cc=roger.pau@citrix.com \ --cc=sstabellini@kernel.org \ --cc=tim@xen.org \ --cc=wei.liu2@citrix.com \ --cc=xen-devel@lists.xen.org \ --cc=xen-devel@lists.xenproject.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).