| Message ID | 20190814131408.57162-5-tony@atomide.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Configure sgx interconnect data for some omap variants | expand |
* Stephen Boyd <sboyd@kernel.org> [190815 18:24]: > Quoting Tony Lindgren (2019-08-14 06:14:06) > > diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi > > --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi > > +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi > > @@ -1146,6 +1146,20 @@ > > }; > > }; > > > > + gpu_cm: gpu_cm@1500 { > > Node names shouldn't have underscores. Maybe clock-controller? OK yeah clock-controller sounds good to me. Thanks, Tony
* Tony Lindgren <tony@atomide.com> [190817 06:56]: > * Stephen Boyd <sboyd@kernel.org> [190815 18:24]: > > Quoting Tony Lindgren (2019-08-14 06:14:06) > > > diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi > > > --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi > > > +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi > > > @@ -1146,6 +1146,20 @@ > > > }; > > > }; > > > > > > + gpu_cm: gpu_cm@1500 { > > > > Node names shouldn't have underscores. Maybe clock-controller? > > OK yeah clock-controller sounds good to me. Below is this one updated to use clock-controller naming. Regards, Tony 8< --------------------------- From tony Mon Sep 17 00:00:00 2001 From: Tony Lindgren <tony@atomide.com> Date: Wed, 14 Aug 2019 05:18:16 -0700 Subject: [PATCH] ARM: dts: Configure sgx for omap5 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I've tested that the interconnect target module enables and idles just fine when probed with ti-sysc with PM runtime control via sys: # echo on > $(find /sys -name control | grep \/5600) # rwmem 0x5600fe00 # OCP Revision 0x5600fe00 = 0x40000000 # echo auto > $(find /sys -name control | grep \/5600) # rwmem 0x5600fe10 # rwmem 0x56000024 Cc: Adam Ford <aford173@gmail.com> Cc: Filip Matijević <filip.matijevic.pz@gmail.com> Cc: "H. Nikolaus Schaller" <hns@goldelico.com> Cc: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com> Cc: moaz korena <moaz@korena.xyz> Cc: Merlijn Wajer <merlijn@wizzup.org> Cc: Paweł Chmiel <pawel.mikolaj.chmiel@gmail.com> Cc: Philipp Rossak <embed3d@gmail.com> Cc: Tomi Valkeinen <tomi.valkeinen@ti.com> Signed-off-by: Tony Lindgren <tony@atomide.com> --- arch/arm/boot/dts/omap5.dtsi | 23 +++++++++++++++++++++++ arch/arm/boot/dts/omap54xx-clocks.dtsi | 14 ++++++++++++++ 2 files changed, 37 insertions(+) diff --git a/arch/arm/boot/dts/omap5.dtsi b/arch/arm/boot/dts/omap5.dtsi --- a/arch/arm/boot/dts/omap5.dtsi +++ b/arch/arm/boot/dts/omap5.dtsi @@ -257,6 +257,29 @@ ports-implemented = <0x1>; }; + target-module@56000000 { + compatible = "ti,sysc-omap4", "ti,sysc"; + reg = <0x5600fe00 0x4>, + <0x5600fe10 0x4>; + reg-names = "rev", "sysc"; + ti,sysc-midle = <SYSC_IDLE_FORCE>, + <SYSC_IDLE_NO>, + <SYSC_IDLE_SMART>; + ti,sysc-sidle = <SYSC_IDLE_FORCE>, + <SYSC_IDLE_NO>, + <SYSC_IDLE_SMART>; + clocks = <&gpu_clkctrl OMAP5_GPU_CLKCTRL 0>; + clock-names = "fck"; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0 0x56000000 0x2000000>; + + /* + * Closed source PowerVR driver, no child device + * binding or driver in mainline + */ + }; + dss: dss@58000000 { compatible = "ti,omap5-dss"; reg = <0x58000000 0x80>; diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi @@ -1146,6 +1146,20 @@ }; }; + gpu_cm: clock-controller@1500 { + compatible = "ti,omap4-cm"; + reg = <0x1500 0x100>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0 0x1500 0x100>; + + gpu_clkctrl: clk@20 { + compatible = "ti,clkctrl"; + reg = <0x20 0x4>; + #clock-cells = <2>; + }; + }; + l3init_cm: l3init_cm@1600 { compatible = "ti,omap4-cm"; reg = <0x1600 0x100>;
Hi Tony, I found some time to update my drivers/stating/pvr/1.14.369993 tree to make use of the new sgx soc integration code and it works on omap5! > root@letux:~# modprobe pvrsrvkm_omap_omap5_sgx544_116 > [ 106.478254] pvrsrvkm_omap_omap5_sgx544_116: module is from the staging directory, the quality is unknown, you have been warned. > [ 106.526201] [drm] Initialized pvr 1.14.3699939 20110701 for 5600fe00.sgx on minor 1 > root@letux:~# pvrsrvctl --no-module --start > [ 261.091978] PVR_K: UM DDK-(3699939) and KM DDK-(3699939) match. [ OK ] > root@letux:~# gles1test1 0 > ^C > root@letux:~# To make it work, I have added a child node to the target-module@56000000 that describes the img,sgx5xx. Here are the work-in-progress patches if you want to take a look inside: https://github.com/openpvrsgx-devgroup/linux_openpvrsgx/commits/letux/omap-pvr-soc-glue-v2 On omap3630 the driver module loads but pvrsrvinit fails: > root@letux:~# modprobe pvrsrvkm_omap_omap3630_sgx530_125 > [ 390.429260] pvrsrvkm_omap_omap3630_sgx530_125: module is from the staging directory, the quality is unknown, you have been warned. > [ 390.474304] [drm] Initialized pvr 1.14.3699939 20110701 for 50000000.sgx on minor 1 > root@letux:~# pvrsrvctl --start --no-module > [ 402.200439] PVR_K: UM DDK-(3699939) and KM DDK-(3699939) match. [ OK ] > [ 402.219604] PVR_K:(Error): BridgedDispatchKM: Initialisation failed. Driver unusable. > PVR:(Error): LoaduKernelProgram : SGX ukernel program Device Addr: 0xe400000 invalid alignment [0, ] > PVR:(Error): SetupuKernel : Failed to load uKernel programs [0, ] > PVR:(Error): SrvInit: Initialisation for device of class 0, type 7, index 0, failed (1) [0, ] > PVR:(Error): PVRSRVBridgeCall: Failed to access device. Function ID:3223086862 (strerror returns no value.). [0, ] > pvrsrvctl: SrvInit failed (already initialized?) (err=PVRSRV_ERROR_OUT_OF_MEMORY - Unable to allocate required memory) > root@letux:~# Maybe I do not have done the address range translation properly. Or I am missing some of your PRM rstctrl patches since I am working on top of linus/master. Because of that I had to comment out prm_gfx in am33xx.dtsi to get it compile and could not test yet on Beagle Bone. A little unclear is how to properly handle omap4. omap4430/4440 and 4460 have the sgx540 while only omap4470 has an sgx544 inside. But omap4420/30 share one .dtsi and omap4460/70 the other. So we might have to define a new omap4470.dtsi and #include in the real omap4470 boards. And am3517 and am43/am57/am65/... support is missing. But that can be added later if we have it running on omap3 and am33xx. Anyways, we again have made a big step forwards. Let's hope that we can get the FLOSS part of the sgx drivers to staging in some not too far future. BR and thanks, Nikolaus > Am 26.08.2019 um 16:55 schrieb Tony Lindgren <tony@atomide.com>: > > * Tony Lindgren <tony@atomide.com> [190817 06:56]: >> * Stephen Boyd <sboyd@kernel.org> [190815 18:24]: >>> Quoting Tony Lindgren (2019-08-14 06:14:06) >>>> diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> @@ -1146,6 +1146,20 @@ >>>> }; >>>> }; >>>> >>>> + gpu_cm: gpu_cm@1500 { >>> >>> Node names shouldn't have underscores. Maybe clock-controller? >> >> OK yeah clock-controller sounds good to me. > > Below is this one updated to use clock-controller naming. > > Regards, > > Tony > > 8< --------------------------- > From tony Mon Sep 17 00:00:00 2001 > From: Tony Lindgren <tony@atomide.com> > Date: Wed, 14 Aug 2019 05:18:16 -0700 > Subject: [PATCH] ARM: dts: Configure sgx for omap5 > MIME-Version: 1.0 > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: 8bit > > I've tested that the interconnect target module enables and idles > just fine when probed with ti-sysc with PM runtime control via sys: > > # echo on > $(find /sys -name control | grep \/5600) > # rwmem 0x5600fe00 # OCP Revision > 0x5600fe00 = 0x40000000 > # echo auto > $(find /sys -name control | grep \/5600) > # rwmem 0x5600fe10 > # rwmem 0x56000024 > > Cc: Adam Ford <aford173@gmail.com> > Cc: Filip Matijević <filip.matijevic.pz@gmail.com> > Cc: "H. Nikolaus Schaller" <hns@goldelico.com> > Cc: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com> > Cc: moaz korena <moaz@korena.xyz> > Cc: Merlijn Wajer <merlijn@wizzup.org> > Cc: Paweł Chmiel <pawel.mikolaj.chmiel@gmail.com> > Cc: Philipp Rossak <embed3d@gmail.com> > Cc: Tomi Valkeinen <tomi.valkeinen@ti.com> > Signed-off-by: Tony Lindgren <tony@atomide.com> > --- > arch/arm/boot/dts/omap5.dtsi | 23 +++++++++++++++++++++++ > arch/arm/boot/dts/omap54xx-clocks.dtsi | 14 ++++++++++++++ > 2 files changed, 37 insertions(+) > > diff --git a/arch/arm/boot/dts/omap5.dtsi b/arch/arm/boot/dts/omap5.dtsi > --- a/arch/arm/boot/dts/omap5.dtsi > +++ b/arch/arm/boot/dts/omap5.dtsi > @@ -257,6 +257,29 @@ > ports-implemented = <0x1>; > }; > > + target-module@56000000 { > + compatible = "ti,sysc-omap4", "ti,sysc"; > + reg = <0x5600fe00 0x4>, > + <0x5600fe10 0x4>; > + reg-names = "rev", "sysc"; > + ti,sysc-midle = <SYSC_IDLE_FORCE>, > + <SYSC_IDLE_NO>, > + <SYSC_IDLE_SMART>; > + ti,sysc-sidle = <SYSC_IDLE_FORCE>, > + <SYSC_IDLE_NO>, > + <SYSC_IDLE_SMART>; > + clocks = <&gpu_clkctrl OMAP5_GPU_CLKCTRL 0>; > + clock-names = "fck"; > + #address-cells = <1>; > + #size-cells = <1>; > + ranges = <0 0x56000000 0x2000000>; > + > + /* > + * Closed source PowerVR driver, no child device > + * binding or driver in mainline > + */ > + }; > + > dss: dss@58000000 { > compatible = "ti,omap5-dss"; > reg = <0x58000000 0x80>; > diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi > --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi > +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi > @@ -1146,6 +1146,20 @@ > }; > }; > > + gpu_cm: clock-controller@1500 { > + compatible = "ti,omap4-cm"; > + reg = <0x1500 0x100>; > + #address-cells = <1>; > + #size-cells = <1>; > + ranges = <0 0x1500 0x100>; > + > + gpu_clkctrl: clk@20 { > + compatible = "ti,clkctrl"; > + reg = <0x20 0x4>; > + #clock-cells = <2>; > + }; > + }; > + > l3init_cm: l3init_cm@1600 { > compatible = "ti,omap4-cm"; > reg = <0x1600 0x100>; > -- > 2.23.0
* H. Nikolaus Schaller <hns@goldelico.com> [190905 15:04]: > Hi Tony, > I found some time to update my drivers/stating/pvr/1.14.369993 tree to make > use of the new sgx soc integration code and it works on omap5! > > > root@letux:~# modprobe pvrsrvkm_omap_omap5_sgx544_116 > > [ 106.478254] pvrsrvkm_omap_omap5_sgx544_116: module is from the staging directory, the quality is unknown, you have been warned. > > [ 106.526201] [drm] Initialized pvr 1.14.3699939 20110701 for 5600fe00.sgx on minor 1 > > root@letux:~# pvrsrvctl --no-module --start > > [ 261.091978] PVR_K: UM DDK-(3699939) and KM DDK-(3699939) match. [ OK ] > > root@letux:~# gles1test1 0 > > ^C > > root@letux:~# Hey that's great, good to hear :) > To make it work, I have added a child node to the target-module@56000000 > that describes the img,sgx5xx. Here are the work-in-progress patches if > you want to take a look inside: > > https://github.com/openpvrsgx-devgroup/linux_openpvrsgx/commits/letux/omap-pvr-soc-glue-v2 OK > On omap3630 the driver module loads but pvrsrvinit fails: > > > root@letux:~# modprobe pvrsrvkm_omap_omap3630_sgx530_125 > > [ 390.429260] pvrsrvkm_omap_omap3630_sgx530_125: module is from the staging directory, the quality is unknown, you have been warned. > > [ 390.474304] [drm] Initialized pvr 1.14.3699939 20110701 for 50000000.sgx on minor 1 > > root@letux:~# pvrsrvctl --start --no-module > > [ 402.200439] PVR_K: UM DDK-(3699939) and KM DDK-(3699939) match. [ OK ] > > [ 402.219604] PVR_K:(Error): BridgedDispatchKM: Initialisation failed. Driver unusable. > > PVR:(Error): LoaduKernelProgram : SGX ukernel program Device Addr: 0xe400000 invalid alignment [0, ] > > PVR:(Error): SetupuKernel : Failed to load uKernel programs [0, ] > > PVR:(Error): SrvInit: Initialisation for device of class 0, type 7, index 0, failed (1) [0, ] > > PVR:(Error): PVRSRVBridgeCall: Failed to access device. Function ID:3223086862 (strerror returns no value.). [0, ] > > pvrsrvctl: SrvInit failed (already initialized?) (err=PVRSRV_ERROR_OUT_OF_MEMORY - Unable to allocate required memory) > > root@letux:~# > > Maybe I do not have done the address range translation properly. Or I am missing > some of your PRM rstctrl patches since I am working on top of linus/master. > Because of that I had to comment out prm_gfx in am33xx.dtsi to get it compile and could > not test yet on Beagle Bone. My guess is that on 36xx the sgx registers are at a different offset that does not start at offset 0? And this might be different again for 34xx? If you enable the module via sysfs, you should be able to just read out some registers to figure out where it might be is my guess. > A little unclear is how to properly handle omap4. omap4430/4440 and 4460 have the sgx540 > while only omap4470 has an sgx544 inside. But omap4420/30 share one .dtsi and omap4460/70 > the other. So we might have to define a new omap4470.dtsi and #include in the real omap4470 > boards. Oh OK. Well for 4460, see omap4460.dtsi that already has some ranges updated compared to 4430 in l4_cfg_segment_300000 and l4_cfg_target_0. I recall that was the only difference for the added 2d accelerator between 4430 and 4460. Adding a omap4470.dtsi in a similar way should work, just also update also &sgx_module there too. > And am3517 and am43/am57/am65/... support is missing. But that can be added later if we > have it running on omap3 and am33xx. Yeah sure at least 335x depends on the pending rstctrl patches. > Anyways, we again have made a big step forwards. Let's hope that we can get > the FLOSS part of the sgx drivers to staging in some not too far future. Yeah great. Regards, Tony
Hi Tony, > Am 26.08.2019 um 16:55 schrieb Tony Lindgren <tony@atomide.com>: > > * Tony Lindgren <tony@atomide.com> [190817 06:56]: >> * Stephen Boyd <sboyd@kernel.org> [190815 18:24]: >>> Quoting Tony Lindgren (2019-08-14 06:14:06) >>>> diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi >>>> @@ -1146,6 +1146,20 @@ >>>> }; >>>> }; >>>> >>>> + gpu_cm: gpu_cm@1500 { >>> >>> Node names shouldn't have underscores. Maybe clock-controller? >> >> OK yeah clock-controller sounds good to me. > > Below is this one updated to use clock-controller naming. > > Regards, > > Tony > > 8< --------------------------- > From tony Mon Sep 17 00:00:00 2001 > From: Tony Lindgren <tony@atomide.com> > Date: Wed, 14 Aug 2019 05:18:16 -0700 > Subject: [PATCH] ARM: dts: Configure sgx for omap5 > MIME-Version: 1.0 > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: 8bit > > I've tested that the interconnect target module enables and idles > just fine when probed with ti-sysc with PM runtime control via sys: > > # echo on > $(find /sys -name control | grep \/5600) Patch has arrived in v5.4-rc1 but does not work in that environment for me: root@letux:~# echo on > $(find /sys -name control | grep \/5600) [ 213.490926] clock-controller:clk:0000:0: failed to enable root@letux:~# uname -a Linux letux 5.4.0-rc1+ #1083 SMP Sat Oct 5 17:47:13 CEST 2019 armv7l GNU/Linux root@letux:~# find /sys -name control | grep \/5600 /sys/devices/platform/44000000.ocp/5600fe00.target-module/power/control root@letux:~# I have tested our own tree but also mainline (plus some minimal patch set to run the Pyra) using omap2plus_defconfig. Does it work on omap5uevm? Is there some related patch needed that did not yet arrive in v5.4-rc1? BR and thanks, Nikolaus
diff --git a/arch/arm/boot/dts/omap5.dtsi b/arch/arm/boot/dts/omap5.dtsi --- a/arch/arm/boot/dts/omap5.dtsi +++ b/arch/arm/boot/dts/omap5.dtsi @@ -257,6 +257,29 @@ ports-implemented = <0x1>; }; + target-module@56000000 { + compatible = "ti,sysc-omap4", "ti,sysc"; + reg = <0x5600fe00 0x4>, + <0x5600fe10 0x4>; + reg-names = "rev", "sysc"; + ti,sysc-midle = <SYSC_IDLE_FORCE>, + <SYSC_IDLE_NO>, + <SYSC_IDLE_SMART>; + ti,sysc-sidle = <SYSC_IDLE_FORCE>, + <SYSC_IDLE_NO>, + <SYSC_IDLE_SMART>; + clocks = <&gpu_clkctrl OMAP5_GPU_CLKCTRL 0>; + clock-names = "fck"; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0 0x56000000 0x2000000>; + + /* + * Closed source PowerVR driver, no child device + * binding or driver in mainline + */ + }; + dss: dss@58000000 { compatible = "ti,omap5-dss"; reg = <0x58000000 0x80>; diff --git a/arch/arm/boot/dts/omap54xx-clocks.dtsi b/arch/arm/boot/dts/omap54xx-clocks.dtsi --- a/arch/arm/boot/dts/omap54xx-clocks.dtsi +++ b/arch/arm/boot/dts/omap54xx-clocks.dtsi @@ -1146,6 +1146,20 @@ }; }; + gpu_cm: gpu_cm@1500 { + compatible = "ti,omap4-cm"; + reg = <0x1500 0x100>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0 0x1500 0x100>; + + gpu_clkctrl: clk@20 { + compatible = "ti,clkctrl"; + reg = <0x20 0x4>; + #clock-cells = <2>; + }; + }; + l3init_cm: l3init_cm@1600 { compatible = "ti,omap4-cm"; reg = <0x1600 0x100>;
I've tested that the interconnect target module enables and idles just fine when probed with ti-sysc with PM runtime control via sys: # echo on > $(find /sys -name control | grep \/5600) # rwmem 0x5600fe00 # OCP Revision 0x5600fe00 = 0x40000000 # echo auto > $(find /sys -name control | grep \/5600) # rwmem 0x5600fe10 # rwmem 0x56000024 Cc: Adam Ford <aford173@gmail.com> Cc: Filip Matijević <filip.matijevic.pz@gmail.com> Cc: "H. Nikolaus Schaller" <hns@goldelico.com> Cc: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com> Cc: moaz korena <moaz@korena.xyz> Cc: Merlijn Wajer <merlijn@wizzup.org> Cc: Paweł Chmiel <pawel.mikolaj.chmiel@gmail.com> Cc: Philipp Rossak <embed3d@gmail.com> Cc: Tomi Valkeinen <tomi.valkeinen@ti.com> Signed-off-by: Tony Lindgren <tony@atomide.com> --- arch/arm/boot/dts/omap5.dtsi | 23 +++++++++++++++++++++++ arch/arm/boot/dts/omap54xx-clocks.dtsi | 14 ++++++++++++++ 2 files changed, 37 insertions(+)