Hi,
I tried booting v5.4 mainline kernel on a stratix10 board with ECC
disabled, and the altera-edac driver (with only SDRAM enabled) is
now crashing the system instead of failing the probe with "No ECC/ECC
disabled".
This seems to have started with commit 08f08bfb7b4c ("EDAC, altera:
Merge Stratix10 into the Arria10 SDRAM probe routine"). With the change,
looks like sdram probe no longer uses SMC calls and instead accesses
the registers directly. The crash looks like this:
[ 4.591371] Internal error: synchronous external abort: 96000010 [#1] PREEMPT SMP
[ 4.598816] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.0.0-aakoskin+ #8
[ 4.605483] Hardware name: SoCFPGA Stratix 10 test (DT)
[ 4.610683] pstate: 20000085 (nzCv daIf -PAN -UAO)
[ 4.615461] pc : regmap_mmio_read32le+0x24/0x48
[ 4.619969] lr : regmap_mmio_read+0x48/0x70
[ 4.624130] sp : ffff00001002b570
[ 4.627428] x29: ffff00001002b570 x28: ffff0000107b558c
[ 4.632715] x27: 0000000000000000 x26: 0000000000000000
[ 4.638002] x25: ffff0000107dc000 x24: ffff800014690c00
[ 4.643288] x23: ffff00001002b6cc x22: ffff800014690c10
[ 4.648576] x21: ffff00001002b6cc x20: ffff800014672800
[ 4.653862] x19: 0000000000000000 x18: ffffffffffffffff
[ 4.659148] x17: 0000000000000007 x16: 0000000000000001
[ 4.664434] x15: ffff000010716000 x14: ffffffffff000000
[ 4.669721] x13: 0000000000001000 x12: 0000000000000038
[ 4.675007] x11: 0000000000000010 x10: 0101010101010101
[ 4.680293] x9 : 0000000000000008 x8 : 0000000000000008
[ 4.685579] x7 : 0000000000000000 x6 : ffff0000107dc548
[ 4.690865] x5 : 0000000000000003 x4 : ffff000010091d98
[ 4.696151] x3 : ffff000010412948 x2 : ffff000010412790
[ 4.701437] x1 : 0000000000000000 x0 : ffff000010045100
[ 4.706725] Process swapper/0 (pid: 1, stack limit = 0x(____ptrval____))
[ 4.713393] Call trace:
[ 4.715828] regmap_mmio_read32le+0x24/0x48
[ 4.719991] regmap_mmio_read+0x48/0x70
[ 4.723811] _regmap_bus_reg_read+0x38/0x48
[ 4.727974] _regmap_read+0x6c/0x1a8
[ 4.731533] regmap_read+0x50/0x78
[ 4.734921] altr_sdram_probe+0x98/0x5f0
[ 4.738828] platform_drv_probe+0x58/0xb8
[ 4.742819] really_probe+0x1d4/0x3d8
[ 4.746463] driver_probe_device+0x70/0x158
[ 4.750627] __device_attach_driver+0xbc/0x140
[ 4.755049] bus_for_each_drv+0x80/0xe0
[ 4.758866] __device_attach+0xe8/0x170
[ 4.762684] device_initial_probe+0x24/0x30
[ 4.766847] bus_probe_device+0xa0/0xa8
[ 4.770666] device_add+0x3d4/0x5f0
[ 4.774140] of_device_add+0x5c/0x70
[ 4.777698] of_platform_device_create_pdata+0xc4/0x128
[ 4.782898] of_platform_bus_create+0x1d0/0x4e0
[ 4.787405] of_platform_populate+0x84/0x150
[ 4.791655] altr_edac_a10_probe+0x388/0x428
[ 4.795904] platform_drv_probe+0x58/0xb8
[ 4.799895] really_probe+0x1d4/0x3d8
[ 4.803539] driver_probe_device+0x70/0x158
[ 4.807702] __driver_attach+0x12c/0x148
[ 4.811606] bus_for_each_dev+0x78/0xe0
[ 4.815422] driver_attach+0x30/0x40
[ 4.818980] bus_add_driver+0x1c0/0x288
[ 4.822797] driver_register+0x68/0x118
[ 4.826614] __platform_driver_register+0x54/0x60
[ 4.831299] altr_edac_a10_driver_init+0x24/0x2c
[ 4.835896] do_one_initcall+0xcc/0x230
[ 4.839714] kernel_init_freeable+0x290/0x384
[ 4.844051] kernel_init+0x18/0x110
[ 4.847525] ret_from_fork+0x10/0x1c
[ 4.851085] Code: 2a0103f3 d503201f f9400280 8b334000 (b9400000)
[ 4.857154] ---[ end trace 416494f6b96edb9b ]---
After the commit 3123c5c4ca15 ("edac: altera: Move Stratix10 SDRAM ECC
to peripheral") the crash location is now at altr_check_ecc_deps(),
but I think the cause is still the same:
[ 4.996439] Internal error: synchronous external abort: 96000010 [#1] PREEMPT SMP
[ 5.003892] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.4.0-aakoskin+ #10
[ 5.010645] Hardware name: SoCFPGA Stratix 10 test (DT)
[ 5.015846] pstate: 80000005 (Nzcv daif -PAN -UAO)
[ 5.020623] pc : altr_check_ecc_deps+0x28/0x70
[ 5.025046] lr : altr_edac_a10_device_add+0x19c/0x388
[ 5.030071] sp : ffff80001002b9f0
[ 5.033368] x29: ffff80001002b9f0 x28: ffff80001002ba88
[ 5.038657] x27: ffff8000104cd000 x26: ffff8000106a82c0
[ 5.043943] x25: ffff800010848000 x24: ffff00007fc007a0
[ 5.049230] x23: ffff00007fc00570 x22: ffff0000146e3000
[ 5.054516] x21: ffff000014728880 x20: 0000000010035100
[ 5.059803] x19: ffff0000146e32a8 x18: ffff800010035fff
[ 5.065089] x17: 000000000000000e x16: 0000000000000007
[ 5.070375] x15: ffff800010035fff x14: 00000000f8012000
[ 5.075661] x13: 0000000000000000 x12: ffff80001077c000
[ 5.080947] x11: 0040000000000001 x10: 0040000000000041
[ 5.086233] x9 : 0400000000000001 x8 : ffff800010036000
[ 5.091520] x7 : 0000000000000000 x6 : 0000000000000001
[ 5.096807] x5 : 0000000000000001 x4 : ffff0000146f4000
[ 5.102094] x3 : ffff0000146f4100 x2 : 0000000000000000
[ 5.107380] x1 : ffff8000106a82c0 x0 : ffff800010035100
[ 5.112667] Call trace:
[ 5.115101] altr_check_ecc_deps+0x28/0x70
[ 5.119178] altr_edac_a10_device_add+0x19c/0x388
[ 5.123859] altr_edac_a10_probe+0x1c0/0x3d0
[ 5.128111] platform_drv_probe+0x58/0xb8
[ 5.132102] really_probe+0x1d4/0x430
[ 5.135746] driver_probe_device+0x70/0x158
[ 5.139908] device_driver_attach+0x74/0x80
[ 5.144071] __driver_attach+0xa0/0x198
[ 5.147889] bus_for_each_dev+0x78/0xe0
[ 5.151706] driver_attach+0x30/0x40
[ 5.155263] bus_add_driver+0x178/0x228
[ 5.159080] driver_register+0x68/0x118
[ 5.162898] __platform_driver_register+0x54/0x60
[ 5.167583] altr_edac_a10_driver_init+0x24/0x2c
[ 5.172181] do_one_initcall+0xd4/0x240
[ 5.175997] kernel_init_freeable+0x1c0/0x260
[ 5.180336] kernel_init+0x18/0x110
[ 5.183811] ret_from_fork+0x10/0x18
[ 5.187373] Code: f9401261 f9400a60 b9802c22 8b020000 (b9400000)
[ 5.193440] ---[ end trace cdc4aebcaf1288ee ]---
Was this register access change intentional? How the driver probe should
work when the ECC is not present?
A.
Hi Aaro,
On 11/29/19 10:57 AM, Aaro Koskinen wrote:
> Hi,
>
> I tried booting v5.4 mainline kernel on a stratix10 board with ECC
> disabled, and the altera-edac driver (with only SDRAM enabled) is
> now crashing the system instead of failing the probe with "No ECC/ECC
> disabled".
>
I apologize for the late reply. I was on vacation.
ECC disabled means the sof/jic that you're loading has ECC disabled,
correct?
> This seems to have started with commit 08f08bfb7b4c ("EDAC, altera:
> Merge Stratix10 into the Arria10 SDRAM probe routine"). With the change,
> looks like sdram probe no longer uses SMC calls and instead accesses
> the registers directly. The crash looks like this:
I haven't seen this. I'd expect both ECC enabled and disabled to fail
with the dumps you have below since they'd both need to use the regmap
functions.
Yes, this does look like it is using the register accesses instead of
the SMC call. Line 2206 sets the SMC call after determining from the
if() statement if it is a Stratix10 or Arria10 and from below it seems
to take the Arria10 path.
>
> [ 4.591371] Internal error: synchronous external abort: 96000010 [#1] PREEMPT SMP
> [ 4.598816] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.0.0-aakoskin+ #8
> [ 4.605483] Hardware name: SoCFPGA Stratix 10 test (DT)
> [ 4.610683] pstate: 20000085 (nzCv daIf -PAN -UAO)
> [ 4.615461] pc : regmap_mmio_read32le+0x24/0x48
> [ 4.619969] lr : regmap_mmio_read+0x48/0x70
> [ 4.624130] sp : ffff00001002b570
> [ 4.627428] x29: ffff00001002b570 x28: ffff0000107b558c
> [ 4.632715] x27: 0000000000000000 x26: 0000000000000000
> [ 4.638002] x25: ffff0000107dc000 x24: ffff800014690c00
> [ 4.643288] x23: ffff00001002b6cc x22: ffff800014690c10
> [ 4.648576] x21: ffff00001002b6cc x20: ffff800014672800
> [ 4.653862] x19: 0000000000000000 x18: ffffffffffffffff
> [ 4.659148] x17: 0000000000000007 x16: 0000000000000001
> [ 4.664434] x15: ffff000010716000 x14: ffffffffff000000
> [ 4.669721] x13: 0000000000001000 x12: 0000000000000038
> [ 4.675007] x11: 0000000000000010 x10: 0101010101010101
> [ 4.680293] x9 : 0000000000000008 x8 : 0000000000000008
> [ 4.685579] x7 : 0000000000000000 x6 : ffff0000107dc548
> [ 4.690865] x5 : 0000000000000003 x4 : ffff000010091d98
> [ 4.696151] x3 : ffff000010412948 x2 : ffff000010412790
> [ 4.701437] x1 : 0000000000000000 x0 : ffff000010045100
> [ 4.706725] Process swapper/0 (pid: 1, stack limit = 0x(____ptrval____))
> [ 4.713393] Call trace:
> [ 4.715828] regmap_mmio_read32le+0x24/0x48
> [ 4.719991] regmap_mmio_read+0x48/0x70
> [ 4.723811] _regmap_bus_reg_read+0x38/0x48
> [ 4.727974] _regmap_read+0x6c/0x1a8
> [ 4.731533] regmap_read+0x50/0x78
> [ 4.734921] altr_sdram_probe+0x98/0x5f0
> [ 4.738828] platform_drv_probe+0x58/0xb8
> [ 4.742819] really_probe+0x1d4/0x3d8
> [ 4.746463] driver_probe_device+0x70/0x158
> [ 4.750627] __device_attach_driver+0xbc/0x140
> [ 4.755049] bus_for_each_drv+0x80/0xe0
> [ 4.758866] __device_attach+0xe8/0x170
> [ 4.762684] device_initial_probe+0x24/0x30
> [ 4.766847] bus_probe_device+0xa0/0xa8
> [ 4.770666] device_add+0x3d4/0x5f0
> [ 4.774140] of_device_add+0x5c/0x70
> [ 4.777698] of_platform_device_create_pdata+0xc4/0x128
> [ 4.782898] of_platform_bus_create+0x1d0/0x4e0
> [ 4.787405] of_platform_populate+0x84/0x150
> [ 4.791655] altr_edac_a10_probe+0x388/0x428
> [ 4.795904] platform_drv_probe+0x58/0xb8
> [ 4.799895] really_probe+0x1d4/0x3d8
> [ 4.803539] driver_probe_device+0x70/0x158
> [ 4.807702] __driver_attach+0x12c/0x148
> [ 4.811606] bus_for_each_dev+0x78/0xe0
> [ 4.815422] driver_attach+0x30/0x40
> [ 4.818980] bus_add_driver+0x1c0/0x288
> [ 4.822797] driver_register+0x68/0x118
> [ 4.826614] __platform_driver_register+0x54/0x60
> [ 4.831299] altr_edac_a10_driver_init+0x24/0x2c
> [ 4.835896] do_one_initcall+0xcc/0x230
> [ 4.839714] kernel_init_freeable+0x290/0x384
> [ 4.844051] kernel_init+0x18/0x110
> [ 4.847525] ret_from_fork+0x10/0x1c
> [ 4.851085] Code: 2a0103f3 d503201f f9400280 8b334000 (b9400000)
> [ 4.857154] ---[ end trace 416494f6b96edb9b ]---
>
> After the commit 3123c5c4ca15 ("edac: altera: Move Stratix10 SDRAM ECC
> to peripheral") the crash location is now at altr_check_ecc_deps(),
> but I think the cause is still the same:
>
> [ 4.996439] Internal error: synchronous external abort: 96000010 [#1] PREEMPT SMP
> [ 5.003892] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.4.0-aakoskin+ #10
> [ 5.010645] Hardware name: SoCFPGA Stratix 10 test (DT)
> [ 5.015846] pstate: 80000005 (Nzcv daif -PAN -UAO)
> [ 5.020623] pc : altr_check_ecc_deps+0x28/0x70
> [ 5.025046] lr : altr_edac_a10_device_add+0x19c/0x388
> [ 5.030071] sp : ffff80001002b9f0
> [ 5.033368] x29: ffff80001002b9f0 x28: ffff80001002ba88
> [ 5.038657] x27: ffff8000104cd000 x26: ffff8000106a82c0
> [ 5.043943] x25: ffff800010848000 x24: ffff00007fc007a0
> [ 5.049230] x23: ffff00007fc00570 x22: ffff0000146e3000
> [ 5.054516] x21: ffff000014728880 x20: 0000000010035100
> [ 5.059803] x19: ffff0000146e32a8 x18: ffff800010035fff
> [ 5.065089] x17: 000000000000000e x16: 0000000000000007
> [ 5.070375] x15: ffff800010035fff x14: 00000000f8012000
> [ 5.075661] x13: 0000000000000000 x12: ffff80001077c000
> [ 5.080947] x11: 0040000000000001 x10: 0040000000000041
> [ 5.086233] x9 : 0400000000000001 x8 : ffff800010036000
> [ 5.091520] x7 : 0000000000000000 x6 : 0000000000000001
> [ 5.096807] x5 : 0000000000000001 x4 : ffff0000146f4000
> [ 5.102094] x3 : ffff0000146f4100 x2 : 0000000000000000
> [ 5.107380] x1 : ffff8000106a82c0 x0 : ffff800010035100
> [ 5.112667] Call trace:
> [ 5.115101] altr_check_ecc_deps+0x28/0x70
> [ 5.119178] altr_edac_a10_device_add+0x19c/0x388
> [ 5.123859] altr_edac_a10_probe+0x1c0/0x3d0
> [ 5.128111] platform_drv_probe+0x58/0xb8
> [ 5.132102] really_probe+0x1d4/0x430
> [ 5.135746] driver_probe_device+0x70/0x158
> [ 5.139908] device_driver_attach+0x74/0x80
> [ 5.144071] __driver_attach+0xa0/0x198
> [ 5.147889] bus_for_each_dev+0x78/0xe0
> [ 5.151706] driver_attach+0x30/0x40
> [ 5.155263] bus_add_driver+0x178/0x228
> [ 5.159080] driver_register+0x68/0x118
> [ 5.162898] __platform_driver_register+0x54/0x60
> [ 5.167583] altr_edac_a10_driver_init+0x24/0x2c
> [ 5.172181] do_one_initcall+0xd4/0x240
> [ 5.175997] kernel_init_freeable+0x1c0/0x260
> [ 5.180336] kernel_init+0x18/0x110
> [ 5.183811] ret_from_fork+0x10/0x18
> [ 5.187373] Code: f9401261 f9400a60 b9802c22 8b020000 (b9400000)
> [ 5.193440] ---[ end trace cdc4aebcaf1288ee ]---
>
> Was this register access change intentional? How the driver probe should
> work when the ECC is not present?
>
The altr_check_ecc_deps() call is checking whether ECC is enabled so the
probe should fail.
I suspect the device tree. Can you verify the following node is in your
device tree?
sdramedac {
compatible = "altr,sdram-edac-s10";
altr,sdr-syscon = <&sdr>;
interrupts = <16 4>;
};
I'll keep looking into this as well.
Thor
> A.
>
Hi,
On Tue, Dec 03, 2019 at 05:09:48PM -0600, Thor Thayer wrote:
> On 11/29/19 10:57 AM, Aaro Koskinen wrote:
> >Hi,
> >
> >I tried booting v5.4 mainline kernel on a stratix10 board with ECC
> >disabled, and the altera-edac driver (with only SDRAM enabled) is
> >now crashing the system instead of failing the probe with "No ECC/ECC
> >disabled".
> >
> I apologize for the late reply. I was on vacation.
>
> ECC disabled means the sof/jic that you're loading has ECC disabled,
> correct?
Yes.
> >This seems to have started with commit 08f08bfb7b4c ("EDAC, altera:
> >Merge Stratix10 into the Arria10 SDRAM probe routine"). With the change,
> >looks like sdram probe no longer uses SMC calls and instead accesses
> >the registers directly. The crash looks like this:
>
> I haven't seen this. I'd expect both ECC enabled and disabled to fail with
> the dumps you have below since they'd both need to use the regmap functions.
With ECC enabled it doesn't fail, as the direct register access appears
to work then (I also checked by reading 0xf8011101 from userspace -
and it works without an abort).
> Yes, this does look like it is using the register accesses instead of the
> SMC call. Line 2206 sets the SMC call after determining from the if()
> statement if it is a Stratix10 or Arria10 and from below it seems to take
> the Arria10 path.
But that's setting the ecc_mgr_map. I think that altr_check_ecc_deps
and altr_sdram_probe use a different mapping. Before commit 08f08bfb7b4c
there was S10 specific altr_s10_sdram_probe() that took care of the SMC,
but I cannot see how the current code doing that unless I'm missing some
special magic.
> The altr_check_ecc_deps() call is checking whether ECC is enabled so the
> probe should fail.
>
> I suspect the device tree. Can you verify the following node is in your
> device tree?
>
> sdramedac {
> compatible = "altr,sdram-edac-s10";
> altr,sdr-syscon = <&sdr>;
> interrupts = <16 4>;
> };
Yes, I'm using the in-tree socfpga_stratix10.dtsi.
A.
Hi Aaro,
On 12/4/19 7:25 AM, Aaro Koskinen wrote:
> Hi,
>
> On Tue, Dec 03, 2019 at 05:09:48PM -0600, Thor Thayer wrote:
>> On 11/29/19 10:57 AM, Aaro Koskinen wrote:
>>> Hi,
>>>
>>> I tried booting v5.4 mainline kernel on a stratix10 board with ECC
>>> disabled, and the altera-edac driver (with only SDRAM enabled) is
>>> now crashing the system instead of failing the probe with "No ECC/ECC
>>> disabled".
>>>
>> I apologize for the late reply. I was on vacation.
>>
>> ECC disabled means the sof/jic that you're loading has ECC disabled,
>> correct?
>
> Yes.
>
>>> This seems to have started with commit 08f08bfb7b4c ("EDAC, altera:
>>> Merge Stratix10 into the Arria10 SDRAM probe routine"). With the change,
>>> looks like sdram probe no longer uses SMC calls and instead accesses
>>> the registers directly. The crash looks like this:
>>
>> I haven't seen this. I'd expect both ECC enabled and disabled to fail with
>> the dumps you have below since they'd both need to use the regmap functions.
>
> With ECC enabled it doesn't fail, as the direct register access appears
> to work then (I also checked by reading 0xf8011101 from userspace -
> and it works without an abort).
>
>> Yes, this does look like it is using the register accesses instead of the
>> SMC call. Line 2206 sets the SMC call after determining from the if()
>> statement if it is a Stratix10 or Arria10 and from below it seems to take
>> the Arria10 path.
>
> But that's setting the ecc_mgr_map. I think that altr_check_ecc_deps
> and altr_sdram_probe use a different mapping. Before commit 08f08bfb7b4c
> there was S10 specific altr_s10_sdram_probe() that took care of the SMC,
> but I cannot see how the current code doing that unless I'm missing some
> special magic.
>
>> The altr_check_ecc_deps() call is checking whether ECC is enabled so the
>> probe should fail.
>>
>> I suspect the device tree. Can you verify the following node is in your
>> device tree?
>>
>> sdramedac {
>> compatible = "altr,sdram-edac-s10";
>> altr,sdr-syscon = <&sdr>;
>> interrupts = <16 4>;
>> };
>
> Yes, I'm using the in-tree socfpga_stratix10.dtsi.
>
> A.
>
This appears to be a setup problem in U-Boot where U-Boot only allows
access to the ECC registers if ECC is enabled. However, the Linux EDAC
driver reads the ECC registers to determine if ECC is enabled or not.
A patch that always allows access to the ECC registers has been posted
to U-Boot.
https://patchwork.ozlabs.org/patch/1205274/
Thanks,
Thor