From: Huacai Chen <chenhuacai@kernel.org>
To: maobibo <maobibo@loongson.cn>
Cc: Tianrui Zhao <zhaotianrui@loongson.cn>,
Juergen Gross <jgross@suse.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Jonathan Corbet <corbet@lwn.net>,
loongarch@lists.linux.dev, linux-kernel@vger.kernel.org,
virtualization@lists.linux.dev, kvm@vger.kernel.org
Subject: Re: [PATCH v8 4/6] LoongArch: KVM: Add vcpu search support from physical cpuid
Date: Mon, 6 May 2024 17:40:35 +0800 [thread overview]
Message-ID: <CAAhV-H7o3oG2KXc2Ou0aWXTLPSNiM3evSB5Z-5dH4bLRd_P_0Q@mail.gmail.com> (raw)
In-Reply-To: <540aa8dd-eada-1f77-0a20-38196fb5472a@loongson.cn>
On Mon, May 6, 2024 at 5:35 PM maobibo <maobibo@loongson.cn> wrote:
>
>
>
> On 2024/5/6 下午4:59, Huacai Chen wrote:
> > On Mon, May 6, 2024 at 4:18 PM maobibo <maobibo@loongson.cn> wrote:
> >>
> >>
> >>
> >> On 2024/5/6 下午3:06, Huacai Chen wrote:
> >>> Hi, Bibo,
> >>>
> >>> On Mon, May 6, 2024 at 2:36 PM maobibo <maobibo@loongson.cn> wrote:
> >>>>
> >>>>
> >>>>
> >>>> On 2024/5/6 上午9:49, Huacai Chen wrote:
> >>>>> Hi, Bibo,
> >>>>>
> >>>>> On Sun, Apr 28, 2024 at 6:05 PM Bibo Mao <maobibo@loongson.cn> wrote:
> >>>>>>
> >>>>>> Physical cpuid is used for interrupt routing for irqchips such as
> >>>>>> ipi/msi/extioi interrupt controller. And physical cpuid is stored
> >>>>>> at CSR register LOONGARCH_CSR_CPUID, it can not be changed once vcpu
> >>>>>> is created and physical cpuid of two vcpus cannot be the same.
> >>>>>>
> >>>>>> Different irqchips have different size declaration about physical cpuid,
> >>>>>> max cpuid value for CSR LOONGARCH_CSR_CPUID on 3A5000 is 512, max cpuid
> >>>>>> supported by IPI hardware is 1024, 256 for extioi irqchip, and 65536
> >>>>>> for MSI irqchip.
> >>>>>>
> >>>>>> The smallest value from all interrupt controllers is selected now,
> >>>>>> and the max cpuid size is defines as 256 by KVM which comes from
> >>>>>> extioi irqchip.
> >>>>>>
> >>>>>> Signed-off-by: Bibo Mao <maobibo@loongson.cn>
> >>>>>> ---
> >>>>>> arch/loongarch/include/asm/kvm_host.h | 26 ++++++++
> >>>>>> arch/loongarch/include/asm/kvm_vcpu.h | 1 +
> >>>>>> arch/loongarch/kvm/vcpu.c | 93 ++++++++++++++++++++++++++-
> >>>>>> arch/loongarch/kvm/vm.c | 11 ++++
> >>>>>> 4 files changed, 130 insertions(+), 1 deletion(-)
> >>>>>>
> >>>>>> diff --git a/arch/loongarch/include/asm/kvm_host.h b/arch/loongarch/include/asm/kvm_host.h
> >>>>>> index 2d62f7b0d377..3ba16ef1fe69 100644
> >>>>>> --- a/arch/loongarch/include/asm/kvm_host.h
> >>>>>> +++ b/arch/loongarch/include/asm/kvm_host.h
> >>>>>> @@ -64,6 +64,30 @@ struct kvm_world_switch {
> >>>>>>
> >>>>>> #define MAX_PGTABLE_LEVELS 4
> >>>>>>
> >>>>>> +/*
> >>>>>> + * Physical cpu id is used for interrupt routing, there are different
> >>>>>> + * definitions about physical cpuid on different hardwares.
> >>>>>> + * For LOONGARCH_CSR_CPUID register, max cpuid size if 512
> >>>>>> + * For IPI HW, max dest CPUID size 1024
> >>>>>> + * For extioi interrupt controller, max dest CPUID size is 256
> >>>>>> + * For MSI interrupt controller, max supported CPUID size is 65536
> >>>>>> + *
> >>>>>> + * Currently max CPUID is defined as 256 for KVM hypervisor, in future
> >>>>>> + * it will be expanded to 4096, including 16 packages at most. And every
> >>>>>> + * package supports at most 256 vcpus
> >>>>>> + */
> >>>>>> +#define KVM_MAX_PHYID 256
> >>>>>> +
> >>>>>> +struct kvm_phyid_info {
> >>>>>> + struct kvm_vcpu *vcpu;
> >>>>>> + bool enabled;
> >>>>>> +};
> >>>>>> +
> >>>>>> +struct kvm_phyid_map {
> >>>>>> + int max_phyid;
> >>>>>> + struct kvm_phyid_info phys_map[KVM_MAX_PHYID];
> >>>>>> +};
> >>>>>> +
> >>>>>> struct kvm_arch {
> >>>>>> /* Guest physical mm */
> >>>>>> kvm_pte_t *pgd;
> >>>>>> @@ -71,6 +95,8 @@ struct kvm_arch {
> >>>>>> unsigned long invalid_ptes[MAX_PGTABLE_LEVELS];
> >>>>>> unsigned int pte_shifts[MAX_PGTABLE_LEVELS];
> >>>>>> unsigned int root_level;
> >>>>>> + spinlock_t phyid_map_lock;
> >>>>>> + struct kvm_phyid_map *phyid_map;
> >>>>>>
> >>>>>> s64 time_offset;
> >>>>>> struct kvm_context __percpu *vmcs;
> >>>>>> diff --git a/arch/loongarch/include/asm/kvm_vcpu.h b/arch/loongarch/include/asm/kvm_vcpu.h
> >>>>>> index 0cb4fdb8a9b5..9f53950959da 100644
> >>>>>> --- a/arch/loongarch/include/asm/kvm_vcpu.h
> >>>>>> +++ b/arch/loongarch/include/asm/kvm_vcpu.h
> >>>>>> @@ -81,6 +81,7 @@ void kvm_save_timer(struct kvm_vcpu *vcpu);
> >>>>>> void kvm_restore_timer(struct kvm_vcpu *vcpu);
> >>>>>>
> >>>>>> int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu, struct kvm_interrupt *irq);
> >>>>>> +struct kvm_vcpu *kvm_get_vcpu_by_cpuid(struct kvm *kvm, int cpuid);
> >>>>>>
> >>>>>> /*
> >>>>>> * Loongarch KVM guest interrupt handling
> >>>>>> diff --git a/arch/loongarch/kvm/vcpu.c b/arch/loongarch/kvm/vcpu.c
> >>>>>> index 3a8779065f73..b633fd28b8db 100644
> >>>>>> --- a/arch/loongarch/kvm/vcpu.c
> >>>>>> +++ b/arch/loongarch/kvm/vcpu.c
> >>>>>> @@ -274,6 +274,95 @@ static int _kvm_getcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 *val)
> >>>>>> return 0;
> >>>>>> }
> >>>>>>
> >>>>>> +static inline int kvm_set_cpuid(struct kvm_vcpu *vcpu, u64 val)
> >>>>>> +{
> >>>>>> + int cpuid;
> >>>>>> + struct loongarch_csrs *csr = vcpu->arch.csr;
> >>>>>> + struct kvm_phyid_map *map;
> >>>>>> +
> >>>>>> + if (val >= KVM_MAX_PHYID)
> >>>>>> + return -EINVAL;
> >>>>>> +
> >>>>>> + cpuid = kvm_read_sw_gcsr(csr, LOONGARCH_CSR_ESTAT);
> >>>>>> + map = vcpu->kvm->arch.phyid_map;
> >>>>>> + spin_lock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + if (map->phys_map[cpuid].enabled) {
> >>>>>> + /*
> >>>>>> + * Cpuid is already set before
> >>>>>> + * Forbid changing different cpuid at runtime
> >>>>>> + */
> >>>>>> + if (cpuid != val) {
> >>>>>> + /*
> >>>>>> + * Cpuid 0 is initial value for vcpu, maybe invalid
> >>>>>> + * unset value for vcpu
> >>>>>> + */
> >>>>>> + if (cpuid) {
> >>>>>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + return -EINVAL;
> >>>>>> + }
> >>>>>> + } else {
> >>>>>> + /* Discard duplicated cpuid set */
> >>>>>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + return 0;
> >>>>>> + }
> >>>>>> + }
> >>>>> I have changed the logic and comments when I apply, you can double
> >>>>> check whether it is correct.
> >>>> I checkout the latest version, the modification in function
> >>>> kvm_set_cpuid() is good for me.
> >>> Now the modified version is like this:
> >>>
> >>> + if (map->phys_map[cpuid].enabled) {
> >>> + /* Discard duplicated CPUID set operation */
> >>> + if (cpuid == val) {
> >>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>> + return 0;
> >>> + }
> >>> +
> >>> + /*
> >>> + * CPUID is already set before
> >>> + * Forbid changing different CPUID at runtime
> >>> + * But CPUID 0 is the initial value for vcpu, so allow
> >>> + * changing from 0 to others
> >>> + */
> >>> + if (cpuid) {
> >>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>> + return -EINVAL;
> >>> + }
> >>> + }
> >>> But I still doubt whether we should allow changing from 0 to others
> >>> while map->phys_map[cpuid].enabled is 1.
> >> It is necessary since the default sw cpuid is zero :-( And we can
> >> optimize it in later, such as set INVALID cpuid in function
> >> kvm_arch_vcpu_create() and logic will be simple in function kvm_set_cpuid().
> > In my opinion, if a vcpu with a uninitialized default physid=0, then
> > map->phys_map[cpuid].enabled should be 0, then code won't come here.
> > And if a vcpu with a real physid=0, then map->phys_map[cpuid].enabled
> > is 1, but we shouldn't allow it to change physid in this case.
> yes, that is actually a problem.
>
> vcpu0 firstly set physid=0, and vcpu0 set physid=1 again is not allowed.
> vcpu0 firstly set physid=0, and vcpu1 set physid=1 is allowed.
So can we simply drop the if (cpuid) checking? That means:
+ if (map->phys_map[cpuid].enabled) {
+ /* Discard duplicated CPUID set operation */
+ if (cpuid == val) {
+ spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
+ return 0;
+ }
+
+ spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
+ return -EINVAL;
+ }
Huacai
>
>
> >
> > Huacai
> >
> >>
> >> Regards
> >> Bibo Mao
> >>
> >>>
> >>> Huacai
> >>>
> >>>>>
> >>>>>> +
> >>>>>> + if (map->phys_map[val].enabled) {
> >>>>>> + /*
> >>>>>> + * New cpuid is already set with other vcpu
> >>>>>> + * Forbid sharing the same cpuid between different vcpus
> >>>>>> + */
> >>>>>> + if (map->phys_map[val].vcpu != vcpu) {
> >>>>>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + return -EINVAL;
> >>>>>> + }
> >>>>>> +
> >>>>>> + /* Discard duplicated cpuid set operation*/
> >>>>>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + return 0;
> >>>>>> + }
> >>>>>> +
> >>>>>> + kvm_write_sw_gcsr(csr, LOONGARCH_CSR_CPUID, val);
> >>>>>> + map->phys_map[val].enabled = true;
> >>>>>> + map->phys_map[val].vcpu = vcpu;
> >>>>>> + if (map->max_phyid < val)
> >>>>>> + map->max_phyid = val;
> >>>>>> + spin_unlock(&vcpu->kvm->arch.phyid_map_lock);
> >>>>>> + return 0;
> >>>>>> +}
> >>>>>> +
> >>>>>> +struct kvm_vcpu *kvm_get_vcpu_by_cpuid(struct kvm *kvm, int cpuid)
> >>>>>> +{
> >>>>>> + struct kvm_phyid_map *map;
> >>>>>> +
> >>>>>> + if (cpuid >= KVM_MAX_PHYID)
> >>>>>> + return NULL;
> >>>>>> +
> >>>>>> + map = kvm->arch.phyid_map;
> >>>>>> + if (map->phys_map[cpuid].enabled)
> >>>>>> + return map->phys_map[cpuid].vcpu;
> >>>>>> +
> >>>>>> + return NULL;
> >>>>>> +}
> >>>>>> +
> >>>>>> +static inline void kvm_drop_cpuid(struct kvm_vcpu *vcpu)
> >>>>>> +{
> >>>>>> + int cpuid;
> >>>>>> + struct loongarch_csrs *csr = vcpu->arch.csr;
> >>>>>> + struct kvm_phyid_map *map;
> >>>>>> +
> >>>>>> + map = vcpu->kvm->arch.phyid_map;
> >>>>>> + cpuid = kvm_read_sw_gcsr(csr, LOONGARCH_CSR_ESTAT);
> >>>>>> + if (cpuid >= KVM_MAX_PHYID)
> >>>>>> + return;
> >>>>>> +
> >>>>>> + if (map->phys_map[cpuid].enabled) {
> >>>>>> + map->phys_map[cpuid].vcpu = NULL;
> >>>>>> + map->phys_map[cpuid].enabled = false;
> >>>>>> + kvm_write_sw_gcsr(csr, LOONGARCH_CSR_CPUID, 0);
> >>>>>> + }
> >>>>>> +}
> >>>>> While kvm_set_cpuid() is protected by a spinlock, do kvm_drop_cpuid()
> >>>>> and kvm_get_vcpu_by_cpuid() also need it?
> >>>>>
> >>>> It is good to me that spinlock is added in function kvm_drop_cpuid().
> >>>> And thinks for the efforts.
> >>>>
> >>>> Regards
> >>>> Bibo Mao
> >>>>>> +
> >>>>>> static int _kvm_setcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 val)
> >>>>>> {
> >>>>>> int ret = 0, gintc;
> >>>>>> @@ -291,7 +380,8 @@ static int _kvm_setcsr(struct kvm_vcpu *vcpu, unsigned int id, u64 val)
> >>>>>> kvm_set_sw_gcsr(csr, LOONGARCH_CSR_ESTAT, gintc);
> >>>>>>
> >>>>>> return ret;
> >>>>>> - }
> >>>>>> + } else if (id == LOONGARCH_CSR_CPUID)
> >>>>>> + return kvm_set_cpuid(vcpu, val);
> >>>>>>
> >>>>>> kvm_write_sw_gcsr(csr, id, val);
> >>>>>>
> >>>>>> @@ -943,6 +1033,7 @@ void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
> >>>>>> hrtimer_cancel(&vcpu->arch.swtimer);
> >>>>>> kvm_mmu_free_memory_cache(&vcpu->arch.mmu_page_cache);
> >>>>>> kfree(vcpu->arch.csr);
> >>>>>> + kvm_drop_cpuid(vcpu);
> >>>>> I think this line should be before the above kfree(), otherwise you
> >>>>> get a "use after free".
> >>>>>
> >>>>> Huacai
> >>>>>
> >>>>>>
> >>>>>> /*
> >>>>>> * If the vCPU is freed and reused as another vCPU, we don't want the
> >>>>>> diff --git a/arch/loongarch/kvm/vm.c b/arch/loongarch/kvm/vm.c
> >>>>>> index 0a37f6fa8f2d..6006a28653ad 100644
> >>>>>> --- a/arch/loongarch/kvm/vm.c
> >>>>>> +++ b/arch/loongarch/kvm/vm.c
> >>>>>> @@ -30,6 +30,14 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
> >>>>>> if (!kvm->arch.pgd)
> >>>>>> return -ENOMEM;
> >>>>>>
> >>>>>> + kvm->arch.phyid_map = kvzalloc(sizeof(struct kvm_phyid_map),
> >>>>>> + GFP_KERNEL_ACCOUNT);
> >>>>>> + if (!kvm->arch.phyid_map) {
> >>>>>> + free_page((unsigned long)kvm->arch.pgd);
> >>>>>> + kvm->arch.pgd = NULL;
> >>>>>> + return -ENOMEM;
> >>>>>> + }
> >>>>>> +
> >>>>>> kvm_init_vmcs(kvm);
> >>>>>> kvm->arch.gpa_size = BIT(cpu_vabits - 1);
> >>>>>> kvm->arch.root_level = CONFIG_PGTABLE_LEVELS - 1;
> >>>>>> @@ -44,6 +52,7 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
> >>>>>> for (i = 0; i <= kvm->arch.root_level; i++)
> >>>>>> kvm->arch.pte_shifts[i] = PAGE_SHIFT + i * (PAGE_SHIFT - 3);
> >>>>>>
> >>>>>> + spin_lock_init(&kvm->arch.phyid_map_lock);
> >>>>>> return 0;
> >>>>>> }
> >>>>>>
> >>>>>> @@ -51,7 +60,9 @@ void kvm_arch_destroy_vm(struct kvm *kvm)
> >>>>>> {
> >>>>>> kvm_destroy_vcpus(kvm);
> >>>>>> free_page((unsigned long)kvm->arch.pgd);
> >>>>>> + kvfree(kvm->arch.phyid_map);
> >>>>>> kvm->arch.pgd = NULL;
> >>>>>> + kvm->arch.phyid_map = NULL;
> >>>>>> }
> >>>>>>
> >>>>>> int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
> >>>>>> --
> >>>>>> 2.39.3
> >>>>>>
> >>>>
> >>
>
next prev parent reply other threads:[~2024-05-06 9:40 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-28 10:05 [PATCH v8 0/6] LoongArch: Add pv ipi support on LoongArch VM Bibo Mao
2024-04-28 10:05 ` [PATCH v8 1/6] LoongArch/smp: Refine some ipi functions on LoongArch platform Bibo Mao
2024-04-28 10:05 ` [PATCH v8 2/6] LoongArch: KVM: Add hypercall instruction emulation support Bibo Mao
2024-05-06 1:54 ` Huacai Chen
2024-05-06 2:41 ` maobibo
2024-04-28 10:05 ` [PATCH v8 3/6] LoongArch: KVM: Add cpucfg area for kvm hypervisor Bibo Mao
2024-04-28 16:38 ` kernel test robot
2024-04-28 10:05 ` [PATCH v8 4/6] LoongArch: KVM: Add vcpu search support from physical cpuid Bibo Mao
2024-05-06 1:49 ` Huacai Chen
2024-05-06 2:29 ` maobibo
2024-05-06 6:36 ` maobibo
2024-05-06 7:06 ` Huacai Chen
2024-05-06 8:18 ` maobibo
2024-05-06 8:59 ` Huacai Chen
2024-05-06 9:35 ` maobibo
2024-05-06 9:40 ` Huacai Chen [this message]
2024-05-06 10:05 ` maobibo
2024-05-06 14:17 ` Huacai Chen
2024-05-07 1:40 ` maobibo
2024-05-07 2:05 ` Huacai Chen
2024-05-07 3:06 ` maobibo
2024-05-08 5:00 ` Huacai Chen
2024-05-08 6:10 ` maobibo
2024-04-28 10:05 ` [PATCH v8 5/6] LoongArch: KVM: Add pv ipi support on kvm side Bibo Mao
2024-04-28 10:05 ` [PATCH v8 6/6] LoongArch: Add pv ipi support on guest kernel side Bibo Mao
2024-05-06 1:53 ` Huacai Chen
2024-05-06 7:00 ` maobibo
2024-05-06 7:04 ` Huacai Chen
2024-05-06 1:45 ` [PATCH v8 0/6] LoongArch: Add pv ipi support on LoongArch VM Huacai Chen
2024-05-06 6:32 ` maobibo
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=CAAhV-H7o3oG2KXc2Ou0aWXTLPSNiM3evSB5Z-5dH4bLRd_P_0Q@mail.gmail.com \
--to=chenhuacai@kernel.org \
--cc=corbet@lwn.net \
--cc=jgross@suse.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=loongarch@lists.linux.dev \
--cc=maobibo@loongson.cn \
--cc=pbonzini@redhat.com \
--cc=virtualization@lists.linux.dev \
--cc=zhaotianrui@loongson.cn \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is 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 read-only IMAP folder(s) and NNTP newsgroup(s).