[PATCH 02/14] drm/amdgpu: add ACA kernel hardware error log support

[Wang, Yang(Kevin)]

[AMD Official Use Only - General]


_____________________________________________
From: Zhang, Hawking <Hawking.Zhang at amd.com>
Sent: Wednesday, January 3, 2024 5:36 PM
To: Wang, Yang(Kevin) <KevinYang.Wang at amd.com>; amd-gfx at lists.freedesktop.org
Cc: Zhou1, Tao <Tao.Zhou1 at amd.com>; Chai, Thomas <YiPeng.Chai at amd.com>
Subject: RE: [PATCH 02/14] drm/amdgpu: add ACA kernel hardware error log support


[AMD Official Use Only - General]



+       dev_info(adev->dev, "[Hardware error] Accelerator Check Architecture events logged\n");
+       /* plus 1 for output format, e.g: ACA[08/08]: xxxx */
+       for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+               dev_info(adev->dev, "[Hardware error] ACA[%02d/%02d].%s=0x%016llx\n",
+                        idx + 1, total, aca_regs[i].name, bank->regs[aca_regs[i].reg_idx]);

We should keep the ACA log format simple since there are tools like crash dumper that grab these logs.

How about formatting log as below
dev_info(adev->dev, "[Hardware error] Accelerator Check Architecture (ACA) events logged\n");
dev_info(adev→dev, "[Hardware error] ACA.%s=0x%016llx\n");

In general, if the idx doesn't convey useful information, then just replace it with ACA.Reg.

[Kevin]:

Agree, will update it in next version.

Best Regards,
Kevin

Thoughts?

Regards,
Hawking

-----Original Message-----
From: Wang, Yang(Kevin) <KevinYang.Wang at amd.com<mailto:KevinYang.Wang at amd.com>>
Sent: Wednesday, January 3, 2024 16:02
To: amd-gfx at lists.freedesktop.org<mailto:amd-gfx at lists.freedesktop.org>
Cc: Zhang, Hawking <Hawking.Zhang at amd.com<mailto:Hawking.Zhang at amd.com>>; Zhou1, Tao <Tao.Zhou1 at amd.com<mailto:Tao.Zhou1 at amd.com>>; Chai, Thomas <YiPeng.Chai at amd.com<mailto:YiPeng.Chai at amd.com>>; Wang, Yang(Kevin) <KevinYang.Wang at amd.com<mailto:KevinYang.Wang at amd.com>>
Subject: [PATCH 02/14] drm/amdgpu: add ACA kernel hardware error log support

add ACA kernel hardware error log support.

Signed-off-by: Yang Wang <kevinyang.wang at amd.com<mailto:kevinyang.wang at amd.com>>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c | 29 +++++++++++++++++++++++++
 1 file changed, 29 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
index 6a6f167b5380..cadeda64eded 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
@@ -100,6 +100,33 @@ static int aca_smu_get_valid_aca_count(struct amdgpu_device *adev, enum aca_erro
        return smu_funcs->get_valid_aca_count(adev, type, count);  }

+static struct aca_regs_dump {
+       const char *name;
+       int reg_idx;
+} aca_regs[] = {
+       {"CONTROL",             ACA_REG_IDX_CTL},
+       {"STATUS",              ACA_REG_IDX_STATUS},
+       {"ADDR",                ACA_REG_IDX_ADDR},
+       {"MISC",                ACA_REG_IDX_MISC0},
+       {"CONFIG",              ACA_REG_IDX_CONFG},
+       {"IPID",                ACA_REG_IDX_IPID},
+       {"SYND",                ACA_REG_IDX_SYND},
+       {"DESTAT",              ACA_REG_IDX_DESTAT},
+       {"DEADDR",              ACA_REG_IDX_DEADDR},
+       {"CONTROL_MASK",        ACA_REG_IDX_CTL_MASK},
+};
+
+static void aca_smu_bank_dump(struct amdgpu_device *adev, int idx, int
+total, struct aca_bank *bank) {
+       int i;
+
+       dev_info(adev->dev, "[Hardware error] Accelerator Check Architecture events logged\n");
+       /* plus 1 for output format, e.g: ACA[08/08]: xxxx */
+       for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+               dev_info(adev->dev, "[Hardware error] ACA[%02d/%02d].%s=0x%016llx\n",
+                        idx + 1, total, aca_regs[i].name, bank->regs[aca_regs[i].reg_idx]);
+}
+
 static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_error_type type,
                                       int start, int count,
                                       struct aca_banks *banks)
@@ -137,6 +164,8 @@ static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_erro
                if (ret)
                        return ret;

+               aca_smu_bank_dump(adev, i, count, &bank);
+
                ret = aca_banks_add_bank(banks, &bank);
                if (ret)
                        return ret;
--
2.34.1

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20240103/0de2f9f3/attachment-0001.htm>