linux-kernel - Re: Regression 2.6.34->2.6.35-rc4: radeaon KMS an RS690 broken

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <AANLkTilWzpu6Re9HG2jTqpHUU7NohULOEgHqGT4yTIaz@mail.gmail.com>
Date:	Wed, 14 Jul 2010 18:16:59 -0400
From:	Alex Deucher <alexdeucher@...il.com>
To:	Torsten Kaiser <just.for.lkml@...glemail.com>
Cc:	Jerome Glisse <glisse@...edesktop.org>,
	linux-kernel@...r.kernel.org,
	DRI mailing list <dri-devel@...ts.freedesktop.org>
Subject: Re: Regression 2.6.34->2.6.35-rc4: radeaon KMS an RS690 broken

On Wed, Jul 14, 2010 at 4:05 PM, Torsten Kaiser
<just.for.lkml@...glemail.com> wrote:
> On Wed, Jul 14, 2010 at 9:30 PM, Jerome Glisse <glisse@...edesktop.org> wrote:
>> On 07/14/2010 02:51 PM, Torsten Kaiser wrote:
>>>
>>> On Tue, Jul 13, 2010 at 9:10 PM, Alex Deucher<alexdeucher@...il.com>
>>>  wrote:
>>>>
>>>> On Tue, Jul 13, 2010 at 2:29 PM, Torsten Kaiser
>>>> <just.for.lkml@...glemail.com>  wrote:
>>>>>
>>>>> But the CP is still broken:
>>>>
>>>> Is this a regression?  If so, can you bisect it?
>>>>
>>>> Alex
>>>
>>> I bisected it to this commit:
>>>
>>> d594e46ace22afa1621254f6f669e65430048153 is the first bad commit
>>> commit d594e46ace22afa1621254f6f669e65430048153
>>> Author: Jerome Glisse<jglisse@...hat.com>
>>> Date:   Wed Feb 17 21:54:29 2010 +0000
>>>
>>>     drm/radeon/kms: simplify memory controller setup V2
>>>
>>>     Get rid of _location and use _start/_end also simplify the
>>>     computation of vram_start|end&  gtt_start|end. For R1XX-R2XX
>>>     we place VRAM at the same address of PCI aperture, those GPU
>>>     shouldn't have much memory and seems to behave better when
>>>     setup that way. For R3XX and newer we place VRAM at 0. For
>>>     R6XX-R7XX AGP we place VRAM before or after AGP aperture this
>>>     might limit to limit the VRAM size but it's very unlikely.
>>>     For IGP we don't change the VRAM placement.
>>>
>>>     Tested on (compiz,quake3,suspend/resume):
>>>     PCI/PCIE:RV280,R420,RV515,RV570,RV610,RV710
>>>     AGP:RV100,RV280,R420,RV350,RV620(RPB*),RV730
>>>     IGP:RS480(RPB*),RS690,RS780(RPB*),RS880
>>>
>>>     RPB: resume previously broken
>>>
>>>     V2 correct commit message to reflect more accurately the bug
>>>     and move VRAM placement to 0 for most of the GPU to avoid
>>>     limiting VRAM.
>>>
>>>     Signed-off-by: Jerome Glisse<jglisse@...hat.com>
>>>     Signed-off-by: Dave Airlie<airlied@...hat.com>
>>>
>>> :040000 040000 05c1e456fcf6565aa8711e4933807956d0055cca
>>> 792c6be2bd161a52500c5e8d685ee651cd5af07e M     drivers
>>>
>>> HTH, Torsten
>>>
>>>>> [    0.426931] Linux agpgart interface v0.103
>>>>> [    0.427092] [drm] Initialized drm 1.1.0 20060810
>>>>> [    0.427196] [drm] radeon defaulting to kernel modesetting.
>>>>> [    0.427255] [drm] radeon kernel modesetting enabled.
>>>>> [    0.427372] radeon 0000:01:05.0: PCI INT A ->  GSI 18 (level, low) ->
>>>>>  IRQ 18
>>>>> [    0.429659] [drm] initializing kernel modesetting (RS690
>>>>> 0x1002:0x791E).
>>>>> [    0.429817] [drm] register mmio base: 0xFE9F0000
>>>>> [    0.429876] [drm] register mmio size: 65536
>>>>> [    0.430457] ATOM BIOS: ATI
>>>>> [    0.430532] radeon 0000:01:05.0: VRAM: 32M 0xDE000000 - 0xDFFFFFFF
>>>>> (32M used)
>>>>> [    0.430592] radeon 0000:01:05.0: GTT: 512M 0xBE000000 - 0xDDFFFFFF
>>>>> [    0.430675] [drm] radeon: irq initialized.
>>>>> [    0.430737] mtrr: type mismatch for fc000000,2000000 old:
>>>>> write-back new: write-comb
>>>>> ining
>>>>> [    0.430811] [drm] Detected VRAM RAM=32M, BAR=32M
>>>>> [    0.430868] [drm] RAM width 128bits DDR
>>>>> [    0.431011] [TTM] Zone  kernel: Available graphics memory: 2010234
>>>>> kiB.
>>>>> [    0.431070] [TTM] Initializing pool allocator.
>>>>> [    0.431147] [drm] radeon: 32M of VRAM memory ready
>>>>> [    0.431205] [drm] radeon: 512M of GTT memory ready.
>>>>> [    0.431266] [drm] GART: num cpu pages 131072, num gpu pages 131072
>>>>> [    0.434654] [drm] radeon: 1 quad pipes, 1 z pipes initialized.
>>>>> [    0.441719] [drm] Loading RS690/RS740 Microcode
>>>>> [    0.441926] [drm] radeon: ring at 0x00000000BE000000
>>>>> [    0.577118] [drm:r100_ring_test] *ERROR* radeon: ring test failed
>>>>> (sracth(0x15E4)=0x
>>>>> CAFEDEAD)
>>>>> [    0.577192] [drm:r100_cp_init] *ERROR* radeon: cp isn't working
>>>>> (-22).
>>>>> [    0.577252] radeon 0000:01:05.0: failled initializing CP (-22).
>>>>> [    0.577310] radeon 0000:01:05.0: Disabling GPU acceleration
>>>>> [    0.577440] [drm] radeon: cp finalized
>>>>> [    0.578078] [drm] Default TV standard: NTSC
>>>>> [    0.578314] [drm] Default TV standard: NTSC
>>>>> [    0.578590] [drm] Radeon Display Connectors
>>>>> [    0.578648] [drm] Connector 0:
>>>>> [    0.578706] [drm]   VGA
>>>>> [    0.578764] [drm]   DDC: 0x7e50 0x7e40 0x7e54 0x7e44 0x7e58 0x7e48
>>>>> 0x7e5c 0x7e4c
>>>>> [    0.578837] [drm]   Encoders:
>>>>> [    0.578894] [drm]     CRT1: INTERNAL_KLDSCP_DAC1
>>>>> [    0.578952] [drm] Connector 1:
>>>>> [    0.579010] [drm]   S-video
>>>>> [    0.579067] [drm]   Encoders:
>>>>> [    0.579124] [drm]     TV1: INTERNAL_KLDSCP_DAC1
>>>>> [    0.579182] [drm] Connector 2:
>>>>> [    0.579239] [drm]   HDMI-A
>>>>> [    0.579297] [drm]   DDC: 0x7e40 0x7e50 0x7e44 0x7e54 0x7e48 0x7e58
>>>>> 0x7e4c 0x7e5c
>>>>> [    0.579369] [drm]   Encoders:
>>>>> [    0.579427] [drm]     DFP3: INTERNAL_LVTM1
>>>>> [    0.773375] [drm] fb mappable at 0xFC040000
>>>>> [    0.773434] [drm] vram apper at 0xFC000000
>>>>> [    0.773491] [drm] size 786432
>>>>> [    0.773549] [drm] fb depth is 8
>>>>> [    0.773606] [drm]    pitch is 1024
>>>>> [    0.773737] fbcon: radeondrmfb (fb0) is primary device
>>>>> [    0.793240] Console: switching to colour frame buffer device 128x48
>>>>> [    0.794833] fb0: radeondrmfb frame buffer device
>>>>> [    0.794852] drm: registered panic notifier
>>>>> [    0.794871] Slow work thread pool: Starting up
>>>>> [    0.794932] Slow work thread pool: Ready
>>>>> [    0.794953] [drm] Initialized radeon 2.5.0 20080528 for
>>>>> 0000:01:05.0 on minor 0
>>>>>
>>>>>
>>>>> Torsten
>>>>>
>>>>
>>
>> Does the attached patch works ? (try to change the if 0 to if 1 too
>
> The patch doesn't compile, but after changing mc->foo to rdev->mc.foo it built.
>

I discussed this with Jerome and I think we found the root cause.
Does this patch help?

Alex


> Result from the original patch ("if 0"):
> [    0.429603] [drm] initializing kernel modesetting (RS690 0x1002:0x791E).
> [    0.429751] [drm] register mmio base: 0xFE9F0000
> [    0.429809] [drm] register mmio size: 65536
> [    0.430385] ATOM BIOS: ATI
> [    0.430460] radeon 0000:01:05.0: VRAM: 32M 0xDE000000 - 0xDFFFFFFF (32M used)
> [    0.430520] radeon 0000:01:05.0: GTT: 512M 0xB0000000 - 0xCFFFFFFF
> [    0.430603] [drm] radeon: irq initialized.
> [    0.430666] mtrr: type mismatch for fc000000,2000000 old:
> write-back new: write-combining
> [    0.430739] [drm] Detected VRAM RAM=32M, BAR=32M
> [    0.430797] [drm] RAM width 128bits DDR
> [    0.430940] [TTM] Zone  kernel: Available graphics memory: 2010234 kiB.
> [    0.430999] [TTM] Initializing pool allocator.
> [    0.431075] [drm] radeon: 32M of VRAM memory ready
> [    0.431133] [drm] radeon: 512M of GTT memory ready.
> [    0.431194] [drm] GART: num cpu pages 131072, num gpu pages 131072
> [    0.434577] [drm] radeon: 1 quad pipes, 1 z pipes initialized.
> [    0.441645] [drm] Loading RS690/RS740 Microcode
> [    0.441853] [drm] radeon: ring at 0x00000000B0000000
> [    0.576773] [drm:r100_ring_test] *ERROR* radeon: ring test failed
> (sracth(0x15E4)=0xCAFEDEAD)
> [    0.576847] [drm:r100_cp_init] *ERROR* radeon: cp isn't working (-22).
> [    0.576907] radeon 0000:01:05.0: failled initializing CP (-22).
> [    0.576965] radeon 0000:01:05.0: Disabling GPU acceleration
>
> Result from patch after changing it to "if 1":
> [    0.400348] [drm] initializing kernel modesetting (RS690 0x1002:0x791E).
> [    0.400497] [drm] register mmio base: 0xFE9F0000
> [    0.400556] [drm] register mmio size: 65536
> [    0.401097] ATOM BIOS: ATI
> [    0.401171] radeon 0000:01:05.0: VRAM: 32M 0xDE000000 - 0xDFFFFFFF (32M used)
> [    0.401231] radeon 0000:01:05.0: GTT: 512M 0x00000000 - 0x1FFFFFFF
> [    0.401314] [drm] radeon: irq initialized.
> [    0.401377] mtrr: type mismatch for fc000000,2000000 old:
> write-back new: write-combining
> [    0.401451] [drm] Detected VRAM RAM=32M, BAR=32M
> [    0.401509] [drm] RAM width 128bits DDR
> [    0.401597] [TTM] Zone  kernel: Available graphics memory: 2010234 kiB.
> [    0.401656] [TTM] Initializing pool allocator.
> [    0.401732] [drm] radeon: 32M of VRAM memory ready
> [    0.401791] [drm] radeon: 512M of GTT memory ready.
> [    0.401852] [drm] GART: num cpu pages 131072, num gpu pages 131072
> [    0.405242] [drm] radeon: 1 quad pipes, 1 z pipes initialized.
> [    0.412298] [drm] Loading RS690/RS740 Microcode
> [    0.412507] [drm] radeon: ring at 0x0000000000000000
> [    0.412582] [drm] ring test succeeded in 1 usecs
> [    0.412726] [drm] radeon: ib pool ready.
> [    0.412792] [drm] ib test succeeded in 0 usecs
>
> Torsten
>

View attachment "0001-drm-radeon-kms-fix-gtt-MC-base-alignment-on-rs4xx-rs.patch" of type "text/x-patch" (7693 bytes)