Using _MASKED_BIT_ENABLE macro to set mask register bits is straight
forward and not likely to go wrong. However when checking which bit(s)
is(are) enabled, simply bitwise AND value and _MASKED_BIT_ENABLE() won't
output expected result. Suppose the register write is disabling bit 1
by setting 0xFFFF0000, however "& _MASKED_BIT_ENABLE(1)" outputs
0x00010000, and the non-zero check will pass which cause the old code
consider the new value set as an enabling operation.
We found guest set 0x80008000 on boot, and set 0xffff8000 during resume.
Both are legal settings but old code will block latter and force vgpu
enter fail-safe mode.
Introduce two new macro and make proper masked bit check in mmio handler:
IS_MASKED_BITS_ENABLED()
IS_MASKED_BITS_DISABLED()
V2: Rebase.
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20200601030721.17129-1-colin.xu@intel.com
The PPGTT in context image can be overridden by LRI cmd with another
PPGTT's pdps. In such case, the load mm is used instead of the one in
the context image. So we need to load its shadow mm in GVT and replace
ppgtt pointers in command.
This feature is used by guest IGD driver to share gfx VM between
different contexts. Verified by IGT "gem_ctx_clone" test.
v4:
- consolidate shadow mm handlers (Yan)
- fix cmd shadow mm pin error path
v3: (Zhenyu Wang)
- Cleanup PDP register offset check
- Add debug check for guest context ppgtt update
- Skip 3-level ppgtt guest handling code. The reason is that all
guests now use 4-level ppgtt table and the only left case for
3-level table is ancient aliasing ppgtt case. But those guest
kernel has no use of PPGTT LRI command. So 3-level ppgtt guest
for this feature becomes simply un-testable.
v2: (Zhenyu Wang)
- Change to list for handling possible multiple ppgtt table loads
in one submission. Make sure shadow mm is to replace for each one.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Cc: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20200508031409.2562-1-zhenyuw@linux.intel.com
Those regs are added in order to slove the following complains:
[70811.201818] gvt: vgpu(1) Invalid FORCE_NONPRIV write 2341 at offset 24d8
[70811.201825] gvt: vgpu(1) Invalid FORCE_NONPRIV write 2351 at offset 24dc
[70811.201831] gvt: vgpu(1) Invalid FORCE_NONPRIV write 10000d82 at offset 24e0
[70811.201837] gvt: vgpu(1) Invalid FORCE_NONPRIV write 10064844 at offset 24e4
So solve them by adding the required regs to the whitelist.
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20200302083130.17831-1-tina.zhang@intel.com
Drm specific drm_WARN* calls include device information in the
backtrace, so we know what device the warnings originate from.
Covert all the calls of WARN* with device specific drm_WARN*
variants in functions where drm_device struct pointer is readily
available.
The conversion was done automatically with below coccinelle semantic
patch. checkpatch errors/warnings are fixed manually.
@@
identifier func, T;
@@
func(struct intel_vgpu *T,...) {
+struct drm_i915_private *i915 = T->gvt->dev_priv;
<+...
(
-WARN(
+drm_WARN(&i915->drm,
...)
|
-WARN_ON(
+drm_WARN_ON(&i915->drm,
...)
|
-WARN_ONCE(
+drm_WARN_ONCE(&i915->drm,
...)
|
-WARN_ON_ONCE(
+drm_WARN_ON_ONCE(&i915->drm,
...)
)
...+>
}
Signed-off-by: Pankaj Bharadiya <pankaj.laxminarayan.bharadiya@intel.com>
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20200220165507.16823-9-pankaj.laxminarayan.bharadiya@intel.com
Now start using parameterized DBUF_CTL instead
of hardcoded, this would allow shorter access
functions when reading or storing entire state.
Tried to implement it in a MMIO_PIPE manner, however
DBUF_CTL1 address is higher than DBUF_CTL2, which
implies that we have to now subtract from base
rather than add.
v2: - Removed unneeded DBUF_CTL_DIST and DBUF_CTL_ADDR
macros. Started to use _PICK construct as suggested
by Matt Roper.
v3: - _DBUF_CTL_S* to DBUF_CTL_S*, changed X to "slice"
in macro(Ville Syrjälä)
- Introduced enum for enumerating DBUF slices(Ville Syrjälä)
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200202230630.8975-5-stanislav.lisovskiy@intel.com
The BitField of non privilege register address is only from bit 2 to 25.
v2: use REG_GENMASK instead. (Zhenyu)
Signed-off-by: Gao, Fred <fred.gao@intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
The region of pvinfo is reserved for communication between a VMM and
the GPU driver executing on a virtual machine. HW doesn't have any
backing mmio store support for the pvinfo region, thus accessing to
this range through MMIO read/write from host side is forbidden which
is regarded as unclaimed register access.
This patch leaves pvinfo range be initialized with zero.
Signed-off-by: Tina Zhang <tina.zhang@intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
We reference DP AUX registers via the DP_AUX_CH_CTL() and
DP_AUX_CH_DATA() macros that calculate all the register offsets for us
automatically; there's no need to explicitly define every offset in
i915_reg.h if they're never going to be used by the driver code.
v2: Apparently GVT was directly using these raw definitions in a couple
places. Switch GVT code over to using our preferred macros.
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com> #v1
Link: https://patchwork.freedesktop.org/patch/msgid/20191026051226.30807-1-matthew.d.roper@intel.com
PSR registers are a mess, some have the full address while others just
have the additional offset from psr_mmio_base.
For BDW+ psr_mmio_base is nothing more than TRANSCODER_EDP_OFFSET +
0x800 and using it makes more difficult for people with an PSR
register address or PSR register name from from BSpec as i915 also
don't match the BSpec names.
For HSW psr_mmio_base is _DDI_BUF_CTL_A + 0x800 and PSR registers are
only available in DDIA.
Other reason to make relative to transcoder is that since BDW every
transcoder have PSR registers, so in theory it should be possible to
have PSR enabled in a non-eDP transcoder.
So for BDW+ we can use _TRANS2() to get the register offset of any
PSR register in any transcoder while for HSW we have _HSW_PSR_ADJ
that will calculate the register offset for the single PSR instance,
noting that we are already guarded about trying to enable PSR in other
port than DDIA on HSW by the 'if (dig_port->base.port != PORT_A)' in
intel_psr_compute_config(), this check should only be valid for HSW
and will be changed in future.
PSR2 registers and PSR_EVENT was added after Haswell so that is why
_PSR_ADJ() is not used in some macros.
The only registers that can not be relative to transcoder are
PSR_IMR and PSR_IIR that are not relative to anything, so keeping it
hardcoded. That changed for TGL but it will be handled in another
patch.
Also removing BDW_EDP_PSR_BASE from GVT because it is not used as it
is the only PSR register that GVT have.
v5:
- Macros changed to be more explicit about HSW (Dhinakaran)
- Squashed with the patch that added the tran parameter to the
macros (Dhinakaran)
v6:
- Checking for interruption errors after module reload in the
transcoder that will be used (Dhinakaran)
- Using lowercase to the registers offsets
v7:
- Removing IS_HASWELL() from registers macros(Jani)
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190820223325.27490-1-jose.souza@intel.com
There is pvinfo writing come from vgpu might be unexpected, like
writing to one unknown address, GVT-g should do as reserved register
to discard any invalid write. Now GVT-g lets it write to the vreg
without prompt error message, should ignore the unexpected pvinfo
write access and leave the vreg as the default value.
For possible guest query GVT-g host feature, this returned proper
value instead of wrong guest setting.
v2: ignore unexpected pvinfo write instead of return predefined value
Fixes: e39c5add32 ("drm/i915/gvt: vGPU MMIO virtualization")
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z.li@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Instead of updating by MMIO write, all of the wa regs are initialized by
wa_ctx. From host side, it should make this behavior as expected, add
'F_CMD_ACCESS' flag to these regs and allow access by commands.
[ 123.557608] gvt: vgpu 2: srm access to non-render register (b11c)
[ 123.563728] gvt: vgpu 2: MI_STORE_REGISTER_MEM handler error
[ 123.569409] gvt: vgpu 2: cmd parser error
[ 123.573424] 0x0
[ 123.573425] 0x24
[ 123.578686] gvt: vgpu 2: scan workload error
[ 123.582958] GVT Internal error for the guest
[ 123.587317] Now vgpu 2 will enter failsafe mode.
[ 123.591938] gvt: vgpu 2: failed to submit desc 0
[ 123.596557] gvt: vgpu 2: fail submit workload on ring 0
[ 123.601786] gvt: vgpu 2: fail to emulate MMIO write 00002230 len 4
Acked-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Weinan Li <weinan.z.li@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Host prints below warning message when guest running some application:
"gvt: vgpu(1) Invalid FORCE_NONPRIV write 2754 at 24f0".
"gvt: vgpu(1) Invalid FORCE_NONPRIV write 28a0 at 24f0".
Registers 0x2754 and 0x28a0 are required by guest so add to whitelist.
Signed-off-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
the vGPU write on TRTTE and 0x4dfc is now write to vreg first. their
values all be restored hardware when context switching.
Fixes: e39c5add32 ("drm/i915/gvt: vGPU MMIO virtualization")
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
This reverts commit f74a6d9a2c.
BXT needs to access 0x141000-0x1417ff register to obtain the dram info.
But after the snapshot range of I915_MCHBAR is refined in f74a6d9a2c,
it only initializes the range of 0x144000-0x147fff for VGPU and then
causes that the guest GPU can't get the initialized value for dram
detection on BXT.
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Typedef is not recommended in the Linux kernel.The klocwork static code
analyzer takes the enumeration as the full range of intel_gvt_gtt_type_t.
But the intel_gvt_gtt_type_t will never be used in full range. For
example, the GTT_TYPE_INVALID will never be used as an index of an array.
Remove the typedef and let the enumeration starts from zero to pass
klocwork analysis.
This patch fixed the critial issues #483, #551, #665 reported by
klockwork.
v3:
- Remove the typedef and let the enumeration starts from zero.
Signed-off-by: Aleksei Gimbitskii <aleksei.gimbitskii@intel.com>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: Zhi Wang <zhi.a.wang@intel.com>
CC: Colin Xu <colin.xu@intel.com>
Reviewed-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Merge back drm-intel-next for engine name definition refinement
and 54939ea0bd ("drm/i915: Switch to use HWS indices rather than addresses")
that would need gvt fixes to depend on.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
According to Intel GFX PRM on 01.org, plane surface address can be updated
synchronously or asynchronously. Synchronous flip will hold plane surface
address update to start of next vsync, which is current implementation.
Asynchronous flip will update the address as soon as possible. Without
async flip, some 3D application could not reach better performance and
the maximum performance is no higher than vsync frequency.
The patch enables the async flip on plane surface address mmio update,
and increment flip count correctly.
With async flip enabled, some 3D applications have significant performance
improvement. i.e. 3DMark Ice Storm has a 300%~400% increment on score.
v2:
Use bit operation definition for flip mode. (zhenyu)
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
in init_skil_mmio_info, replaced register address with the known
name from i915_reg.h definition to improve code readbility.
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Xiaolin Zhang <xiaolin.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
code cleanup. sreg is not used now. remove it for code cleanness.
v3: remove unnecessary array_size in vreg's memory allocation (min he)
v2: do not allocate memory for sreg. (min he)
Reviewed-by: He, Min <min.he@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
In the next patch, we are introducing a broad virtual engine to encompass
multiple physical engines, losing the 1:1 nature of BIT(engine->id). To
reflect the broader set of engines implied by the virtual instance, lets
store the full bitmask.
v2: Use intel_engine_mask_t (s/ring_mask/engine_mask/)
v3: Tvrtko voted for moah churn so teach everyone to not mention ring
and use $class$instance throughout.
v4: Comment upon the disparity in bspec for using VCS1,VCS2 in gen8 and
VCS[0-4] in later gen. We opt to keep the code consistent and use
0-index naming throughout.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190305180332.30900-1-chris@chris-wilson.co.uk
Currently it will take the snapshot of the MCHBAR registers for gvt-g
initialization so that it can be used for guest vgpu. And it will cover
from 0x140000 to 0x17ffff. In fact based on the HW spec most of them are
meanlingless and some time is wasted to read these register.
Only the range of 0x144000 to 0x147fff contains the valid definition.
So the range of capturing I915 MCHBAR register is refined, which helps
to optimize the gvt-g boot time.
V1->V2: Move the register definition into reg.h
Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
The newly updated guest driver could program 0x21f0 via lri and srm,
without F_CMD_ACCESS flag cmd parser will stop parser reset cmd.
0x21f0 applies to BDW, SKL, KBL, BXT and CFL.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Host print below warning message when creating guest:
"gvt: vgpu(2) Invalid FORCE_NONPRIV write 83a8".
Register 0x83a8 should be in force-to-nonpriv whitelist as required by
guest
v2: update commit message to describe purpose of this patch in detail
(zhenyu wang)
Signed-off-by: Zhao Yan <yan.y.zhao@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
GVT-g only simulates DP port for guest and leaves EDP_PSR_IMR
and EDP_PSR_IIR registers as default MMIO read/write.
So guest won't get expected initial values of these registers when
initializing the gpu driver, which results in following warning and logs.
--------
Interrupt register 0x64838 is not zero: 0xffffffff
WARNING: CPU: 1 PID: 157 at drivers/gpu/drm/i915/i915_irq.c:177
gen3_assert_iir_is_zero+0x38/0xa0
Call Trace:
gen8_de_irq_postinstall+0xa7/0x400
gen8_irq_postinstall+0x27/0x80
drm_irq_install+0xbc/0x140
i915_driver_load+0xa9d/0xd50
--------
Because GVT-g does not handle EDP(embedded DP) simulation for guests,
always set EDP_PSR_IMR and EDP_PSR_IIR to value 0.
Signed-off-by: Longhe Zheng <longhe.zheng@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Host prints lots of untracked MMIO at 0x4653c when creating linux guest.
"gvt: vgpu 2: untracked MMIO 0004653c len 4"
GEN9_CLKGATE_DIS_4 (0x4653c) is accessed by i915 for gmbus clockgating.
However vgpu doesn't support any clockgating powergating operations
on related mmio access trap so need add it to default handler.
GEN9_CLKGATE_DIS_4 is accessed in bxt_gmbus_clock_gating() which only
applies to GEN9_LP so doens't show the warning on other platforms.
The solution is to add it to default handler init_bxt_mmio_info().
Reviewed-by: He, Min <min.he@intel.com>
Signed-off-by: Colin Xu <colin.xu@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Merge tag 'gvt-next-2018-09-04'
drm-intel-next-2018-09-06-1:
UAPI Changes:
- GGTT coherency GETPARAM: GGTT has turned out to be non-coherent for some
platforms, which we've failed to communicate to userspace so far. SNA was
modified to do extra flushing on non-coherent GGTT access, while Mesa will
mitigate by always requiring WC mapping (which is non-coherent anyway).
- Neuter Resource Streamer uAPI: There never really were users for the feature,
so neuter it while keeping the interface bits for compatibility. This is a
long due item from past.
Cross-subsystem Changes:
- Backmerge of branch drm-next-4.19 for DP_DPCD_REV_14 changes
Core Changes:
- None
Driver Changes:
- A load of Icelake (ICL) enabling patches (Paulo, Manasi)
- Enabled full PPGTT for IVB,VLV and HSW (Chris)
- Bugzilla #107113: Distribute DDB based on display resolutions (Mahesh)
- Bugzillas #100023,#107476,#94921: Support limited range DP displays (Jani)
- Bugzilla #107503: Increase LSPCON timeout (Fredrik)
- Avoid boosting GPU due to an occasional stall in interactive workloads (Chris)
- Apply GGTT coherency W/A only for affected systems instead of all (Chris)
- Fix for infinite link training loop for faulty USB-C MST hubs (Nathan)
- Keep KMS functional on Gen4 and earlier when GPU is wedged (Chris)
- Stop holding ppGTT reference from closed VMAs (Chris)
- Clear error registers after error capture (Lionel)
- Various Icelake fixes (Anusha, Jyoti, Ville, Tvrtko)
- Add missing Coffeelake (CFL) PCI IDs (Rodrigo)
- Flush execlists tasklet directly from reset-finish (Chris)
- Fix LPE audio runtime PM (Chris)
- Fix detection of out of range surface positions (GLK/CNL) (Ville)
- Remove wait-for-idle for PSR2 (Dhinakaran)
- Power down existing display hardware resources when display is disabled (Chris)
- Don't allow runtime power management if RC6 doesn't exist (Chris)
- Add debugging checks for runtime power management paths (Imre)
- Increase symmetry in display power init/fini paths (Imre)
- Isolate GVT specific macros from i915_reg.h (Lucas)
- Increase symmetry in power management enable/disable paths (Chris)
- Increase IP disable timeout to 100 ms to avoid DRM_ERROR (Imre)
- Fix memory leak from HDMI HDCP write function (Brian, Rodrigo)
- Reject Y/Yf tiling on interlaced modes (Ville)
- Use a cached mapping for the physical HWS on older gens (Chris)
- Force slow path of writing relocations to buffer if unable to write to userspace (Chris)
- Do a full device reset after being wedged (Chris)
- Keep forcewake counts over reset (in case of debugfs user) (Imre, Chris)
- Avoid false-positive errors from power wells during init (Imre)
- Reset engines forcibly in exchange of declaring whole device wedged (Mika)
- Reduce context HW ID lifetime in preparation for Icelake (Chris)
- Attempt to recover from module load failures (Chris)
- Keep select interrupts over a reset to avoid missing/losing them (Chris)
- GuC submission backend improvements (Jakub)
- Terminate context images with BB_END (Chris, Lionel)
- Make GCC evaluate GGTT view struct size assertions again (Ville)
- Add selftest to exercise suspend/hibernate code-paths for GEM (Chris)
- Use a full emulation of a user ppgtt context in selftests (Chris)
- Exercise resetting in the middle of a wait-on-fence in selftests (Chris)
- Fix coherency issues on selftests for Baytrail (Chris)
- Various other GEM fixes / self-test updates (Chris, Matt)
- GuC doorbell self-tests (Daniele)
- PSR mode control through debugfs for IGTs (Maarten)
- Degrade expected WM latency errors to DRM_DEBUG_KMS (Chris)
- Cope with errors better in MST link training (Dhinakaran)
- Fix WARN on KBL external displays (Azhar)
- Power well code cleanups (Imre)
- Fixes to PSR debugging (Dhinakaran)
- Make forcewake errors louder for easier catching in CI (WARNs) (Chris)
- Fortify tiling code against programmer errors (Chris)
- Bunch of fixes for CI exposed corner cases (multiple authors, mostly Chris)
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907105446.GA22860@jlahtine-desk.ger.corp.intel.com
