Upcoming commits make supervisor handling share code between the NV50
and GF119 implementations. Because of this, and a few other cleanups,
we need to allow some additional customisation.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
In order to properly support the SOR -> SOR + pad macro separation
that occurred with GM20x GPUs, we need to separate OR handling out
of the output path code.
This will be used as the base to support ORs (DAC, SOR, PIOR).
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Primarily intended as a way to pass per-head state around during
supervisor handling, and share logic between NV50/GF119.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This is to allow hw-specific code to instantiate output resources first,
so we can cull unsupported output paths based on them.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Not all users of nvkm_output_dp have been changed here. The remaining
ones belong to code that's disappearing in upcoming commits.
This also modifies the debug level of some messages.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This isn't technically "output", but, "display/output path".
Not all users of nvkm_output have been changed here. The remaining
ones belong to code that's disappearing in upcoming commits.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Upcoming changes to split OR from output path drastically change the
placement of various operations.
In order to make the real changes clearer, do the moving around part
ahead of time.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
As of DCB 4.1, these are not the same thing.
Compatibility temporarily in place until callers have been updated.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
We already have a subdev pointer, from which we can locate the device's
BIOS subdev. No need for a separate pointer.
Structure/callers not updated yet, as I want to batch more changes and
only touch the callers once.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
nvkm_timer_alarm() already handles this as part of protecting against
callers passing in no timeout value.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
I only saw those values inside the vbios: 0xff, 0xfd, 0xfc, 0xfa for valid
rails.
No idea what the lower value does, but at least we get power readings on
a lot of Fermi GPUs with that.
v2: add missing parentheses
Signed-off-by: Karol Herbst <karolherbst@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
This is according to what we have in nvbios.
Fixes "ERROR: Can't get value of subfeature in0_min: Can't read" errors
in sensors for some GPUs.
Signed-off-by: Karol Herbst <karolherbst@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Enable stereoscopic output for HDMI and DisplayPort connectors on
NV50+ (G80+) hardware. We do not enable stereoscopy on older
hardware in case there is some older board that still has HDMI
output but for which we have no logic for setting the Vendor
InfoFrame.
With this, I get an obvious 3D output when using the "testdisplay"
program from intel-gpu-tools with the "-3" parameter and outputting
to a 3D-capable HDMI display, for all available 3D modes (be they
TB, SBSH, or FP) on all four G80+ DISPs.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Frame-packing modes add an extra vtotal raster lines to each frame
above and beyond what the basic mode description calls for.
Account for this during scaler configuration (possibly a bit of a
hack), during CRTC configuration (clearly not a hack), and when
checking that a mode is valid for a given connector (cribbed from
the i915 driver).
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Now that we have the InfoFrame data being provided, for the most
part, program the hardware to use it.
While we're here, and since the functionality will come in handy
for supporting 3D stereoscopy, implement setting the Vendor
("generic"?) InfoFrame.
Also don't enable any InfoFrame that is not provided, and disable
the Vendor InfoFrame when disabling the output.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Now that we have the InfoFrame data being provided, for the most
part, program the hardware to use it.
While we're here, and since the functionality will come in handy
for supporting 3D stereoscopy, implement setting the Vendor
("generic"?) InfoFrame.
Also don't enable any InfoFrame that is not provided, and disable
the Vendor InfoFrame when disabling the output.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Now that we have the InfoFrame data being provided, for the most
part, program the hardware to use it.
While we're here, and since the functionality will come in handy
for supporting 3D stereoscopy, implement setting the Vendor
("generic") InfoFrame.
Also don't enable any AVI or Vendor InfoFrame that is not provided,
and disable the Vendor InfoFrame when disabling the output.
Ignore the Audio InfoFrame: We don't supply it, and altering HDMI
audio semantics (for better or worse) on this hardware is out of
scope for me at this time.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Now that we have the InfoFrame data being provided, for the most
part, program the hardware to use it.
While we're here, and since the functionality will come in handy
for supporting 3D stereoscopy, implement setting the Vendor
("generic"?) InfoFrame.
Also don't enable any AVI or Vendor InfoFrame that is not provided,
and disable the Vendor InfoFrame when disabling the output.
Ignore the Audio InfoFrame: We don't supply it, and altering HDMI
audio semantics (for better or worse) on this hardware is out of
scope for me at this time.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
HDMI InfoFrames are passed to NVKM as bags of bytes, but the
hardware needs them to be packed into words. Rather than having
four (or more) copies of the packing logic introduce a single copy
now, in a central place.
We currently need these for AVI and Vendor InfoFrames, but we may
also expect to need them for Audio InfoFrames at some point.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Now that we have mechanism by which to pass mode-dependent HDMI
InfoFrames to the low-level hardware driver, it is incumbent upon
us to do so.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
The nouveau driver, in the Linux 3.7 days, used to try and set the
AVI InfoFrame based on the selected display mode. These days, it
uses a fixed set of InfoFrames. Start to correct that, by
providing a mechanism whereby InfoFrame data may be passed to the
NVKM functions that do the actual configuration.
At this point, only establish the new parameters and their parsing,
don't actually use the data anywhere yet (since it's not supplied
anywhere).
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
drm_mode_set_crtcinfo() does compensation for interlace and
doublescan timing effects already, so do it first and use the
compensated figures instead of the constant "vscan / ilace" terms
that we had before.
And then it turns out that the hardware model for how the timing
parameters are configured is basically the standard model, but
starting one clock before the sync pulse rather than at the start
of the display area, which lets us drastically simplify the
overall timing calculations (verifying the changes by algebraic
operations is left as an exercise for the reader).
Finally, there were a couple of issues with the computation of
m->v.blankus that are addressed here. Interlaced modes would
generate a negative intermediate result. Double scan modes would
generate an overestimate rather than an underestimate. And when
enabling frame-packing modes, a rather extreme overestimate would
be generated. Fixed, by using the timings as adjusted for the
CRTC to find the length of the vertical blanking period instead of
mixing adjusted and pre-adjustment timing parameters.
Signed-off-by: Alastair Bridgewater <alastair.bridgewater@gmail.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Tariq Toukan says:
====================
mlx4 XDP performance improvements
This patchset contains data-path improvements, mainly for XDP_DROP
and XDP_TX cases.
Main patches:
* Patch 2 by Saeed allows enabling optimized A0 RX steering (in HW) when
setting a single RX ring.
With this configuration, HW packet-rate dramatically improves,
reaching 28.1 Mpps in XDP_DROP case for both IPv4 (37% gain)
and IPv6 (53% gain).
* Patch 6 enhances the XDP xmit function. Among other changes, now we
ring one doorbell per NAPI. Patch gives 17% gain in XDP_TX case.
* Patch 7 obsoletes the NAPI of XDP_TX completion queue and integrates its
poll into the respective RX NAPI. Patch gives 15% gain in XDP_TX case.
Series generated against net-next commit:
f7aec129a3 rxrpc: Cache the congestion window setting
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Some code re-ordering, functionally equivalent.
- The !tx_info->inl check is evaluated anyway in both flows
(common case/end case). Run it first, this might finish
the flows earlier.
- dma_unmap calls are identical in both flows, get it out
of the if block into the common area.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Gain is too small to be measurable, no degradation sensed.
Results are similar for IPv4 and IPv6.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Define LOG_TXBB_SIZE, log of TXBB_SIZE, and use it with a shift
operation instead of a multiplication with TXBB_SIZE.
Operations are equivalent as TXBB_SIZE is a power of two.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Gain is too small to be measurable, no degradation sensed.
Results are similar for IPv4 and IPv6.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Increase the default TX ring size (from 512 to 1024) to match
the RX ring size.
This gives the XDP TX ring a better chance to keep up with the
rate of its RX ring in case of a high load of XDP_TX actions.
Tested:
Ethtool counter rx_xdp_tx_full used to increase, after applying this
patch it stopped.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of having their own NAPIs, XDP TX completion queues get
polled within the corresponding RX NAPI.
This prevents any possible race on TX ring prod/cons indices,
between the context that issues the transmits (RX NAPI) and the
context that handles the completions (was previously done in
a separate NAPI).
This also improves performance, as it decreases the number
of NAPIs running on a CPU, saving the overhead of syncing
and switching between the contexts.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Single queue no-RSS optimization ON.
XDP_TX packet rate:
-------------------------------------
| Before | After | Gain |
IPv4 | 12.0 Mpps | 13.8 Mpps | 15% |
IPv6 | 12.0 Mpps | 13.8 Mpps | 15% |
-------------------------------------
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several performance improvements in XDP TX datapath,
including:
- Ring a single doorbell for XDP TX ring per NAPI budget,
instead of doing it per a lower threshold (was 8).
This includes removing the flow of immediate doorbell ringing
in case of a full TX ring.
- Compiler branch predictor hints.
- Calculate values in compile time rather than in runtime.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Single queue no-RSS optimization ON.
XDP_TX packet rate:
-------------------------------------
| Before | After | Gain |
IPv4 | 10.3 Mpps | 12.0 Mpps | 17% |
IPv6 | 10.3 Mpps | 12.0 Mpps | 17% |
-------------------------------------
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several small code and performance improvements in stack TX datapath,
including:
- Compiler branch predictor hints.
- Minimize variables scope.
- Move tx_info non-inline flow handling to a separate function.
- Calculate data_offset in compile time rather than in runtime
(for !lso_header_size branch).
- Avoid trinary-operator ("?") when value can be preset in a matching
branch.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Gain is too small to be measurable, no degradation sensed.
Results are similar for IPv4 and IPv6.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several small performance improvements in TX CQ polling,
including:
- Compiler branch predictor hints.
- Minimize variables scope.
- More proper check of cq type.
- Use boolean instead of int for a binary indication.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Packet-rate tests for both regular stack and XDP use cases:
No noticeable gain, no degradation.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several small performance improvements in RX datapath,
including:
- Compiler branch predictor hints.
- Replace a multiplication with a shift operation.
- Minimize variables scope.
- Write-prefetch for packet header.
- Avoid trinary-operator ("?") when value can be preset in a matching
branch.
- Save a branch by updating RX ring doorbell within
mlx4_en_refill_rx_buffers(), which now returns void.
Performance tests:
Tested on ConnectX3Pro, Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Single queue no-RSS optimization ON
(enable by ethtool -L <interface> rx 1).
XDP_DROP packet rate:
Same (28.1 Mpps), lower CPU utilization (from ~100% to ~92%).
Drop packets in TC:
-------------------------------------
| Before | After | Gain |
IPv4 | 4.14 Mpps | 4.18 Mpps | 1% |
-------------------------------------
XDP_TX packet rate:
-------------------------------------
| Before | After | Gain |
IPv4 | 10.1 Mpps | 10.3 Mpps | 2% |
IPv6 | 10.1 Mpps | 10.3 Mpps | 2% |
-------------------------------------
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Cc: kernel-team@fb.com
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Saeed Mahameed says:
====================
Mellanox mlx5 fixes 2017-06-14
This series contains some fixes for the mlx5 core and netdev driver.
Please pull and let me know if there's any problem.
For -stable:
("net/mlx5: Wait for FW readiness before initializing command interface") kernels >= 4.4
("net/mlx5e: Fix timestamping capabilities reporting") kernels >= 4.5
("net/mlx5e: Avoid doing a cleanup call if the profile doesn't have it") kernels >= 4.9
("net/mlx5e: Fix min inline value for VF rep SQs") kernels >= 4.11
The "net/mlx5e: Fix min inline .." (a oneliner patch) doesn't cleanly apply
to 4.11, it hits a contextual conflict and can be easily resolved by:
+ mlx5_query_min_inline(mdev, &priv->params.tx_min_inline_mode);
to the end of mlx5e_build_rep_netdev_priv. Note the 2nd parameter of
mlx5_query_min_inline is slightly different from the original one.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
At present, HV KVM on POWER8 and POWER9 machines loses any instruction
or data breakpoint set in the host whenever a guest is run.
Instruction breakpoints are currently only used by xmon, but ptrace
and the perf_event subsystem can set data breakpoints as well as xmon.
To fix this, we save the host values of the debug registers (CIABR,
DAWR and DAWRX) before entering the guest and restore them on exit.
To provide space to save them in the stack frame, we expand the stack
frame allocated by kvmppc_hv_entry() from 112 to 144 bytes.
Fixes: b005255e12 ("KVM: PPC: Book3S HV: Context-switch new POWER8 SPRs", 2014-01-08)
Cc: stable@vger.kernel.org # v3.14+
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
sysfs-firmware-ofw describes the /sys/firmware/devicetree/ hierarchy
and /proc/device-tree so add a file line for it to the entry
OPEN FIRMWARE AND FLATTENED DEVICE TREE
Signed-off-by: Frank Rowand <frank.rowand@sony.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Make the three tests that did use the old ksft_ext_skip()
(breakpoints/breakpoint_test_arm64, breakpoints/step_after_suspend_test,
and membarrier_test) use the new one, with an output for the
reason for skipping all the tests.
Signed-off-by: Paul Elder <paul.elder@pitt.edu>
Signed-off-by: Shuah Khan <shuahkh@osg.samsung.com>
