Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
I sent a patch to the mailing list and wanted to have some review on that from at least Ben, but no idea if Ben already picked it and if it's good enough for sending it to stable yet.
Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
On Saturday 10 October 2020 02:02:42 Karol Herbst wrote: > On Sat, Oct 10, 2020 at 12:23 AM Ilia Mirkin wrote: > > > > On Fri, Oct 9, 2020 at 5:54 PM Karol Herbst wrote: > > > > > > On Fri, Oct 9, 2020 at 11:35 PM Ondrej Zary wrote: > > > > > > > > Hello, > > > > I'm testing 5.9.0-rc8 and found that Riva TNT2 stopped working: > > > > [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian > > > > 8.3.0-6) 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri > > > > Oct 9 22:31:40 CEST 2020 > > > > ... > > > > [ 14.771464] nouveau :01:00.0: GPU not supported on big-endian > > > > [ 14.771782] nouveau: probe of :01:00.0 failed with error -38 > > > > > > > > big-endian? WTF? The machine is x86. > > > > > > > > > > mhh, we reworked the endianess checks a bit and apparently that broke > > > something... I will give it some thoughts, but could you be so kind > > > and create an mmiotrace under 5.9 with nouveau? You won't need to > > > start X or anything while doing it. Just enable the trace and modprobe > > > nouveau and collect the trace. > > > > Looks like nvkm_device_endianness unconditionally reads out 0x4. I > > don't think that reg is there pre-NV11. At least NV4, NV5, NV10 and > > maybe NV15 (which is logically pre-NV11) don't support big-endian > > mode. Not sure about NV1A, which was the IGP of the series and IIRC > > logically pre-NV11 as well (but clearly could only be used with x86 > > chips, since it was part of the motherboard). > > > > Aha, it's documented in rnndb: > > > > https://github.com/envytools/envytools/blob/master/rnndb/bus/pmc.xml > > > > > > ohh, I should have checked there.. yeah, will write a fix for it then. > Before my patch we just always tried to switch it, but never threw an > error. Any progress with the patch? -- Ondrej Zary
Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
On Saturday 10 October 2020 00:23:38 Ilia Mirkin wrote: > On Fri, Oct 9, 2020 at 5:54 PM Karol Herbst wrote: > > > > On Fri, Oct 9, 2020 at 11:35 PM Ondrej Zary wrote: > > > > > > Hello, > > > I'm testing 5.9.0-rc8 and found that Riva TNT2 stopped working: > > > [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian 8.3.0-6) > > > 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri Oct 9 > > > 22:31:40 CEST 2020 > > > ... > > > [ 14.771464] nouveau :01:00.0: GPU not supported on big-endian > > > [ 14.771782] nouveau: probe of :01:00.0 failed with error -38 > > > > > > big-endian? WTF? The machine is x86. > > > > > > > mhh, we reworked the endianess checks a bit and apparently that broke > > something... I will give it some thoughts, but could you be so kind > > and create an mmiotrace under 5.9 with nouveau? You won't need to > > start X or anything while doing it. Just enable the trace and modprobe > > nouveau and collect the trace. > > Looks like nvkm_device_endianness unconditionally reads out 0x4. I > don't think that reg is there pre-NV11. At least NV4, NV5, NV10 and > maybe NV15 (which is logically pre-NV11) don't support big-endian > mode. Not sure about NV1A, which was the IGP of the series and IIRC > logically pre-NV11 as well (but clearly could only be used with x86 > chips, since it was part of the motherboard). Yes, you're right. Forcing nvkm_device_endianness to return true allows 5.9.0-rc8 to work: [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian 8.3.0-6) 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri Oct 9 22:31:40 CEST 2020 ... [ 12.311258] nouveau :01:00.0: bios: DCB table not found [ 12.311583] nouveau :01:00.0: bios: DCB table not found [ 12.311834] nouveau :01:00.0: bios: DCB table not found [ 12.311847] nouveau :01:00.0: bios: DCB table not found [ 12.311989] agpgart-intel :00:00.0: AGP 3.0 bridge [ 12.312017] agpgart-intel :00:00.0: bridge is in legacy mode, falling back to 2.x [ 12.312031] agpgart-intel :00:00.0: putting AGP V2 device into 4x mode [ 12.312066] nouveau :01:00.0: putting AGP V2 device into 4x mode [ 12.312162] agpgart-intel :00:00.0: AGP 3.0 bridge [ 12.312182] agpgart-intel :00:00.0: bridge is in legacy mode, falling back to 2.x [ 12.312195] agpgart-intel :00:00.0: putting AGP V2 device into 4x mode [ 12.312230] nouveau :01:00.0: putting AGP V2 device into 4x mode [ 12.312247] nouveau :01:00.0: tmr: unknown input clock freq [ 12.318341] nouveau :01:00.0: fb: 32 MiB SDRAM [ 12.76] [TTM] Zone kernel: Available graphics memory: 385048 KiB [ 12.92] [TTM] Initializing pool allocator [ 12.333434] nouveau :01:00.0: DRM: VRAM: 31 MiB [ 12.333443] nouveau :01:00.0: DRM: GART: 128 MiB [ 12.333453] nouveau :01:00.0: DRM: BMP version 5.6 [ 12.333460] nouveau :01:00.0: DRM: No DCB data found in VBIOS [ 12.335355] nouveau :01:00.0: DRM: MM: using M2MF for buffer copies [ 12.335443] nouveau :01:00.0: bios: DCB table not found [ 12.336033] nouveau :01:00.0: DRM: Saving VGA fonts [ 12.376420] nouveau :01:00.0: DRM: No DCB data found in VBIOS [ 12.410397] nouveau :01:00.0: DRM: allocated 1280x1024 fb: 0x4000, bo b68d2ac4 [ 12.441217] fbcon: nouveaudrmfb (fb0) is primary device [ 12.591964] Console: switching to colour frame buffer device 160x64 [ 12.593876] nouveau :01:00.0: [drm] fb0: nouveaudrmfb frame buffer device [ 12.594944] [drm] Initialized nouveau 1.3.1 20120801 for :01:00.0 on minor 0 BTW. 5.8 kernel (that appeared today in Debian packports) is broken the same way. > Aha, it's documented in rnndb: > > https://github.com/envytools/envytools/blob/master/rnndb/bus/pmc.xml > > > -ilia > -- Ondrej Zary
Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
On Sat, Oct 10, 2020 at 12:23 AM Ilia Mirkin wrote: > > On Fri, Oct 9, 2020 at 5:54 PM Karol Herbst wrote: > > > > On Fri, Oct 9, 2020 at 11:35 PM Ondrej Zary wrote: > > > > > > Hello, > > > I'm testing 5.9.0-rc8 and found that Riva TNT2 stopped working: > > > [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian 8.3.0-6) > > > 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri Oct 9 > > > 22:31:40 CEST 2020 > > > ... > > > [ 14.771464] nouveau :01:00.0: GPU not supported on big-endian > > > [ 14.771782] nouveau: probe of :01:00.0 failed with error -38 > > > > > > big-endian? WTF? The machine is x86. > > > > > > > mhh, we reworked the endianess checks a bit and apparently that broke > > something... I will give it some thoughts, but could you be so kind > > and create an mmiotrace under 5.9 with nouveau? You won't need to > > start X or anything while doing it. Just enable the trace and modprobe > > nouveau and collect the trace. > > Looks like nvkm_device_endianness unconditionally reads out 0x4. I > don't think that reg is there pre-NV11. At least NV4, NV5, NV10 and > maybe NV15 (which is logically pre-NV11) don't support big-endian > mode. Not sure about NV1A, which was the IGP of the series and IIRC > logically pre-NV11 as well (but clearly could only be used with x86 > chips, since it was part of the motherboard). > > Aha, it's documented in rnndb: > > https://github.com/envytools/envytools/blob/master/rnndb/bus/pmc.xml > > ohh, I should have checked there.. yeah, will write a fix for it then. Before my patch we just always tried to switch it, but never threw an error. > -ilia >
Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
On Fri, Oct 9, 2020 at 5:54 PM Karol Herbst wrote: > > On Fri, Oct 9, 2020 at 11:35 PM Ondrej Zary wrote: > > > > Hello, > > I'm testing 5.9.0-rc8 and found that Riva TNT2 stopped working: > > [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian 8.3.0-6) > > 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri Oct 9 22:31:40 > > CEST 2020 > > ... > > [ 14.771464] nouveau :01:00.0: GPU not supported on big-endian > > [ 14.771782] nouveau: probe of :01:00.0 failed with error -38 > > > > big-endian? WTF? The machine is x86. > > > > mhh, we reworked the endianess checks a bit and apparently that broke > something... I will give it some thoughts, but could you be so kind > and create an mmiotrace under 5.9 with nouveau? You won't need to > start X or anything while doing it. Just enable the trace and modprobe > nouveau and collect the trace. Looks like nvkm_device_endianness unconditionally reads out 0x4. I don't think that reg is there pre-NV11. At least NV4, NV5, NV10 and maybe NV15 (which is logically pre-NV11) don't support big-endian mode. Not sure about NV1A, which was the IGP of the series and IIRC logically pre-NV11 as well (but clearly could only be used with x86 chips, since it was part of the motherboard). Aha, it's documented in rnndb: https://github.com/envytools/envytools/blob/master/rnndb/bus/pmc.xml -ilia
Re: [Nouveau] nouveau broken on Riva TNT2 in 5.9.0-rc8: GPU not supported on big-endian
On Fri, Oct 9, 2020 at 11:35 PM Ondrej Zary wrote: > > Hello, > I'm testing 5.9.0-rc8 and found that Riva TNT2 stopped working: > [0.00] Linux version 5.9.0-rc8+ (zary@gsql) (gcc (Debian 8.3.0-6) > 8.3.0, GNU ld (GNU Binutils for Debian) 2.31.1) #326 SMP Fri Oct 9 22:31:40 > CEST 2020 > ... > [ 14.771464] nouveau :01:00.0: GPU not supported on big-endian > [ 14.771782] nouveau: probe of :01:00.0 failed with error -38 > > big-endian? WTF? The machine is x86. > mhh, we reworked the endianess checks a bit and apparently that broke something... I will give it some thoughts, but could you be so kind and create an mmiotrace under 5.9 with nouveau? You won't need to start X or anything while doing it. Just enable the trace and modprobe nouveau and collect the trace. > It works fine with Debian 5.7 kernel (5.7.10-1~bpo10+1): > [0.00] Linux version 5.7.0-0.bpo.2-686 > (debian-ker...@lists.debian.org) (gcc version 8.3.0 (Debian 8.3.0-6), GNU ld > (GNU Binutils for Debian) 2.31.1) #1 SMP Debian 5.7.10-1~bpo10+1 (2020-07-30) > ... > [ 23.266196] nouveau :01:00.0: NVIDIA NV05 (20154000) > [ 23.288582] nouveau :01:00.0: bios: version 02.05.20.02.00 > [ 23.288869] nouveau :01:00.0: bios: DCB table not found > [ 23.289595] nouveau :01:00.0: bios: DCB table not found > [ 23.289956] nouveau :01:00.0: bios: DCB table not found > [ 23.290015] nouveau :01:00.0: bios: DCB table not found > [ 23.290215] agpgart-intel :00:00.0: AGP 3.0 bridge > [ 23.290287] agpgart-intel :00:00.0: bridge is in legacy mode, falling > back to 2.x > [ 23.290351] agpgart-intel :00:00.0: putting AGP V2 device into 4x mode > [ 23.290430] nouveau :01:00.0: putting AGP V2 device into 4x mode > [ 23.290565] agpgart-intel :00:00.0: AGP 3.0 bridge > [ 23.290627] agpgart-intel :00:00.0: bridge is in legacy mode, falling > back to 2.x > [ 23.290690] agpgart-intel :00:00.0: putting AGP V2 device into 4x mode > [ 23.290768] nouveau :01:00.0: putting AGP V2 device into 4x mode > [ 23.290830] nouveau :01:00.0: tmr: unknown input clock freq > [ 23.293026] nouveau :01:00.0: fb: 32 MiB SDRAM > [ 23.301269] [TTM] Zone kernel: Available graphics memory: 382728 KiB > [ 23.301327] [TTM] Initializing pool allocator > [ 23.301414] nouveau :01:00.0: DRM: VRAM: 31 MiB > [ 23.301465] nouveau :01:00.0: DRM: GART: 128 MiB > [ 23.301518] nouveau :01:00.0: DRM: BMP version 5.6 > [ 23.301570] nouveau :01:00.0: DRM: No DCB data found in VBIOS > [ 23.303594] nouveau :01:00.0: DRM: MM: using M2MF for buffer copies > [ 23.303719] nouveau :01:00.0: bios: DCB table not found > [ 23.304904] nouveau :01:00.0: DRM: Saving VGA fonts > [ 23.349089] nouveau :01:00.0: DRM: No DCB data found in VBIOS > [ 23.349681] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). > [ 23.383066] nouveau :01:00.0: DRM: allocated 1280x1024 fb: 0x4000, bo > b10d2f17 > [ 23.413903] fbcon: nouveaudrmfb (fb0) is primary device > [ 23.569851] Console: switching to colour frame buffer device 160x64 > [ 23.571050] nouveau :01:00.0: fb0: nouveaudrmfb frame buffer device > > > -- > Ondrej Zary > ___ > Nouveau mailing list > nouv...@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/nouveau >
