[PATCH 1/3] drm/amd/ras: skip poison banks read through the UE channel

Outside of UMC a bank with the poison bit set describes the consumption
of an error that was already reported, not a new uncorrectable one.
Injecting an uncorrectable error while a workload is running therefore
buries the real bank under a dump of GFX poison banks collected from
every other device of the hive.

Signed-off-by: Xiang Liu <[email protected]>
---
 drivers/gpu/drm/amd/ras/core/aca.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/amd/ras/core/aca.c b/drivers/gpu/drm/amd/ras/core/aca.c
index 0f5e57965ac8..f316f49c24bd 100644
--- a/drivers/gpu/drm/amd/ras/core/aca.c
+++ b/drivers/gpu/drm/amd/ras/core/aca.c
@@ -403,6 +403,13 @@ static int aca_banks_update(struct ras_core_context *ras_core,
 
 		memset(&bank_ecc, 0, sizeof(bank_ecc));
 		aca_blk = aca_get_bank_aca_block(ras_core, &bank);
+
+		/* Outside of UMC a poison bank is a consumption, not a new error. */
+		if (ecc_type == RAS_ERR_TYPE__UE &&
+		    ACA_REG_STATUS_POISON(bank.regs[ACA_REG_IDX__STATUS]) &&
+		    (!aca_blk || aca_blk->blk_info->ras_block_id != RAS_BLOCK_ID__UMC))
+			continue;
+
 		if (aca_blk)
 			ret = aca_parse_bank(ras_core, aca_blk, &bank, &bank_ecc);
 
-- 
2.34.1