linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Andy Shevchenko <andy.shevchenko@gmail.com>
To: Vitaly Kuznetsov <vkuznets@redhat.com>
Cc: devel@linuxdriverproject.org, "x86@kernel.org" <x86@kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"K. Y. Srinivasan" <kys@microsoft.com>,
	Haiyang Zhang <haiyangz@microsoft.com>,
	Stephen Hemminger <sthemmin@microsoft.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>, "H. Peter Anvin" <hpa@zytor.com>,
	Steven Rostedt <rostedt@goodmis.org>,
	Jork Loeser <Jork.Loeser@microsoft.com>,
	Simon Xiao <sixiao@microsoft.com>,
	Andy Lutomirski <luto@kernel.org>
Subject: Re: [PATCH v4 09/10] x86/hyper-v: support extended CPU ranges for TLB flush hypercalls
Date: Sat, 27 May 2017 21:13:51 +0300	[thread overview]
Message-ID: <CAHp75VfmgOWnTJziYfO-ZQw0+ABQcFWbOcGGMqz9NzvKK_AKtQ@mail.gmail.com> (raw)
In-Reply-To: <20170524120405.19079-10-vkuznets@redhat.com>

On Wed, May 24, 2017 at 3:04 PM, Vitaly Kuznetsov <vkuznets@redhat.com> wrote:
> Hyper-V hosts may support more than 64 vCPUs, we need to use
> HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX/LIST_EX hypercalls in this
> case.

> +{

> +       /*
> +        * We can't be sure that translated vcpu numbers will always be
> +        * in ascending order, so iterate over all possible banks and
> +        * check all vcpus in it instead.

vcpu -> vCPU
vcpus -> vCPUs

> +        */

> +       for (cur_bank = 0; cur_bank < ms_hyperv.max_vp_index/64; cur_bank++) {
> +               has_cpus = false;
> +               for_each_cpu(cpu, cpus) {

int vcpu_bank = vcpu / 64;
int vcpu_offset = vcpu % 64;

> +                       vcpu = hv_cpu_number_to_vp_number(cpu);

> +                       if (vcpu/64 != cur_bank)

if (vcpu_bank != cur_bank)

> +                               continue;
> +                       if (!has_cpus) {

> +                               flush->hv_vp_set.valid_bank_mask |=
> +                                       1 << vcpu / 64;

__set_bit(vcpu_bank, &mask);

> +                               flush->hv_vp_set.bank_contents[nr_bank] =
> +                                       1 << vcpu % 64;

Ditto. (vcpu_offset)

> +                               has_cpus = true;
> +                       } else {

> +                               flush->hv_vp_set.bank_contents[nr_bank] |=
> +                                       1 << vcpu % 64;

Ditto.

> +                       }
> +               }
> +               if (has_cpus)
> +                       nr_bank++;
> +       }
> +
> +       return nr_bank;
> +}

> +static void hyperv_flush_tlb_others_ex(const struct cpumask *cpus,
> +                                      struct mm_struct *mm,
> +                                      unsigned long start,
> +                                      unsigned long end)
> +{
> +       struct hv_flush_pcpu_ex *flush;
> +       unsigned long cur, flags;

> +       u64 status = -1ULL;

U64_MAX

> +       int nr_bank = 0, max_gvas, gva_n;

> +       /*
> +        * We can flush not more than max_gvas with one hypercall. Flush the
> +        * whole address space if we were asked to do more.
> +        */

#define XXX    (PAGE_SIZE * PAGE_SIZE)

> +       max_gvas = (PAGE_SIZE - sizeof(*flush) - nr_bank*8) / 8;

> +
> +       if (end == TLB_FLUSH_ALL ||
> +           (end && ((end - start)/(PAGE_SIZE*PAGE_SIZE)) > max_gvas)) {
> +               if (end == TLB_FLUSH_ALL)
> +                       flush->flags |= HV_FLUSH_NON_GLOBAL_MAPPINGS_ONLY;
> +
> +               status = hv_do_rep_hypercall(
> +                       HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX,
> +                       0, nr_bank + 2, flush, NULL);

if (end == _FLASH_ALL) {
              flush->flags |= HV_FLUSH_NON_GLOBAL_MAPPINGS_ONLY;
              status = hv_do_rep_hypercall(
                      HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX,
                      0, nr_bank + 2, flush, NULL);
} else if (end && ((end - start) / XXX > max_gvas)) {
               status = hv_do_rep_hypercall(
                      HVCALL_FLUSH_VIRTUAL_ADDRESS_SPACE_EX,
                      0, nr_bank + 2, flush, NULL);
} else {
...

Yes, a bit more code, but IMO much more understandable.

> +       } else {
> +               cur = start;
> +               gva_n = nr_bank;
> +               do {

> +                       flush->gva_list[gva_n] = cur & PAGE_MASK;

> +                       /*
> +                        * Lower 12 bits encode the number of additional
> +                        * pages to flush (in addition to the 'cur' page).
> +                        */
> +                       if (end >= cur + PAGE_SIZE * PAGE_SIZE)

if (end >= cur + XXX)

> +                               flush->gva_list[gva_n] |= ~PAGE_MASK;

> +                       else if (end > cur)
> +                               flush->gva_list[gva_n] |=
> +                                       (end - cur - 1) >> PAGE_SHIFT;
> +

> +                       cur += PAGE_SIZE * PAGE_SIZE;

+= XXX;

> +                       ++gva_n;
> +
> +               } while (cur < end);

> +}

-- 
With Best Regards,
Andy Shevchenko

  reply	other threads:[~2017-05-27 18:13 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-24 12:03 [PATCH v4 00/10] Hyper-V: praravirtualized remote TLB flushing and hypercall improvements Vitaly Kuznetsov
2017-05-24 12:03 ` [PATCH v4 01/10] x86/hyper-v: include hyperv/ only when CONFIG_HYPERV is set Vitaly Kuznetsov
2017-05-24 12:03 ` [PATCH v4 02/10] x86/hyper-v: stash the max number of virtual/logical processor Vitaly Kuznetsov
2017-05-27 17:43   ` Andy Shevchenko
2017-06-05 15:24     ` Steven Rostedt
2017-06-05 15:30       ` Stephen Hemminger
2017-05-24 12:03 ` [PATCH v4 03/10] x86/hyper-v: make hv_do_hypercall() inline Vitaly Kuznetsov
2017-05-27 17:46   ` Andy Shevchenko
2017-05-24 12:03 ` [PATCH v4 04/10] x86/hyper-v: fast hypercall implementation Vitaly Kuznetsov
2017-05-27 17:49   ` Andy Shevchenko
2017-05-24 12:04 ` [PATCH v4 05/10] hyper-v: use fast hypercall for HVCALL_SIGNAL_EVENT Vitaly Kuznetsov
2017-05-27 17:50   ` Andy Shevchenko
2017-05-24 12:04 ` [PATCH v4 06/10] x86/hyper-v: implement rep hypercalls Vitaly Kuznetsov
2017-05-24 12:04 ` [PATCH v4 07/10] hyper-v: globalize vp_index Vitaly Kuznetsov
2017-05-24 12:04 ` [PATCH v4 08/10] x86/hyper-v: use hypercall for remote TLB flush Vitaly Kuznetsov
2017-05-24 12:04 ` [PATCH v4 09/10] x86/hyper-v: support extended CPU ranges for TLB flush hypercalls Vitaly Kuznetsov
2017-05-27 18:13   ` Andy Shevchenko [this message]
2017-05-24 12:04 ` [PATCH v4 10/10] tracing/hyper-v: trace hyperv_mmu_flush_tlb_others() Vitaly Kuznetsov
2017-06-05 15:50   ` Steven Rostedt
2017-06-05 16:19     ` Vitaly Kuznetsov
2017-06-07 13:52       ` Steven Rostedt
2017-06-07 14:03         ` Vitaly Kuznetsov

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=CAHp75VfmgOWnTJziYfO-ZQw0+ABQcFWbOcGGMqz9NzvKK_AKtQ@mail.gmail.com \
    --to=andy.shevchenko@gmail.com \
    --cc=Jork.Loeser@microsoft.com \
    --cc=devel@linuxdriverproject.org \
    --cc=haiyangz@microsoft.com \
    --cc=hpa@zytor.com \
    --cc=kys@microsoft.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=luto@kernel.org \
    --cc=mingo@redhat.com \
    --cc=rostedt@goodmis.org \
    --cc=sixiao@microsoft.com \
    --cc=sthemmin@microsoft.com \
    --cc=tglx@linutronix.de \
    --cc=vkuznets@redhat.com \
    --cc=x86@kernel.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: 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 NNTP newsgroup(s).