Dear Sascha, Thank you for your reply. On 2024/10/18 下午 05:40, Sascha Hauer wrote:
On Fri, Oct 18, 2024 at 02:25:19AM +0000, Hui-Ping Chen wrote:+static int ma35_nfi_ecc_check(struct nand_chip *chip, u8 *addr) +{ + struct ma35_nand_info *nand = nand_get_controller_data(chip); + struct mtd_info *mtd = nand_to_mtd(chip); + int i, j, nchunks = 0; + int report_err = 0; + int err_cnt = 0; + u32 status; + + nchunks = mtd->writesize / chip->ecc.steps; + if (nchunks < 4) + nchunks = 1; + else + nchunks /= 4; + + for (j = 0; j < nchunks; j++) { + status = readl(nand->regs + MA35_NFI_REG_NANDECCES0 + j * 4); + if (!status) + continue; + + for (i = 0; i < 4; i++) { + if ((status & ECC_STATUS_MASK) == 0x01) { + /* Correctable error */ + err_cnt = (status >> 2) & ECC_ERR_CNT_MASK; + ma35_nfi_correct(chip, j * 4 + i, err_cnt, addr); + report_err += err_cnt;This is still wrong. This should be report_err = max(report_err, err_cnt); You have to search for the subpage that has the most bitflips and return the number of bitflips on that subpage, *not* the total amount of errors on the whole page. That said, the statistic counter should still count the total number of bitflips occurred.
I re-examined the return value of read_page, and it indeed needs to return the
maximum bitflips for the subpage. I will correct this return value. Thank you.
Sascha+ } else { + /* Uncorrectable error */ + dev_warn(nand->dev, "uncorrectable error! 0x%4x\n", status); + return -EBADMSG; + } + status >>= 8; + } + } + return report_err; +} +
Best regards, Hui-Ping Chen
