Current vblank emulator uses single hrtimer at 16ms period for all vGPUs,
which introduces three major issues:
- 16ms matches the refresh rate at 62.5Hz (instead of 60Hz) which
doesn't follow standard timing. This leads to some frame drop or glitch
issue during video playback. SW expects a vsync interval of 16.667ms or
higher precision for an accurate 60Hz refresh rate. However current
vblank emulator only works at 16ms.
- Doesn't respect the fact that with current virtual EDID timing set,
not all resolutions are running at 60Hz. For example, current virtual
EDID also supports refresh rate at 56Hz, 59.97Hz, 60Hz, 75Hz, etc.
- Current vblank emulator use single hrtimer for all vGPUs. Regardsless
the possibility that different guests could run in different
resolutions, all vsync interrupts are injected at 16ms interval with
same hrtimer.
Based on previous patch which decode guest expected refresh rate from
vreg, the vblank emulator refactor patch makes following changes:
- Change the vblank emulator hrtimer from gvt global to per-vGPU.
By doing this, each vGPU display can operates at different refresh
rates. Currently only one dislay is supported for each vGPU so per-vGPU
hrtimer is enough. If multiple displays are supported per-vGPU in
future, we can expand to per-PIPE further.
- Change the fixed hrtimer period from 16ms to dynamic based on vreg.
GVT is expected to emulate the HW as close as possible. So reflacting
the accurate vsync interrupt interval is more correct than fixed 16ms.
- Change the vblank timer period and start the timer on PIPECONF change.
The initial period is updated to 16666667 based on 60Hz refresh rate.
According to PRM, PIPECONF controls the timing generator of the
connected display on this pipe, so it's safe to stop hrtimer on
PIPECONF disabling, and re-start hrtimer at new period on enabling.
Other changes including:
- Move vblank_timer_fn from irq.c into display.c.
- Clean per-vGPU vblank timer at clean_display instead of clean_irq.
To run quick test, launch a web browser and goto URL: www.displayhz.com
The actual refresh rate from guest can now always match guest settings.
V2:
Rebase to 5.11.
Remove unused intel_gvt_clean_irq().
Simplify enable logic in update_vblank_emulation(). (zhenyu)
Loop all vGPU by idr when check all vblank timer. (zhenyu)
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/20210226044630.284269-1-colin.xu@intel.com
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Guest OS builds up its timing mode list based on the virtual EDID as
simulated by GVT. However since there are several timings supported in
the virtual EDID, and each timing can also support several modes
(resolution and refresh rate), current emulated vblank period (16ms)
may not always be correct and could lead to miss-sync behavior in guest.
Guest driver will setup new resolution and program vregs accordingly and
it should always follows GEN PRM. Based on the simulated display regs by
GVT, it's safe to decode the actual refresh rate using by guest from
vreg only.
Current implementation only enables PIPE_A and PIPE_A is always tied to
TRANSCODER_A in HW. GVT may simulate DP monitor on PORT_B or PORT_D
based on the caller. So we can find out which DPLL is used by PORT_x
which connected to TRANSCODER_A and calculate the DP bit rate from the
DPLL frequency. Then DP stream clock (pixel clock) can be calculated
from DP link M/N and DP bit rate. Finally, get the refresh rate from
pixel clock, H total and V total.
The per-vGPU accurate refresh rate is not used yet but only stored,
until per-vGPU vblank timer is enabled. Then each vGPU can have
different and accurate refresh rate per-guest driver configuration.
Refer to PRM for GEN display and VESA timing standard for more details.
V2:
Rebase to 5.11.
Correctly calculate DP link rate for BDW and BXT.
Use GVT_DEFAULT_REFRESH_RATE instead of hardcoded to 60 as init refresh.
Typo fix. (zhenyu)
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/20210226044559.283622-1-colin.xu@intel.com
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
In Bspec the TGL TypeC ports are TC1-6, the AUX power well request flags
are USBC1-6/TBT1-6, so for clarity use these names in the port power
domain names instead of the D-I terminology (which Bspec uses only for
the ICL TypeC ports).
A domain name should follow the <domain>_<pipe/transcoder/port/aux_ch>
format. Add the new aliases based on this, leaving a change to rename
all the rest accordingly for a follow-up.
No functional change.
v2: Add comment to commit log about unifying domain names. (Jose)
Cc: Souza Jose <jose.souza@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210222210400.940158-1-imre.deak@intel.com
Even though GEN12+ HW supports PSR + DC3CO, DMC's HW DC3CO exit mechanism
has an issue with using of Selective Fecth and PSR2 manual tracking.
And as some GEN12+ platforms (RKL, ADL-S) don't support PSR2 HW tracking,
Selective Fetch will be enabled by default on that platforms.
Therefore if the system enables PSR Selective Fetch / PSR manual tracking,
it does not allow DC3CO dc state, in that case.
When this DC3CO exit issue is addressed while PSR Selective Fetch is
enabled, this restriction should be removed.
v2: Address Jose's review comment.
- Fix typo
- Move check routine of DC3CO ability to
tgl_dc3co_exitline_compute_config()
v3: Change the check routine of enablement of psr2 sel fetch. (Jose)
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Signed-off-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210222213006.1609085-1-gwan-gyeong.mun@intel.com
There is no support for two pipes one transcoder for PSR and if we had
that the current code should not use cpu_transcoder.
Also I can't see a scenario where crtc_state->enable_psr2_sel_fetch is
set and PSR is not enabled and if by a bug it happens PSR HW will just
ignore any value in set in PSR2_MAN_TRK_CTL.
So dropping all the rest and keeping the same behavior that we have
with intel_psr2_program_plane_sel_fetch().
Cc: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210209181439.215104-2-jose.souza@intel.com
We tend to use output_format!=RGB as a shorthand for YCbCr, but
this fails if we have a disabled crtc where output_format==INVALID.
We're now getting some fail from intel_color_check() when we have:
hw.enable==false
hw.ctm!=NULL
output_format==INVALID
Let's avoid that by throwing INTEL_OUTPUT_FORMAT_INVALID to the
dumpster, and thus everything defaults to RGB when the crtc
is disabled.
This does beg the deeper question of how much of the state
should we in fact be validating when hw/uapi.enable==false.
And should we even be doing the uapi->hw copy when
uapi.enable==false? So far I've not been able to come up with
satisfactory answers for myself, so I'm putting it off for the
moment.
Cc: Lee Shawn C <shawn.c.lee@intel.com>
Fixes: 0aa5c3835c ("drm/i915: support two CSC module on gen11 and later")
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2964
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210205202322.27608-1-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Instead of every new platform having yet another masive
copy of the whole PLL sanitation code, let's just reuse the
.disable_clock() hook for this purpose. We do need to plug
this into the ICL+ DSI code for that, but fortunately it
already has a suitable function we can use.
We do lose the debug message though on account of not bothering
to check if the clock is actually enabled or not before turning
it off. We could introduce yet another vfunc to query the current
state, but not sure it's worth the hassle?
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210205214634.19341-14-ville.syrjala@linux.intel.com
For ICL+ we have several styles of clock routing for DDIs:
1) TC DDI + TC PHY
-> needs DDI_CLK_SEL==MG/TBT part form intel_ddi_clk_{select,disable}()
and ICL_DPCLKA_CFGCR0_TC_CLK_OFF part form icl_{map,unmap}_plls_to_ports()
2) ICL/TGL combo DDI + combo PHY
-> just need the stuff from icl_{map,unmap}_plls_to_ports()
3) JSL/EHL TC DDI + combo PHY
-> needs DDI_CLK_SEL==MG part from intel_ddi_clk_{select,disable}() and
the full combo style clock selection from icl_{map,unmap}_plls_to_ports()
4) ADLS/RKL
-> these use both TC and combo DDIs with combo PHYs, however they
always use the full combo style clock selection as per
icl_{map,unmap}_plls_to_ports() and do not use DDI_CLK_SEL at all,
thus get treated the same as 2)
We extract all that from the current mess in the following way:
1) icl_ddi_tc_{enable,disable}_clock()
2) icl_ddi_combo_{enable,disable}_clock()
3) jsl_ddi_tc_{enable,disable}_clock()
4) for now we reuse icl_ddi_combo_{enable,disable}_clock() here
v2: s/dev_priv/i915/ (Lucas)
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210205214634.19341-8-ville.syrjala@linux.intel.com
intel_gt_driver_register() may be called earlier than
intel_opregion_register() and acpi_video_register(), so move it up.
intel_display_debugfs_register() may be called later, together with the
other display-related initializations. There is a slight change in
behavior that sysfs files will show up before the display-related
debugfs files, but that shouldn't be a problem - userspace shouldn't be
relying in debugfs.
This allows us to group all the display-related calls under a single
check for "HAS_DISPLAY()" that can be later moved to a better place.
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210213042756.953007-2-lucas.demarchi@intel.com
In rare circumstances bugs in PCI programming, broken BIOS, or failing
hardware can cause the CPU to lose access to the MMIO BAR on dgfx
platforms. This is a pretty catastrophic failure since all register
reads come back with values of 0xFFFFFFFF. Let's check for this special
case while doing our usual checks for unclaimed registers; the FPGA_DBG
register we use for those checks on modern platforms has some unused
bits that will always read back as 0 when things are behaving properly;
we can use them as canaries to detect when MMIO itself has suddenly
broken and try to print a more informative error message in the logs.
v2: Let the detection function still return 'true' if we've lost our
MMIO access. We'll still get an extra false positive message about
an unclaimed register access, but we'll still honor the 'mmio_debug'
limit and not spam the log. (Lucas)
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210212211925.3418280-2-matthew.d.roper@intel.com
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Although the bspec's description doesn't make it very clear, the
hardware architects have confirmed that the FPGA_DBG register that we
use to check for unclaimed MMIO accesses is display-specific and will
only properly flag unclaimed MMIO transactions for registers in the
display range. If a platform doesn't have display, FPGA_DBG itself will
not be available and should not be checked. Let's move the feature flag
into intel_device_info.display to more accurately reflect this.
Given that we now know FPGA_DBG is display-specific, it could be argued
that we should only check it on out intel_de_*() functions. However
let's not make that change right now; keeping the checks in all of the
existing locations still helps us catch cases where regular
intel_uncore_*() functions use bad MMIO offset math / base addresses and
accidentally wind up landing within an unused area within the display
MMIO range. It will also help catch cases where userspace-initiated
MMIO (e.g., IGT's intel_reg tool) attempt to read bad offsets within the
display range.
v2: Add missing hunk with the update to the HAS_FPGA_DBG_UNCLAIMED
macro. (CI)
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210212222049.3516344-1-matthew.d.roper@intel.com
Next, let's start introducing the HPD pin mappings for Intel's new gen9_bc
platform in order to make hotplugging display connectors work. Since
gen9_bc is just a TGP PCH along with a CML CPU, except with the same HPD
mappings as ICL, we simply add a skl_hpd_pin function that is shared
between gen9 and gen9_bc which handles both the traditional gen9 HPD pin
mappings and the Icelake HPD pin mappings that gen9_bc uses.
Changes since v4:
* Split this into its own commit
* Introduce skl_hpd_pin() like vsyrjala suggested and use that instead of
sticking our HPD pin mappings in TGP code
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
[originally from Tejas's work]
Signed-off-by: Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210209212832.1401815-4-lyude@redhat.com
