From mboxrd@z Thu Jan 1 00:00:00 1970 From: Suraj Jitindar Singh Subject: Re: [PATCH] kvm/stats: Update kvm stats to clear on write to their debugfs entry Date: Tue, 18 Oct 2016 17:20:57 +1100 Message-ID: <1476771657.2150.3.camel@gmail.com> References: <1476424962-7610-1-git-send-email-sjitindarsingh@gmail.com> <1551761682.4111493.1476723030110.JavaMail.zimbra@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8bit Cc: kvm@vger.kernel.org, rkrcmar@redhat.com To: Paolo Bonzini Return-path: Received: from mail-pf0-f193.google.com ([209.85.192.193]:33338 "EHLO mail-pf0-f193.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750761AbcJRGVD (ORCPT ); Tue, 18 Oct 2016 02:21:03 -0400 Received: by mail-pf0-f193.google.com with SMTP id i85so10844587pfa.0 for ; Mon, 17 Oct 2016 23:21:03 -0700 (PDT) In-Reply-To: <1551761682.4111493.1476723030110.JavaMail.zimbra@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Mon, 2016-10-17 at 12:50 -0400, Paolo Bonzini wrote: > > ----- Original Message ----- > > > > From: "Suraj Jitindar Singh" > > To: kvm@vger.kernel.org > > Cc: pbonzini@redhat.com, rkrcmar@redhat.com, "Suraj Jitindar Singh" > > > > Sent: Friday, October 14, 2016 8:02:42 AM > > Subject: [PATCH] kvm/stats: Update kvm stats to clear on write to > > their debugfs entry > > > > Various kvm vm and vcpu stats are provided via debugfs entries. > > Currently there is no way to reset these stats back to zero for the > > system > > except by stopping all vms. > > > > Add the ability to clear (reset back to zero) these stats on a per > > stat > > basis by writing to the debugfs files. The stats are just reset to > > zero > > irrespective of what is actually written to the file. > Even though this is a debug interface, people will still rely on > underspecified APIs.  Please make it fail with EINVAL unless 0 is > written.  Apart from this, it's a good idea. Ok, I'll update this and send a V2 > > Paolo > > > > > > > Signed-off-by: Suraj Jitindar Singh > > --- > >  virt/kvm/kvm_main.c | 66 > >  ++++++++++++++++++++++++++++++++++++++++++++++++----- > >  1 file changed, 60 insertions(+), 6 deletions(-) > > > > diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c > > index 81dfc73..141fc03 100644 > > --- a/virt/kvm/kvm_main.c > > +++ b/virt/kvm/kvm_main.c > > @@ -595,7 +595,7 @@ static int kvm_create_vm_debugfs(struct kvm > > *kvm, int fd) > >   stat_data->kvm = kvm; > >   stat_data->offset = p->offset; > >   kvm->debugfs_stat_data[p - debugfs_entries] = > > stat_data; > > - if (!debugfs_create_file(p->name, 0444, > > + if (!debugfs_create_file(p->name, 0644, > >    kvm->debugfs_dentry, > >    stat_data, > >    stat_fops_per_vm[p- > > >kind])) > > @@ -3658,11 +3658,20 @@ static int vm_stat_get_per_vm(void *data, > > u64 *val) > >   return 0; > >  } > >   > > +static int vm_stat_clear_per_vm(void *data, u64 val) > > +{ > > + struct kvm_stat_data *stat_data = (struct kvm_stat_data > > *)data; > > + > > + *(ulong *)((void *)stat_data->kvm + stat_data->offset) = > > 0; > > + > > + return 0; > > +} > > + > >  static int vm_stat_get_per_vm_open(struct inode *inode, struct > > file *file) > >  { > >   __simple_attr_check_format("%llu\n", 0ull); > >   return kvm_debugfs_open(inode, file, vm_stat_get_per_vm, > > - NULL, "%llu\n"); > > + vm_stat_clear_per_vm, "%llu\n"); > >  } > >   > >  static const struct file_operations vm_stat_get_per_vm_fops = { > > @@ -3688,11 +3697,23 @@ static int vcpu_stat_get_per_vm(void *data, > > u64 *val) > >   return 0; > >  } > >   > > +static int vcpu_stat_clear_per_vm(void *data, u64 val) > > +{ > > + int i; > > + struct kvm_stat_data *stat_data = (struct kvm_stat_data > > *)data; > > + struct kvm_vcpu *vcpu; > > + > > + kvm_for_each_vcpu(i, vcpu, stat_data->kvm) > > + *(u64 *)((void *)vcpu + stat_data->offset) = 0; > > + > > + return 0; > > +} > > + > >  static int vcpu_stat_get_per_vm_open(struct inode *inode, struct > > file *file) > >  { > >   __simple_attr_check_format("%llu\n", 0ull); > >   return kvm_debugfs_open(inode, file, vcpu_stat_get_per_vm, > > -  NULL, "%llu\n"); > > +  vcpu_stat_clear_per_vm, > > "%llu\n"); > >  } > >   > >  static const struct file_operations vcpu_stat_get_per_vm_fops = { > > @@ -3727,7 +3748,23 @@ static int vm_stat_get(void *_offset, u64 > > *val) > >   return 0; > >  } > >   > > -DEFINE_SIMPLE_ATTRIBUTE(vm_stat_fops, vm_stat_get, NULL, > > "%llu\n"); > > +static int vm_stat_clear(void *_offset, u64 val) > > +{ > > + unsigned offset = (long)_offset; > > + struct kvm *kvm; > > + struct kvm_stat_data stat_tmp = {.offset = offset}; > > + > > + spin_lock(&kvm_lock); > > + list_for_each_entry(kvm, &vm_list, vm_list) { > > + stat_tmp.kvm = kvm; > > + vm_stat_clear_per_vm((void *)&stat_tmp, 0); > > + } > > + spin_unlock(&kvm_lock); > > + > > + return 0; > > +} > > + > > +DEFINE_SIMPLE_ATTRIBUTE(vm_stat_fops, vm_stat_get, vm_stat_clear, > > "%llu\n"); > >   > >  static int vcpu_stat_get(void *_offset, u64 *val) > >  { > > @@ -3747,7 +3784,24 @@ static int vcpu_stat_get(void *_offset, u64 > > *val) > >   return 0; > >  } > >   > > -DEFINE_SIMPLE_ATTRIBUTE(vcpu_stat_fops, vcpu_stat_get, NULL, > > "%llu\n"); > > +static int vcpu_stat_clear(void *_offset, u64 val) > > +{ > > + unsigned offset = (long)_offset; > > + struct kvm *kvm; > > + struct kvm_stat_data stat_tmp = {.offset = offset}; > > + > > + spin_lock(&kvm_lock); > > + list_for_each_entry(kvm, &vm_list, vm_list) { > > + stat_tmp.kvm = kvm; > > + vcpu_stat_clear_per_vm((void *)&stat_tmp, 0); > > + } > > + spin_unlock(&kvm_lock); > > + > > + return 0; > > +} > > + > > +DEFINE_SIMPLE_ATTRIBUTE(vcpu_stat_fops, vcpu_stat_get, > > vcpu_stat_clear, > > + "%llu\n"); > >   > >  static const struct file_operations *stat_fops[] = { > >   [KVM_STAT_VCPU] = &vcpu_stat_fops, > > @@ -3765,7 +3819,7 @@ static int kvm_init_debug(void) > >   > >   kvm_debugfs_num_entries = 0; > >   for (p = debugfs_entries; p->name; ++p, > > kvm_debugfs_num_entries++) { > > - if (!debugfs_create_file(p->name, 0444, > > kvm_debugfs_dir, > > + if (!debugfs_create_file(p->name, 0644, > > kvm_debugfs_dir, > >    (void *)(long)p->offset, > >    stat_fops[p->kind])) > >   goto out_dir; > > -- > > 2.5.5 > > > >