[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [PATCH RFC 31/49] xen/sched: use sched_resource cpu instead smp_processor_id in schedulers
- To: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx
- From: Juergen Gross <jgross@xxxxxxxx>
- Date: Sat, 30 Mar 2019 11:22:25 +0100
- Autocrypt: addr=jgross@xxxxxxxx; prefer-encrypt=mutual; keydata= mQENBFOMcBYBCACgGjqjoGvbEouQZw/ToiBg9W98AlM2QHV+iNHsEs7kxWhKMjrioyspZKOB ycWxw3ie3j9uvg9EOB3aN4xiTv4qbnGiTr3oJhkB1gsb6ToJQZ8uxGq2kaV2KL9650I1SJve dYm8Of8Zd621lSmoKOwlNClALZNew72NjJLEzTalU1OdT7/i1TXkH09XSSI8mEQ/ouNcMvIJ NwQpd369y9bfIhWUiVXEK7MlRgUG6MvIj6Y3Am/BBLUVbDa4+gmzDC9ezlZkTZG2t14zWPvx XP3FAp2pkW0xqG7/377qptDmrk42GlSKN4z76ELnLxussxc7I2hx18NUcbP8+uty4bMxABEB AAG0H0p1ZXJnZW4gR3Jvc3MgPGpncm9zc0BzdXNlLmNvbT6JATkEEwECACMFAlOMcK8CGwMH CwkIBwMCAQYVCAIJCgsEFgIDAQIeAQIXgAAKCRCw3p3WKL8TL8eZB/9G0juS/kDY9LhEXseh mE9U+iA1VsLhgDqVbsOtZ/S14LRFHczNd/Lqkn7souCSoyWsBs3/wO+OjPvxf7m+Ef+sMtr0 G5lCWEWa9wa0IXx5HRPW/ScL+e4AVUbL7rurYMfwCzco+7TfjhMEOkC+va5gzi1KrErgNRHH kg3PhlnRY0Udyqx++UYkAsN4TQuEhNN32MvN0Np3WlBJOgKcuXpIElmMM5f1BBzJSKBkW0Jc Wy3h2Wy912vHKpPV/Xv7ZwVJ27v7KcuZcErtptDevAljxJtE7aJG6WiBzm+v9EswyWxwMCIO RoVBYuiocc51872tRGywc03xaQydB+9R7BHPuQENBFOMcBYBCADLMfoA44MwGOB9YT1V4KCy vAfd7E0BTfaAurbG+Olacciz3yd09QOmejFZC6AnoykydyvTFLAWYcSCdISMr88COmmCbJzn sHAogjexXiif6ANUUlHpjxlHCCcELmZUzomNDnEOTxZFeWMTFF9Rf2k2F0Tl4E5kmsNGgtSa aMO0rNZoOEiD/7UfPP3dfh8JCQ1VtUUsQtT1sxos8Eb/HmriJhnaTZ7Hp3jtgTVkV0ybpgFg w6WMaRkrBh17mV0z2ajjmabB7SJxcouSkR0hcpNl4oM74d2/VqoW4BxxxOD1FcNCObCELfIS auZx+XT6s+CE7Qi/c44ibBMR7hyjdzWbABEBAAGJAR8EGAECAAkFAlOMcBYCGwwACgkQsN6d 1ii/Ey9D+Af/WFr3q+bg/8v5tCknCtn92d5lyYTBNt7xgWzDZX8G6/pngzKyWfedArllp0Pn fgIXtMNV+3t8Li1Tg843EXkP7+2+CQ98MB8XvvPLYAfW8nNDV85TyVgWlldNcgdv7nn1Sq8g HwB2BHdIAkYce3hEoDQXt/mKlgEGsLpzJcnLKimtPXQQy9TxUaLBe9PInPd+Ohix0XOlY+Uk QFEx50Ki3rSDl2Zt2tnkNYKUCvTJq7jvOlaPd6d/W0tZqpyy7KVay+K4aMobDsodB3dvEAs6 ScCnh03dDAFgIq5nsB11j3KPKdVoPlfucX2c7kGNH+LUMbzqV6beIENfNexkOfxHf4kBrQQY AQgAIBYhBIUSZ3Lo9gSUpdCX97DendYovxMvBQJa3fDQAhsCAIEJELDendYovxMvdiAEGRYI AB0WIQRTLbB6QfY48x44uB6AXGG7T9hjvgUCWt3w0AAKCRCAXGG7T9hjvk2LAP99B/9FenK/ 1lfifxQmsoOrjbZtzCS6OKxPqOLHaY47BgEAqKKn36YAPpbk09d2GTVetoQJwiylx/Z9/mQI CUbQMg1pNQf9EjA1bNcMbnzJCgt0P9Q9wWCLwZa01SnQWFz8Z4HEaKldie+5bHBL5CzVBrLv 81tqX+/j95llpazzCXZW2sdNL3r8gXqrajSox7LR2rYDGdltAhQuISd2BHrbkQVEWD4hs7iV 1KQHe2uwXbKlguKPhk5ubZxqwsg/uIHw0qZDk+d0vxjTtO2JD5Jv/CeDgaBX4Emgp0NYs8IC UIyKXBtnzwiNv4cX9qKlz2Gyq9b+GdcLYZqMlIBjdCz0yJvgeb3WPNsCOanvbjelDhskx9gd 6YUUFFqgsLtrKpCNyy203a58g2WosU9k9H+LcheS37Ph2vMVTISMszW9W8gyORSgmw==
- Cc: Stefano Stabellini <sstabellini@xxxxxxxxxx>, Wei Liu <wei.liu2@xxxxxxxxxx>, 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: Sat, 30 Mar 2019 10:22:40 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
- Openpgp: preference=signencrypt
On 29/03/2019 20:36, Andrew Cooper wrote:
> On 29/03/2019 15:09, Juergen Gross wrote:
>> diff --git a/xen/common/schedule.c b/xen/common/schedule.c
>> index 9b5527c1eb..0b5e5e566b 100644
>> --- a/xen/common/schedule.c
>> +++ b/xen/common/schedule.c
>> @@ -347,7 +347,7 @@ int sched_init_vcpu(struct vcpu *v)
>> if ( (item = sched_alloc_item(v)) == NULL )
>> return 1;
>>
>> - if ( is_idle_domain(d) || d->is_pinned )
>> + if ( is_idle_domain(d) )
>> processor = v->vcpu_id;
>> else
>> processor = sched_select_initial_cpu(v);
>
> This looks like a spurious change. I also don't an obvious other patch
> that it might fit into.
That should be in patch 30.
> As for the field itself, it is also fairly objectionable. It is only
> ever set for the hardware domain, and only if dom0_vcpus_pin is used,
> but the actual pinning information is also reflected in dom0's hard
> affinity mask.
Right.
> In practice, all this flag does is permit the use of VCPUOP_get_physid,
> disallow the use of vcpu_set_hard_affinity(), and allow dom0 to attempt
> to actually write to MSR_AMD64_NB_CFG, MSR_FAM10H_MMIO_CONF_BASE,
> MSR_IA32_UCODE_REV, MSR_IA32_THERM_CONTROL and
> MSR_IA32_ENERGY_PERF_BIAS, rather than having the write silently discarded.
>
> Dom0's use of those MSRs is dubious at best, and disabled by default,
> *and* when active, also cross-checks with the hard affinity mask. Does
> anyone use dom0_vcpus_pin in production?
I have seen it on customer systems.
> I think there is quite a lot of value in getting rid of d->is_pinned and
> is_pinned_vcpu() entirely, with will remove an extreme
> corner-case-x86-ism out of the common code.
Right. I don't see a reason why we need anything else but the hard
affinity for this option.
Allowing to re-pin (or unpin) vcpus should be fine. And use of the
said MSRs could be restricted to the correct hard affinity being
active.
Juergen
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel
|
