linux-kernel - Re: [PATCH v1 2/3] locking/pvqspinlock, hv: Enable PV qspinlock for Hyper-V

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] [day] [month] [year] [list]

Date:   Thu, 20 Sep 2018 09:56:44 +0800
From:   Yi Sun <yi.y.sun@...ux.intel.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     linux-kernel@...r.kernel.org, x86@...nel.org,
        chao.p.peng@...el.com, chao.gao@...el.com,
        isaku.yamahata@...el.com, michael.h.kelly@...rosoft.com,
        tianyu.lan@...rosoft.com, "K. Y. Srinivasan" <kys@...rosoft.com>,
        Haiyang Zhang <haiyangz@...rosoft.com>,
        Stephen Hemminger <sthemmin@...rosoft.com>
Subject: Re: [PATCH v1 2/3] locking/pvqspinlock, hv: Enable PV qspinlock for
 Hyper-V

Hi, 

On 18-09-17 20:54:52, Yi Sun wrote:
> On 18-09-14 12:42:33, Thomas Gleixner wrote:
> > On Fri, 14 Sep 2018, Yi Sun wrote:
> > > > > +static void hv_notify_long_spin_wait(void)
> > > > > +{
> > > > > +	u64 input = spin_wait_info | 0x00000000ffffffff;
> > > > 
> > > > What? The result is always 0x00000000ffffffff .....
> > > > 
> > > Oh, sorry for such error.
> > > 
> > > > > +	spin_wait_info++;
> > > > > +	hv_do_fast_hypercall8(HVCALL_NOTIFY_LONG_SPIN_WAIT, input);
> > > > > +}
> > > > > +
> > > > > +static void hv_qlock_kick(int cpu)
> > > > > +{
> > > > > +	spin_wait_info--;
> > > > 
> > > > Looking at the above this is completely pointless and I have no idea what
> > > > that variable is supposed to do.
> > > > 
> > > Per Microsoft Hypervisor Top Level Functional Specification, the input
> > > parameter of HVCALL_NOTIFY_LONG_SPIN_WAIT is defined as below:
> > > 
> > > SpinwaitInfo – Specifies the accumulated count the guest was spinning.
> > > 
> > > So, it is increased when guest is spinning and reduced when spinlock is
> > > released.
> > 
> > But that's a global variable, so it might be incremented and decremented by
> > several CPUs at once. I don't have the spec and have no time to study it
> > either, but global does not make any sense to me. The spin wait info comes
> > from a single guest CPU and the wakeup is targeted at that guest CPU as
> > well. So why global? It might be defined that way, but then you really want
> > to explain it proper.
> > 
> > Thanks,
> > 
> > 	tglx
> > 
> Let me check this more. Will reply to you later. Thanks!
> 
> BRs,
> Yi Sun

I have got the details from Microsoft.

Notify_Long_Spin_Wait hypercall should be called after a specific times
spinning which can be got through CPUID. When hypervisor receives this,
it will try to ensure that all vcpus are scheduled. 

Notify_Long_Spin_Wait is a standalone feature which should be split out
as a new patch set.

So I would like split it into a new patch set and submit it after PV
Hyper-V Spinlock patch set merged.