similar to: [PATCH 0/2] drm/nouveau: avoid usage of list iterator after loop

Reviewed-by: Lyude Paul <lyude at redhat.com> Will push upstream in just a moment On Wed, 2023-03-01 at 18:25 +0100, Jakob Koschel wrote: > This patch set includes two instances where the list iterator variable > 'pstate' is implicitly assumed to be valid after the iterator loop. > While in pratice that is most likely the case (if >

On Wed, 2023-03-01 at 18:25 +0100, Jakob Koschel wrote: > If potentially no valid element is found, 'pstate' would contain an > invalid pointer past the iterator loop. To ensure 'pstate' is always > valid, we only set it if the correct element was found. That allows > adding a BUG_ON in case the code works incorrectly, exposing currently > undetectable potential

This series addresses various issues inside the reclocking code: 1. after resume the set clocks are reset 2. reclocking not possible while GPU is suspended 3. nouveau always does full reclocks even if only a change of the voltage is required Some of the patches were part of the bigger reclocking series I sent months ago, some things have changed though. This is also preparation work of

This series addresses various issues inside the reclocking code: 1. after resume the set clocks are reset 2. reclocking not possible while GPU is suspended Some of the patches were part of the bigger reclocking series I sent months ago, some things have changed though. This is also preparation work of changing the clock state due to temperature changes and dynamic reclocking. v2: remove commits

This series addresses various issues inside the reclocking code: 1. after resume the set clocks are reset 2. reclocking not possible while GPU is suspended 3. racy reclocking while GPU is suspending and leading to hangs Some of the patches were part of the bigger reclocking series I sent months ago, some things have changed though. This is also preparation work of changing the clock state due to

On 2017-09-15 — 17:11, Karol Herbst wrote: > This makes the code easier, because we can compare the id with > pstate->pstate and saves us from the trouble of iterating over the pstates > to match the index. I don’t remember whether I have already done this comment before, but I am not sure where your iterating over the pstates savings are, as in this patch at least, you iterate as

This is an updated series for the old clocking improvement one. I think I got everything needed in place and also a simple update mechanism for updating the cstates/voltage on temperature changes. If anything is unclear how I REed or got the information, please leave a note so that I can provide additional information in the commits. Besides that I think we are pretty close now and only some

The patch seems fine but I found it super confusing that sometimes `pstate` is a pointer (for example `clk->pstate`), sometimes it is an int (for example `args->v0.pstate`). On 2017-09-15 — 17:11, Karol Herbst wrote: > We will access the current cstate at least every second and this saves us > some CPU cycles looking them up every second. > > v2: Rewording commit message. >

On 18/04/16 22:14, Karol Herbst wrote: > we will access the current set cstate at least every second and this safes us saves > some CPU cycles looking them up every second. > > Signed-off-by: Karol Herbst <nouveau at karolherbst.de> > --- > drm/nouveau/include/nvkm/subdev/clk.h | 2 +- > drm/nouveau/nvkm/engine/device/ctrl.c | 5 ++++- >

Just wanted to post updated versions of my last series/patches. Reviews welcomed. It would be also nice if we agree on features I should focus upstreaming, so that this work can be better splitted or reordered. Sadly most of my patches depend on the rather big clk subdev rework and I think those patches shows best, why I think this rework is actually needed and makes things much easier to add

Adds Nouveau controlled thermal throttling for Kepler+ GPUs. With this I feel safe enough to add support for Maxwell2 reclocking later on (still hidden behind a switch, but we can be fairly sure to not overheat hardware if a user isn't carefull enough) Contains all patches from my clk update series, but I thought it makes sense to include those in this series as well for completness. Please

Last update here: https://lists.freedesktop.org/archives/nouveau/2017-September/028848.html Basically big cleanup, reordering, simplifications and some renaming to make the code easier to read and to review. I also moved some bugfixes to the front so they can be merged prior the other patches. There was also a bug related to the therm daemon triggering a pstate change leading to PMU lockups,

We are slowly getting there! v4 of the series with some realy good improvements, so I am sure this is like 95% done and only needs some proper polishing and proper Reviews! I also added the NvVoltOffsetmV module parameter, so that a user is able to over and !under!-volt the GPU. Overvolting makes sense, when there are still some reclocking issues left, which might be solved by a higher voltage.

This series fixes most of the issues regarding volting on GPUs with any form of GPU Boost inside their vbios, which is mainly Kepler and newer, but we find some boosting related tables in Fermi vbios' already In the end reclocking should work on most Kepler cards without any issues v2: boost_mode can now be changed at runtime minor fixups Karol Herbst (22): bios/volt: handle voltage

this patch series implements PCIe speed changes for Tesla and newer. The Kepler and Fermi bits are tested on my cards at home. Karol Herbst (9): pci: add gk104 variant pci: add gf106 variant pci: implement generic code for PCIe speed change pci: implement pcie speed change for tesla pci: implement pcie speed change on Fermi pci: implement PCIe speed change for kepler+ bios/perf:

On Fri, 13 Sep 2019 at 21:33, Karol Herbst <kherbst at redhat.com> wrote: > > Apperantly things go south if we suspend the device with a PCIe link speed > set to 2.5. Fixes runtime suspend on my gp107. > > This all looks like some bug inside the pci subsystem and I would prefer a > fix there instead of nouveau, but maybe there is no real nice way of doing > that outside

not much changed since the last time I sent those patches out, but there are a couple of annoying bug fixes, which users would probably never hit unless they do rmmod/modprobe nouveau cycles. Biggest change is that I force the link to a 8.0 speed rather than the speed the GPU came up with. Also this series depends on the PCIe improvement patches I sent out recently. Karol Herbst (4): pci:

On Tue, 17 Sep 2019 at 18:07, Karol Herbst <kherbst at redhat.com> wrote: > > On Tue, Sep 17, 2019 at 8:01 AM Ben Skeggs <skeggsb at gmail.com> wrote: > > > > On Fri, 13 Sep 2019 at 21:33, Karol Herbst <kherbst at redhat.com> wrote: > > > > > > Apperantly things go south if we suspend the device with a PCIe link speed > > > set to 2.5.

I've splitted my big series between the part which actually fixes the engine reclocking bits and the part handling voltage/clock updates on temperature change, so that the more reviewed parts can be merged in faster. This series fixes a lot of Engine reclocking issues found on Fermi, Kepler and all Maxwell generation GPUs. It does _not_ fix memory reclocking on Fermi. It mostly contains of

overall it is for the most part the same as my older version. I cleaned up some copyright things, so that it is more like the others. Also I moved the print about the max speed supported into preinit and did some other minor cleanups in the 3rd commit. Happy testing (and performance for prime offloading setups) Karol Herbst (9): pci: add gk104 variant pci: add gf106 variant pci: implement