[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 03/60] xen/sched: let sched_switch_sched() return new lock address


  • To: Juergen Gross <jgross@xxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • Date: Wed, 12 Jun 2019 09:05:22 +0100
  • Authentication-results: esa1.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none; spf=None smtp.pra=andrew.cooper3@xxxxxxxxxx; spf=Pass smtp.mailfrom=Andrew.Cooper3@xxxxxxxxxx; spf=None smtp.helo=postmaster@xxxxxxxxxxxxxxx
  • Autocrypt: addr=andrew.cooper3@xxxxxxxxxx; prefer-encrypt=mutual; keydata= mQINBFLhNn8BEADVhE+Hb8i0GV6mihnnr/uiQQdPF8kUoFzCOPXkf7jQ5sLYeJa0cQi6Penp VtiFYznTairnVsN5J+ujSTIb+OlMSJUWV4opS7WVNnxHbFTPYZVQ3erv7NKc2iVizCRZ2Kxn srM1oPXWRic8BIAdYOKOloF2300SL/bIpeD+x7h3w9B/qez7nOin5NzkxgFoaUeIal12pXSR Q354FKFoy6Vh96gc4VRqte3jw8mPuJQpfws+Pb+swvSf/i1q1+1I4jsRQQh2m6OTADHIqg2E ofTYAEh7R5HfPx0EXoEDMdRjOeKn8+vvkAwhviWXTHlG3R1QkbE5M/oywnZ83udJmi+lxjJ5 YhQ5IzomvJ16H0Bq+TLyVLO/VRksp1VR9HxCzItLNCS8PdpYYz5TC204ViycobYU65WMpzWe LFAGn8jSS25XIpqv0Y9k87dLbctKKA14Ifw2kq5OIVu2FuX+3i446JOa2vpCI9GcjCzi3oHV e00bzYiHMIl0FICrNJU0Kjho8pdo0m2uxkn6SYEpogAy9pnatUlO+erL4LqFUO7GXSdBRbw5 gNt25XTLdSFuZtMxkY3tq8MFss5QnjhehCVPEpE6y9ZjI4XB8ad1G4oBHVGK5LMsvg22PfMJ ISWFSHoF/B5+lHkCKWkFxZ0gZn33ju5n6/FOdEx4B8cMJt+cWwARAQABtClBbmRyZXcgQ29v cGVyIDxhbmRyZXcuY29vcGVyM0BjaXRyaXguY29tPokCOgQTAQgAJAIbAwULCQgHAwUVCgkI CwUWAgMBAAIeAQIXgAUCWKD95wIZAQAKCRBlw/kGpdefoHbdD/9AIoR3k6fKl+RFiFpyAhvO 59ttDFI7nIAnlYngev2XUR3acFElJATHSDO0ju+hqWqAb8kVijXLops0gOfqt3VPZq9cuHlh IMDquatGLzAadfFx2eQYIYT+FYuMoPZy/aTUazmJIDVxP7L383grjIkn+7tAv+qeDfE+txL4 SAm1UHNvmdfgL2/lcmL3xRh7sub3nJilM93RWX1Pe5LBSDXO45uzCGEdst6uSlzYR/MEr+5Z JQQ32JV64zwvf/aKaagSQSQMYNX9JFgfZ3TKWC1KJQbX5ssoX/5hNLqxMcZV3TN7kU8I3kjK mPec9+1nECOjjJSO/h4P0sBZyIUGfguwzhEeGf4sMCuSEM4xjCnwiBwftR17sr0spYcOpqET ZGcAmyYcNjy6CYadNCnfR40vhhWuCfNCBzWnUW0lFoo12wb0YnzoOLjvfD6OL3JjIUJNOmJy RCsJ5IA/Iz33RhSVRmROu+TztwuThClw63g7+hoyewv7BemKyuU6FTVhjjW+XUWmS/FzknSi dAG+insr0746cTPpSkGl3KAXeWDGJzve7/SBBfyznWCMGaf8E2P1oOdIZRxHgWj0zNr1+ooF /PzgLPiCI4OMUttTlEKChgbUTQ+5o0P080JojqfXwbPAyumbaYcQNiH1/xYbJdOFSiBv9rpt TQTBLzDKXok86LkCDQRS4TZ/ARAAkgqudHsp+hd82UVkvgnlqZjzz2vyrYfz7bkPtXaGb9H4 Rfo7mQsEQavEBdWWjbga6eMnDqtu+FC+qeTGYebToxEyp2lKDSoAsvt8w82tIlP/EbmRbDVn 7bhjBlfRcFjVYw8uVDPptT0TV47vpoCVkTwcyb6OltJrvg/QzV9f07DJswuda1JH3/qvYu0p vjPnYvCq4NsqY2XSdAJ02HrdYPFtNyPEntu1n1KK+gJrstjtw7KsZ4ygXYrsm/oCBiVW/OgU g/XIlGErkrxe4vQvJyVwg6YH653YTX5hLLUEL1NS4TCo47RP+wi6y+TnuAL36UtK/uFyEuPy wwrDVcC4cIFhYSfsO0BumEI65yu7a8aHbGfq2lW251UcoU48Z27ZUUZd2Dr6O/n8poQHbaTd 6bJJSjzGGHZVbRP9UQ3lkmkmc0+XCHmj5WhwNNYjgbbmML7y0fsJT5RgvefAIFfHBg7fTY/i kBEimoUsTEQz+N4hbKwo1hULfVxDJStE4sbPhjbsPCrlXf6W9CxSyQ0qmZ2bXsLQYRj2xqd1 bpA+1o1j2N4/au1R/uSiUFjewJdT/LX1EklKDcQwpk06Af/N7VZtSfEJeRV04unbsKVXWZAk uAJyDDKN99ziC0Wz5kcPyVD1HNf8bgaqGDzrv3TfYjwqayRFcMf7xJaL9xXedMcAEQEAAYkC HwQYAQgACQUCUuE2fwIbDAAKCRBlw/kGpdefoG4XEACD1Qf/er8EA7g23HMxYWd3FXHThrVQ HgiGdk5Yh632vjOm9L4sd/GCEACVQKjsu98e8o3ysitFlznEns5EAAXEbITrgKWXDDUWGYxd pnjj2u+GkVdsOAGk0kxczX6s+VRBhpbBI2PWnOsRJgU2n10PZ3mZD4Xu9kU2IXYmuW+e5KCA vTArRUdCrAtIa1k01sPipPPw6dfxx2e5asy21YOytzxuWFfJTGnVxZZSCyLUO83sh6OZhJkk b9rxL9wPmpN/t2IPaEKoAc0FTQZS36wAMOXkBh24PQ9gaLJvfPKpNzGD8XWR5HHF0NLIJhgg 4ZlEXQ2fVp3XrtocHqhu4UZR4koCijgB8sB7Tb0GCpwK+C4UePdFLfhKyRdSXuvY3AHJd4CP 4JzW0Bzq/WXY3XMOzUTYApGQpnUpdOmuQSfpV9MQO+/jo7r6yPbxT7CwRS5dcQPzUiuHLK9i nvjREdh84qycnx0/6dDroYhp0DFv4udxuAvt1h4wGwTPRQZerSm4xaYegEFusyhbZrI0U9tJ B8WrhBLXDiYlyJT6zOV2yZFuW47VrLsjYnHwn27hmxTC/7tvG3euCklmkn9Sl9IAKFu29RSo d5bD8kMSCYsTqtTfT6W4A3qHGvIDta3ptLYpIAOD2sY3GYq2nf3Bbzx81wZK14JdDDHUX2Rs 6+ahAA==
  • Cc: Stefano Stabellini <sstabellini@xxxxxxxxxx>, Wei Liu <wl@xxxxxxx>, Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>, George Dunlap <george.dunlap@xxxxxxxxxxxxx>, Tim Deegan <tim@xxxxxxx>, Ian Jackson <ian.jackson@xxxxxxxxxxxxx>, Robert VanVossen <robert.vanvossen@xxxxxxxxxxxxxxx>, Dario Faggioli <dfaggioli@xxxxxxxx>, Julien Grall <julien.grall@xxxxxxx>, Josh Whitehead <josh.whitehead@xxxxxxxxxxxxxxx>, Meng Xu <mengxu@xxxxxxxxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>
  • Delivery-date: Wed, 12 Jun 2019 08:05:39 +0000
  • Ironport-sdr: gqUjvhJxtIg1CuknlAhU+q2UNjqqJVG9HNU5b6qdeHEsAN1jppWuict9MPOnRKwbSEeZPG9PrH QxGje1kNqLPL/0oLRaAusB7m4QAuJb9O+DZEI0Q5kSTZDzL12+qil+i1QOuQvOsH0XYyTgoqSs 9TYzWWSTIbeF53u0uuUbGbnYEHOM3aQC8CokVKkpVyOWwwny0Wak/DNosIEiVWaBgdc1BIyy2A +kr6LZSF5X3u1j2w+3TlwxUsq2+DUbsZrWQhf6UKuE5xHmYdH32g/L6NM6dW/TJfLmA+yPZF5D ugQ=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
  • Openpgp: preference=signencrypt

On 28/05/2019 11:32, Juergen Gross wrote:
> @@ -1870,8 +1871,19 @@ int schedule_cpu_switch(unsigned int cpu, struct 
> cpupool *c)
>      old_lock = pcpu_schedule_lock_irq(cpu);
>  
>      vpriv_old = idle->sched_priv;
> -    ppriv_old = per_cpu(schedule_data, cpu).sched_priv;
> -    sched_switch_sched(new_ops, cpu, ppriv, vpriv);
> +    ppriv_old = sd->sched_priv;
> +    new_lock = sched_switch_sched(new_ops, cpu, ppriv, vpriv);
> +
> +    per_cpu(scheduler, cpu) = new_ops;
> +    sd->sched_priv = ppriv;
> +
> +    /*
> +     * (Re?)route the lock to the per pCPU lock as /last/ thing. In fact,
> +     * if it is free (and it can be) we want that anyone that manages
> +     * taking it, finds all the initializations we've done above in place.
> +     */
> +    smp_mb();

I realise you're just moving existing code, but this barrier sticks out
like a sore thumb.

A full memory barrier is a massive overhead for what should be
smp_wmb().  The matching barrier is actually hidden in the implicit
semantics of managing to lock sd->schedule_lock (which is trial an error
anyway), but the only thing that matters here is that all other written
data is in place first.

Beyond that, local causality will cause all reads to be in order (not
that the are important) due to logic dependencies.  Any that miss out on
this are a optimisation-waiting-to-happen as the compiler could elide
them fully.

~Andrew

> +    sd->schedule_lock = new_lock;
>  
>      /* _Not_ pcpu_schedule_unlock(): schedule_lock may have changed! */
>      spin_unlock_irq(old_lock);


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.