Skip to contentSkip to conversation

drm/i915: Do not enable movntdqa optimization in hypervisor guest

PATCH Superseded [email protected] 3 participants 5 messages Dec 22, 2017 – Dec 25, 2017

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Do not enable movntdqa optimization in hypervisor guest

From: Changbin Du <[email protected]>

Our QA reported a problem caused by movntdqa instructions. Currently, the KVM hypervisor doesn't support VEX-prefix instructions emulation. If users passthrough a GPU to guest with vfio option 'x-no-mmap=on', then all access to the BARs will be trapped and emulated. The KVM hypervisor would raise an inertal error to qemu which cause the guest killed. (Since 'movntdqa' ins is not supported.)

This patch try not to enable movntdqa optimization if the driver is running in hypervisor guest.

---
 drivers/gpu/drm/i915/i915_memcpy.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_memcpy.c +6 -1
index 49a0794..8921f40 100644
@@ -96,6 +96,11 @@ bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len)
 
 void i915_memcpy_init_early(struct drm_i915_private *dev_priv)
 {
-	if (static_cpu_has(X86_FEATURE_XMM4_1))
+	/**
+	 * Some hypervisors (e.g. KVM) don't support VEX-prefix instructions
+	 * emulation. So don't enable movntdqa in hypervisor guest.
+	 */
+	if (static_cpu_has(X86_FEATURE_XMM4_1) &&
+	    !boot_cpu_has(X86_FEATURE_HYPERVISOR))
 		static_branch_enable(&has_movntdqa);
 }
Signature · 6 lines
2.7.4

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

✓ Fi.CI.BAT: success for drm/i915: Do not enable movntdqa optimization in hypervisor guest

== Series Details ==

== Series Details ==

Series: drm/i915: Do not enable movntdqa optimization in hypervisor guest
URL : https://patchwork.freedesktop.org/series/35711/
State : success

== Summary ==

Series 35711v1 drm/i915: Do not enable movntdqa optimization in hypervisor guest
https://patchwork.freedesktop.org/api/1.0/series/35711/revisions/1/mbox/

Test debugfs_test:

        Subgroup read_all_entries:
                incomplete -> PASS       (fi-snb-2520m) fdo#103713
Test gem_exec_suspend:
        Subgroup basic-s4-devices:
                dmesg-warn -> PASS       (fi-elk-e7500) fdo#103989
Test kms_pipe_crc_basic:
        Subgroup suspend-read-crc-pipe-b:
                dmesg-warn -> PASS       (fi-kbl-r) fdo#104172 +1
Test kms_psr_sink_crc:
        Subgroup psr_basic:
                pass       -> DMESG-WARN (fi-skl-6700hq) fdo#101144

fdo#103713 https://bugs.freedesktop.org/show_bug.cgi?id=103713
fdo#103989 https://bugs.freedesktop.org/show_bug.cgi?id=103989
fdo#104172 https://bugs.freedesktop.org/show_bug.cgi?id=104172
fdo#101144 https://bugs.freedesktop.org/show_bug.cgi?id=101144

fi-bdw-5557u     total:288  pass:267  dwarn:0   dfail:0   fail:0   skip:21  time:438s
fi-bdw-gvtdvm    total:288  pass:264  dwarn:0   dfail:0   fail:0   skip:24  time:436s
fi-blb-e6850     total:288  pass:223  dwarn:1   dfail:0   fail:0   skip:64  time:383s
fi-bsw-n3050     total:288  pass:242  dwarn:0   dfail:0   fail:0   skip:46  time:496s
fi-bwr-2160      total:288  pass:183  dwarn:0   dfail:0   fail:0   skip:105 time:277s
fi-bxt-dsi       total:288  pass:258  dwarn:0   dfail:0   fail:0   skip:30  time:491s
fi-bxt-j4205     total:288  pass:259  dwarn:0   dfail:0   fail:0   skip:29  time:502s
fi-byt-j1900     total:288  pass:253  dwarn:0   dfail:0   fail:0   skip:35  time:477s
fi-byt-n2820     total:288  pass:249  dwarn:0   dfail:0   fail:0   skip:39  time:463s
fi-elk-e7500     total:224  pass:164  dwarn:14  dfail:0   fail:0   skip:45 
fi-gdg-551       total:288  pass:179  dwarn:0   dfail:0   fail:1   skip:108 time:261s
fi-glk-1         total:288  pass:260  dwarn:0   dfail:0   fail:0   skip:28  time:529s
fi-hsw-4770      total:288  pass:261  dwarn:0   dfail:0   fail:0   skip:27  time:406s
fi-hsw-4770r     total:288  pass:261  dwarn:0   dfail:0   fail:0   skip:27  time:414s
fi-ivb-3520m     total:288  pass:259  dwarn:0   dfail:0   fail:0   skip:29  time:468s
fi-ivb-3770      total:288  pass:255  dwarn:0   dfail:0   fail:0   skip:33  time:424s
fi-kbl-7500u     total:288  pass:263  dwarn:1   dfail:0   fail:0   skip:24  time:483s
fi-kbl-7560u     total:288  pass:268  dwarn:1   dfail:0   fail:0   skip:19  time:520s
fi-kbl-7567u     total:288  pass:268  dwarn:0   dfail:0   fail:0   skip:20  time:467s
fi-kbl-r         total:288  pass:260  dwarn:1   dfail:0   fail:0   skip:27  time:519s
fi-pnv-d510      total:288  pass:222  dwarn:1   dfail:0   fail:0   skip:65  time:581s
fi-skl-6260u     total:288  pass:268  dwarn:0   dfail:0   fail:0   skip:20  time:439s
fi-skl-6600u     total:288  pass:260  dwarn:1   dfail:0   fail:0   skip:27  time:529s
fi-skl-6700hq    total:288  pass:261  dwarn:1   dfail:0   fail:0   skip:26  time:552s
fi-skl-6700k2    total:288  pass:264  dwarn:0   dfail:0   fail:0   skip:24  time:508s
fi-skl-6770hq    total:288  pass:268  dwarn:0   dfail:0   fail:0   skip:20  time:500s
fi-skl-gvtdvm    total:288  pass:265  dwarn:0   dfail:0   fail:0   skip:23  time:445s
fi-snb-2520m     total:288  pass:248  dwarn:0   dfail:0   fail:0   skip:40  time:546s
fi-snb-2600      total:288  pass:248  dwarn:0   dfail:0   fail:0   skip:40  time:411s
Blacklisted hosts:
fi-cfl-s2        total:288  pass:262  dwarn:0   dfail:0   fail:0   skip:26  time:593s
fi-cnl-y         total:288  pass:262  dwarn:0   dfail:0   fail:0   skip:26  time:624s
fi-glk-dsi       total:288  pass:189  dwarn:1   dfail:4   fail:0   skip:94  time:404s

c0a64101df89fe312cb41d27e184555400d0e3b9 drm-tip: 2017y-12m-21d-18h-02m-56s UTC integration manifest af8be38582f2 drm/i915: Do not enable movntdqa optimization in hypervisor guest

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_7562/issues.html
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#2 lore raw

✗ Fi.CI.IGT: warning for drm/i915: Do not enable movntdqa optimization in hypervisor guest

In reply to [email protected]
== Series Details ==

== Series Details ==

Series: drm/i915: Do not enable movntdqa optimization in hypervisor guest
URL : https://patchwork.freedesktop.org/series/35711/
State : warning

== Summary ==

Test pm_rpm:

        Subgroup system-suspend:
                skip       -> PASS       (shard-hsw) fdo#103375
        Subgroup system-suspend-modeset:
                pass       -> SKIP       (shard-hsw)
Test gem_tiled_swapping:
        Subgroup non-threaded:
                pass       -> INCOMPLETE (shard-snb) fdo#104218 +1
Test drv_suspend:
        Subgroup fence-restore-tiled2untiled:
                skip       -> PASS       (shard-snb)
        Subgroup sysfs-reader:
                skip       -> PASS       (shard-hsw)
Test gem_exec_suspend:
        Subgroup basic-s3:
                pass       -> SKIP       (shard-hsw) fdo#103540

fdo#103375 https://bugs.freedesktop.org/show_bug.cgi?id=103375
fdo#104218 https://bugs.freedesktop.org/show_bug.cgi?id=104218
fdo#103540 https://bugs.freedesktop.org/show_bug.cgi?id=103540

shard-hsw        total:2712 pass:1534 dwarn:1   dfail:0   fail:10  skip:1167 time:9395s
shard-snb        total:2640 pass:1283 dwarn:1   dfail:0   fail:10  skip:1345 time:7824s
Blacklisted hosts:
shard-apl        total:2712 pass:1685 dwarn:1   dfail:0   fail:24  skip:1001 time:13756s
shard-kbl        total:2653 pass:1753 dwarn:10  dfail:1   fail:24  skip:864 time:10677s

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_7562/shards.html
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reviewed-by on v1 1/1
#3 lore raw
In reply to [email protected]

Quoting [email protected] (2017-12-22 06:31:49)

Quoted 22 lines

From: Changbin Du <[email protected]>

Our QA reported a problem caused by movntdqa instructions. Currently, the KVM hypervisor doesn't support VEX-prefix instructions emulation. If users passthrough a GPU to guest with vfio option 'x-no-mmap=on', then all access to the BARs will be trapped and emulated. The KVM hypervisor would raise an inertal error to qemu which cause the guest killed. (Since 'movntdqa' ins is not supported.)

This patch try not to enable movntdqa optimization if the driver is running in hypervisor guest.

---
 drivers/gpu/drm/i915/i915_memcpy.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_memcpy.c +0 -0
index 49a0794..8921f40 100644

drivers/gpu/drm/i915/i915_memcpy.c bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len) · line 99

Quoted 5 lines
drivers/gpu/drm/i915/i915_memcpy.c +1 -1
@@ -96,6 +96,11 @@ bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len)
 
 void i915_memcpy_init_early(struct drm_i915_private *dev_priv)
 {
-       if (static_cpu_has(X86_FEATURE_XMM4_1))
+       /**

This isn't a kerneldoc, so just /* for a regular comment.

drivers/gpu/drm/i915/i915_memcpy.c bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len) · line 105

Quoted 6 lines
drivers/gpu/drm/i915/i915_memcpy.c +5 -0
+        * Some hypervisors (e.g. KVM) don't support VEX-prefix instructions
+        * emulation. So don't enable movntdqa in hypervisor guest.
+        */
+       if (static_cpu_has(X86_FEATURE_XMM4_1) &&
+           !boot_cpu_has(X86_FEATURE_HYPERVISOR))
                static_branch_enable(&has_movntdqa);

Code checks out, and I believe you that some hypervisor setups may trap those instructions, so

We could try and jigger the init order about, but we do need this setup quite early so that we use the knowledge in other setup. Considering that, an explicit check for a hypervisor seems sensible. -Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw
Quoted 24 lines · Chris Wilson

Quoting [email protected] (2017-12-22 06:31:49)

Quoted 23 lines

From: Changbin Du <[email protected]>

Our QA reported a problem caused by movntdqa instructions. Currently, the KVM hypervisor doesn't support VEX-prefix instructions emulation. If users passthrough a GPU to guest with vfio option 'x-no-mmap=on', then all access to the BARs will be trapped and emulated. The KVM hypervisor would raise an inertal error to qemu which cause the guest killed. (Since 'movntdqa' ins is not supported.)

This patch try not to enable movntdqa optimization if the driver is running in hypervisor guest.

---
 drivers/gpu/drm/i915/i915_memcpy.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/i915_memcpy.c +0 -0
index 49a0794..8921f40 100644
@@ -96,6 +96,11 @@ bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len)
 
Quoted 7 lines

drivers/gpu/drm/i915/i915_memcpy.c bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len) · line 99

Quoted 4 lines
drivers/gpu/drm/i915/i915_memcpy.c +1 -1
 void i915_memcpy_init_early(struct drm_i915_private *dev_priv)
 {
-       if (static_cpu_has(X86_FEATURE_XMM4_1))
+       /**

This isn't a kerneldoc, so just /* for a regular comment.

Thanks, will update.

Quoted 10 lines

drivers/gpu/drm/i915/i915_memcpy.c bool i915_memcpy_from_wc(void *dst, const void *src, unsigned long len) · line 105

Quoted 6 lines
drivers/gpu/drm/i915/i915_memcpy.c +5 -0
+        * Some hypervisors (e.g. KVM) don't support VEX-prefix instructions
+        * emulation. So don't enable movntdqa in hypervisor guest.
+        */
+       if (static_cpu_has(X86_FEATURE_XMM4_1) &&
+           !boot_cpu_has(X86_FEATURE_HYPERVISOR))
                static_branch_enable(&has_movntdqa);

Code checks out, and I believe you that some hypervisor setups may trap those instructions, so

Quoted 6 lines

We could try and jigger the init order about, but we do need this setup quite early so that we use the knowledge in other setup. Considering that, an explicit check for a hypervisor seems sensible. -Chris

And this check is not only for GVTg, but also for physical GPU passthrough. So it is beyong intel_vgpu_active().

Signature · 6 lines
Thanks,
Changbin Du
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx