git: 1b008564a31b - stable/15 - e1000: Report corrected I210 and I211 ECC errors

The branch stable/15 has been updated by kbowling:

URL: https://cgit.FreeBSD.org/src/commit/?id=1b008564a31b4baaaa3f9cb16fdf83c2b28967ba

commit 1b008564a31b4baaaa3f9cb16fdf83c2b28967ba
Author:     Kevin Bowling <kbowling@FreeBSD.org>
AuthorDate: 2026-08-12 18:34:15 +0000
Commit:     Kevin Bowling <kbowling@FreeBSD.org>
CommitDate: 2026-08-30 01:16:01 +0000

    e1000: Report corrected I210 and I211 ECC errors
    
    I210 and I211 do not interrupt for corrected internal ECC errors.
    Instead, the DMA packet-buffer and PCIe memories expose sticky status
    bits in PBECCSTS and PCIEECCSTS.
    
    Sample these bits with the regular hardware statistics update, preserve
    the I210/I211 PBECCSTS enable state while clearing its RW1C indication,
    and expose separate counters for the DMA packet buffer, PCIe transmit
    data, and PCIe retry buffer.
    
    The counters represent observed indications rather than exact error
    counts because multiple corrections between samples collapse into one
    sticky status bit.
    
    Hardware validation used an I210 revision 3.  Unlike I225 and I226, the
    published I210/I211 register definitions do not expose self-clearing
    injectors for these corrected ECC memories.  The three counter sysctls
    were present and remained zero under line-rate traffic and three fatal
    LAN parity recoveries.  PBECCSTS.ECC_ENABLE remained set after every
    reset.  Actual corrected-error accounting was therefore not injected.
    
    Sponsored by:   BBOX.io
    
    (cherry picked from commit 0ea53a7123ffc1ea11daa748e7148ac8413fd2de)
---
 sys/dev/e1000/if_em.c | 40 ++++++++++++++++++++++++++++++++++++++++
 sys/dev/e1000/if_em.h |  3 +++
 2 files changed, 43 insertions(+)

diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
index b9168a9a5bbe..0618995b5587 100644
--- a/sys/dev/e1000/if_em.c
+++ b/sys/dev/e1000/if_em.c
@@ -1925,6 +1925,32 @@ em_update_pch_ecc_stats(struct e1000_softc *sc, u32 pbeccsts)
 	    E1000_PBECCSTS_UNCORR_ERR_CNT_SHIFT;
 }
 
+static void
+em_update_i210_ecc_stats(struct e1000_softc *sc)
+{
+	struct e1000_hw *hw;
+	u32 pbeccsts, pcieeccsts;
+
+	hw = &sc->hw;
+	pbeccsts = E1000_READ_REG(hw, E1000_PBECCSTS_I210);
+	if (pbeccsts & E1000_PBECCSTS_I210_CORR_ERR) {
+		sc->corrected_error_dma_count++;
+		/* Preserve the enable bit while clearing the RW1C status. */
+		E1000_WRITE_REG(hw, E1000_PBECCSTS_I210,
+		    pbeccsts & (E1000_PBECCSTS_I210_ECC_ENABLE |
+		    E1000_PBECCSTS_I210_CORR_ERR));
+	}
+
+	pcieeccsts = E1000_READ_REG(hw, E1000_PCIEECCSTS) &
+	    E1000_PCIEECCSTS_CORR_MASK;
+	if (pcieeccsts & E1000_PCIEECCSTS_TX_WR_DATA)
+		sc->corrected_error_pcie_tx_data_count++;
+	if (pcieeccsts & E1000_PCIEECCSTS_RETRY_BUF)
+		sc->corrected_error_pcie_retry_count++;
+	if (pcieeccsts != 0)
+		E1000_WRITE_REG(hw, E1000_PCIEECCSTS, pcieeccsts);
+}
+
 /*
  * Fatal internal-memory errors stop part or all of the MAC.  Capture the
  * read-clear indication before handing recovery to the iflib admin task.
@@ -5371,6 +5397,8 @@ em_update_stats_counters(struct e1000_softc *sc)
 	if (em_has_pch_ecc(&sc->hw))
 		em_update_pch_ecc_stats(sc,
 		    E1000_READ_REG(&sc->hw, E1000_PBECCSTS));
+	else if (em_has_i210_memory_errors(&sc->hw))
+		em_update_i210_ecc_stats(sc);
 }
 
 static bool
@@ -5717,6 +5745,18 @@ em_add_hw_stats(struct e1000_softc *sc)
 			    "fatal_unknown", CTLFLAG_RD,
 			    &sc->fatal_error_unknown_count,
 			    "Fatal memory errors without a reported region");
+			SYSCTL_ADD_UQUAD(ctx, memerr_list, OID_AUTO,
+			    "corrected_dma", CTLFLAG_RD,
+			    &sc->corrected_error_dma_count,
+			    "Corrected DMA memory error indications");
+			SYSCTL_ADD_UQUAD(ctx, memerr_list, OID_AUTO,
+			    "corrected_pcie_tx_data", CTLFLAG_RD,
+			    &sc->corrected_error_pcie_tx_data_count,
+			    "Corrected PCIe transmit-data memory indications");
+			SYSCTL_ADD_UQUAD(ctx, memerr_list, OID_AUTO,
+			    "corrected_pcie_retry", CTLFLAG_RD,
+			    &sc->corrected_error_pcie_retry_count,
+			    "Corrected PCIe retry-buffer memory indications");
 		}
 	}
 
diff --git a/sys/dev/e1000/if_em.h b/sys/dev/e1000/if_em.h
index a75b3b1b290e..de0d078a4e72 100644
--- a/sys/dev/e1000/if_em.h
+++ b/sys/dev/e1000/if_em.h
@@ -629,6 +629,9 @@ struct e1000_softc {
 	u64			fatal_error_pcie_count;
 	u64			fatal_error_dma_count;
 	u64			fatal_error_unknown_count;
+	u64			corrected_error_dma_count;
+	u64			corrected_error_pcie_tx_data_count;
+	u64			corrected_error_pcie_retry_count;
 	u64			corrected_error_packet_buffer_count;
 	u64			uncorrected_error_packet_buffer_count;