Toss a bunch if constants into .rodata drom the stack. Also
shrink the types of some of the arrays to reduce the size.
bloat-o-meter -c intel_dpll_mgr.o:
add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-86 (-86)
Function old new delta
icl_get_dplls 3393 3372 -21
skl_get_dpll 2069 2004 -65
Total: Before=28029, After=27943, chg -0.31%
add/remove: 0/0 grow/shrink: 0/0 up/down: 0/0 (0)
Data old new delta
Total: Before=17, After=17, chg +0.00%
add/remove: 2/0 grow/shrink: 0/2 up/down: 28/-129 (-101)
RO Data old new delta
dco_central_freq - 24 +24
div1_vals - 4 +4
odd_dividers 28 7 -21
even_dividers 144 36 -108
Total: Before=3600, After=3499, chg -2.81%
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220301173128.6988-3-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Now that we track framestart_delay in the crtc state with readout
and state checker support we can remove the explicit framestart_delay
sanitation code.
Also I'm not convinced reprogramming this while the pipe is running
is even valid. CHICKEN_TRANS (hsw+) and TRANS_CHICKEN2 (cpt+) docs
at least make no mention of double buffering which seems to imply
that live reprogramming is not supported. On older platforms
PIPECONF and PCH_TRANSCONF (ibx) are double buffered though, so
might be that we could do this on the older platforms. But doesn't
really make sense to special case old platforms for this.
So from now on if the BIOS has misprogrammed this we shall simply do
a full modeset at boot to fix it up. Such systems will of course lose
fastboot, but I think less code (and less uncertainty what
reprogramming this on a running pipe will even do) outweighs that.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220221110356.5532-2-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
The current implementation of the async flip wm0/ddb optimization
does not work at all. The biggest problem is that we skip the
whole intel_pipe_update_{start,end}() dance and thus never actually
complete the commit that is trying to do the wm/ddb change.
To fix this we need to move the do_async_flip flag to the crtc
state since we handle commits per-pipe, not per-plane.
Also since all planes can now be included in the first/last
"async flip" (which gets converted to a sync flip to do the
wm/ddb mangling) we need to be more careful when checking if
the plane state is async flip comptatible. Only planes doing
the async flip should be checked and other planes are perfectly
fine not adhereing to any async flip related limitations.
However for subsequent commits which are actually going do the
async flip in hardware we want to make sure no other planes
are in the state. That should never happen assuming we did our
job correctly, so we'll toss in a WARN to make sure we catch
any bugs here.
Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Fixes: c3639f3be4 ("drm/i915: Use wm0 only during async flips for DG2")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220214105532.13049-4-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Since the async flip state check is done very late and
thus it can see potentially all the planes in the state
(due to the wm/ddb optimization) we need to move the
"can the requested plane do async flips at all?" check
much earlier. For this purpose we introduce
intel_async_flip_check_uapi() that gets called early during
the atomic check.
And for good measure we'll throw in a couple of basic checks:
- is the crtc active?
- was a modeset flagged?
- is+was the plane enabled?
Though atm all of those should be guaranteed by the fact
that the async flip can only be requested through the legacy
page flip ioctl.
Cc: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Fixes: c3639f3be4 ("drm/i915: Use wm0 only during async flips for DG2")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220214105532.13049-3-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Some users are suffering with PSR2 issues that are under debug or
issues that were root caused to panel firmware bugs, to make life of
those users easier here adding a option to disable PSR2 with kernel
parameters so they can still benefit from PSR1 power savings.
Using the same enable_psr that is current used to turn the whole
feature on or off and allowing user to select up to what PSR version
it should enable.
Right now users only set this parameter to 0 when they want to disable
PSR1 and PSR2 or don't add it at all leaving it to per-chip behavior
so it should not cause a bad impact on users.
v2:
- changing enable_psr values (Ville and Rodrigo)
Link: https://gitlab.freedesktop.org/drm/intel/-/issues/4951
Cc: Jouni Högander <jouni.hogander@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220224202523.993560-1-jose.souza@intel.com
Currently we are observing occasional screen flickering when
PSR2 selective fetch is enabled. More specifically glitch seems
to happen on full frame update when cursor moves to coords
x = -1 or y = -1.
According to Bspec SF Single full frame should not be set if
SF Partial Frame Enable is not set. This happened to be true for
ADLP as PSR2_MAN_TRK_CTL_ENABLE is always set and for ADL_P it's
actually "SF Partial Frame Enable" (Bit 31).
Setting "SF Partial Frame Enable" bit also on full update seems to
fix screen flickering.
Also make code more clear by setting PSR2_MAN_TRK_CTL_ENABLE
only if not on ADL_P. Bit 31 has different meaning in ADL_P.
Bspec: 49274
v2: Fix Mihai Harpau email address
v3: Modify commit message and remove unnecessary comment
Tested-by: Lyude Paul <lyude@redhat.com>
Fixes: 7f6002e580 ("drm/i915/display: Enable PSR2 selective fetch by default")
Reported-by: Lyude Paul <lyude@redhat.com>
Cc: Mihai Harpau <mharpau@gmail.com>
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Bugzilla: https://gitlab.freedesktop.org/drm/intel/-/issues/5077
Signed-off-by: Jouni Högander <jouni.hogander@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/20220225070228.855138-1-jouni.hogander@intel.com
Commit d5ce34da31
("drm/i915: Add state verification for the TypeC port mode")
added a verification to the TypeC AUX power well enable()/disable()
hooks to check if the TypeC port related to this power well is properly
locked. If the disabling happens asynchronously the verification is
skipped, since in this case the port is unlocked. The detection of
asnychronous disabling doesn't work as intended though, since the power
well's reference count is always 0 when its disable() hook is called
(and since there won't be any domain reference held for this power well
either, the verification is always skipped); remove the verification
from the disable() hook for now. In the power well's enable() hook the
power well's reference will be always >0 and there won't be any
asynchronous disabling pending for it, so we can drop the async refcount
check from there.
No functional change.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220222165137.1004194-4-imre.deak@intel.com
Instead of open-coding the call of the power wells' enable()/disable()
hooks use the corresponding helper functions. This will also ensure that
the power well's cached-enable state is always up-to-date. Luckily the
lack of this updating hasn't been a problem, since the state either
didn't change (in intel_display_power_set_target_dc_state()), or got
updated subsequently (for vlv_cmnlane_wa(), in the following
intel_power_domains_sync_hw()).
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220222165137.1004194-3-imre.deak@intel.com
The POWER_DOMAIN_TRANSCODER() macro depends on the
POWER_DOMAIN_TRANSCODER_A/B .. DSI_A/C enum values to be consecutive,
move POWER_DOMAIN_TRANSCODER_VDSC_PW2 after these to ensure this. The
wrong order didn't cause a problem, since the DSI_A/C domains are in
always-on power wells on all relevant platforms. The same power well
ends up being enabled/disabled when the VDSC_PW2 domain is selected
incorrectly.
While at it add a code comment about enum values that need to stay
consecutive.
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220222165137.1004194-2-imre.deak@intel.com
When calculating pipe_mode and when doing readout we need
to order our steps correctly.
1. We start with adjusted_mode crtc timings being populated
with the transcoder timings (either via readout or
compute_config(). These will be per-segment for MSO.
2. For all other uses we want the full crtc timings so
we ask intel_splitter_adjust_timings() to expand
the per-segment numbers to their full glory
3. If bigjoiner is used we the divide the full numbers
down to per-pipe numbers using intel_bigjoiner_adjust_timings()
During readout we also have to reconstruct the adjusted_mode
normal timings (ie. not the crtc_ stuff). These are supposed
to reflect the full timings of the display. So we grab these
between steps 2 and 3.
The "user" mode readout (mainly done for fastboot purposes)
should be whatever mode the user would have used had they
asked us to do a modeset. We want the full timings for this
as the per-segment timings are not suppoesed to be user visible.
Also the user mode normal timings hdisplay/vdisplay need to
match PIPESRC (that is where we get our PIPESRC size
we doing a modeset with a user supplied mode).
And we end up with
- adjusted_mode normal timigns == full timings
- adjusted_mode crtc timings == transcoder timings
(per-segment timings for MSO, full timings otherwise)
- pipe_mode normal/crtc timings == pipe timings
(full timings divided by the number of bigjoiner pipes, if any)
- user mode normal timings == full timings with
hdisplay/vdisplay replaced with PIPESRC size
- user mode crtc timings == full timings
Yes, that is a lot of timings. One day we'll try to remove
some of the ones we don't actually need to keep around...
Reviewed-by: Manasi Navare <manasi.d.navare@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220223131315.18016-10-ville.syrjala@linux.intel.com
bigjoiner_pipes==0 leads bigjoiner_master_pipe() to
do BIT(ffs(0)-1) which is undefined behaviour. The code should
actually still work fine since the only place we provoke
that is intel_crtc_bigjoiner_slave_pipes() and it'll bitwise
AND the result with 0, so doesn't really matter what we get
out of bigjoiner_master_pipe(). But best not provoke undefined
behaviour anyway.
Reported-by: kernel test robot <oliver.sang@intel.com>
Fixes: a6e7a006f5 ("drm/i915: Change bigjoiner state tracking to use the pipe bitmask")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220223131315.18016-2-ville.syrjala@linux.intel.com
Reviewed-by: Manasi Navare <manasi.d.navare@intel.com>
Checking by >= DISPLAY_VER(12) made sense when the support for Tiger
Lake was added. However now it only leads to wrong behavior when adding
more platforms since it's expected they either don't have DMC to load
or they have their own blob.
Logs from DG2 loading on a CFL host, without having a DMC firmware
defined:
<6>[ 0.000000] DMI: Intel Corporation CoffeeLake Client Platform/CoffeeLake S UDIMM RVP, BIOS CNLSFWR1.R00.X220.B00.2103302221 03/30/2021
...
<6>[ 2.706607] pci 0000:03:00.0: [8086:56a0] type 00 class 0x030000
...
<7>[ 6.340397] i915 0000:03:00.0: [drm:intel_dmc_ucode_init [i915]] Loading i915/tgl_dmc_ver2_12.bin
<7>[ 6.341841] i915 0000:03:00.0: [drm:intel_fbc_init [i915]] Sanitized enable_fbc value: 1
<3>[ 6.342432] genirq: Flags mismatch irq 0. 00000080 (i915) vs. 00015a00 (timer)
<6>[ 6.346283] i915 0000:03:00.0: [drm] Finished loading DMC firmware i915/tgl_dmc_ver2_12.bin (v2.12)
<3>[ 6.385756] i915 0000:03:00.0: Device initialization failed (-16)
<5>[ 6.385778] i915 0000:03:00.0: Please file a bug on drm/i915; see https://gitlab.freedesktop.org/drm/intel/-/wikis/How-to-file-i915-bugs for details.
<4>[ 6.385782] i915: probe of 0000:03:00.0 failed with error -16
TGL is the only platform left with DISPLAY_VER() == 12 that is not
handled already in the if/else ladder, so handle it specifically.
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220223210933.3049143-1-lucas.demarchi@intel.com
Drop the locks around most primary plane register writes.
The lock isn't needed since each plane's register are neatly
contained on their own cachelines.
The one exception we have to make is DSPADDR/DSPSURF which is
(ab)used to also trigger FBC nukes on pre-snb (since the
hardware doesn't seem to have any dedicated mechanism to
trigger nukes). So we need to keep the lock around it to
protect against the rmw performed by the fbc code.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220210062403.18690-5-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Drop the locks around cursor plane register writes. The
lock isn't needed since each plane's register are neatly
contained on their own cachelines.
The locking did have a secondary effect of disabling
interrupts around the cursor registers writes though.
If we drop that then we open outselves up for sceduling
delays and whatnot while on the middle of the register
writes. That increases the chance of not all the register
writes land during the same frame. For normal atomic
commits this is not a concern as the vblank evade mechanism
anyway disables interrupts around the update, but the legacy
cursor codepath does not. Technically we should do a vblank
evade there as well, but so far no one has bothered to hook
that up. So in the meantime let's put an explicit local irq
disable/enable around the legacy cursor update to keep the
race window minimal.
v2: local_irq_{disable,enable}() for legacy cursor ioctl
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220211092604.393-1-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
On icl+ all plane registers are armed by PLANE_SURF, so we can
move almost everything over into the update_noarm() hook.
The PLANE_CTL write has to stay in the icl_update_arm() hook though
as it still exhibits the somewhat annoying self-arming behaviour
when the plane transitioning from disabled to enabled.
We could either do a full split for skl+ vs. icl+, or we could try
some other kind of split where we'd eg. keep most things in the skl+
functions and call them from the icl+ functions. I think a full split
is probably the cleaner approach since we've anyway accumulated quite
a bit of icl+ specific things, so that is what I opted to do.
Some i915_update_info stats for tgl:
before: after:
Updates: 5043 Updates: 5043
| |
1us | 1us |
|** |***
4us |****** 4us |********
|********** |***********
16us |*********** 16us |**********
|**** |*
66us | 66us |
| |
262us | 262us |
| |
1ms | 1ms |
| |
4ms | 4ms |
| |
17ms | 17ms |
| |
Min update: 3494ns Min update: 2983ns
Max update: 49491ns Max update: 39986ns
Average update: 18031ns Average update: 13423ns
Overruns > 100us: 0 Overruns > 100us: 0
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220210062403.18690-2-ville.syrjala@linux.intel.com
Reviewed-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
