Tool/software:
Hi, Dear Expert
This is extension discussion of this thread, because of the original thread was locked.
(1) Issues : We found linux uart log output "mmc0: running CQE recovery" , and some files crash. eMMC works on HS200, and we already make sure disable DDR mode in uboot / linux DTS, because AM62 do not support eMMC DDR mode.
(2) SDK : 10.01.10.04
(3) eMMC parts : THGBMUG6C1LBAIL (KIOXIA)
(4) SDIO0 (eMMC) signal integrity : HS200 Pass (hardware)
(5) SDIO0 (eMMC) switching characteristics : HS200 Pass (hardware)
(6) Clue : If we reduce speed to HS mode, "CQE recovery" disappear
(7) Not every board have the same issues, as far as I know, failed rate about 15%, we are still keep tracking.
(8) Error log.
* Log 1 :
[ 4.711702] mmc0: running CQE recovery
[ 4.735046] mmc0: running CQE recovery
[ 4.758311] mmc0: running CQE recovery
[ 4.766138] I/O error, dev mmcblk0, sector 102744 op 0x0:(READ) flags 0x3000 phys_seg 1 prio class 2
[ 4.775291] EXT4-fs warning (device mmcblk0p1): htree_dirblock_to_tree:1082: inode #880: lblock 0: comm
systemd: error -5 reading directory block
[ 4.788391] systemd[1]: tmp.mount: Failed to check directory /tmp: Input/output error
[ 4.815487] mmc0: running CQE recovery
[ 4.838846] mmc0: running CQE recovery
[ 4.862113] mmc0: running CQE recovery
[ 4.869930] I/O error, dev mmcblk0, sector 291832 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
* Log 2 :
[ 11.193981] mmc0: running CQE recovery
[ 11.199428] Buffer I/O error on dev mmcblk0p1, logical block 0, lost sync page write
[ 11.207195] EXT4-fs (mmcblk0p1): I/O error while writing superblock
[ 11.229373] mmc0: running CQE recovery
[ 11.252665] mmc0: running CQE recovery
[ 11.275931] mmc0: running CQE recovery
* Log 3 :
[ 1.309542] mmc0: Command Queue Engine enabled [ 1.314030] mmc0: new HS200 MMC card at address 0001 [ 1.319843] mmcblk0: mmc0:0001 008GB1 1.16 GiB [ 1.325811] mmc0: running CQE recovery [ 1.332062] mmc0: running CQE recovery [ 1.338170] GPT:Primary header thinks Alt. header is not at the end of the disk. [ 1.345632] GPT:2113535 != 2441215 [ 1.349132] GPT:Alternate GPT header not at the end of the disk. [ 1.355233] GPT:2113535 != 2441215 [ 1.358647] GPT: Use GNU Parted to correct GPT errors. [ 1.363807] mmcblk0: p1 p2 [ 1.367355] mmcblk0boot0: mmc0:0001 008GB1 4.00 MiB [ 1.373287] mmcblk0boot1: mmc0:0001 008GB1 4.00 MiB [ 1.379213] mmcblk0gp0: mmc0:0001 008GB1 2.55 GiB [ 1.385927] mmc0: running CQE recovery [ 1.392012] mmc0: running CQE recovery [ 1.397284] mmc0: running CQE recovery [ 1.402967] mmcblk0gp0: p1 p2 [ 1.406607] mmcblk0rpmb: mmc0:0001 008GB1 4.00 MiB, chardev (243:0) [ 1.474043] mmc2: Failed to initialize a non-removable card [ 2.730108] sdhci-am654 fa00000.mmc: Power on failed [ 2.771213] mmc1: SDHCI controller on fa00000.mmc [fa00000.mmc] using ADMA 64-bit [ 2.780358] mmc0: running CQE recovery [ 2.785587] mmc0: running CQE recovery [ 2.791585] mmc0: running CQE recovery\
* Log 4 :
1.205849] mmc2: CQHCI version 5.10 [ 1.230661] mmc0: SDHCI controller on fa10000.mmc [fa10000.mmc] using ADMA 64-bit [ 1.238247] mmc2: SDHCI controller on fa20000.mmc [fa20000.mmc] using ADMA 64-bit [ 1.316005] mmc0: Command Queue Engine enabled [ 1.320483] mmc0: new HS200 MMC card at address 0001 [ 1.326164] mmcblk0: mmc0:0001 008GB1 1.16 GiB [ 1.332067] mmc0: running CQE recovery [ 1.337534] mmc0: running CQE recovery [ 1.343268] mmc0: running CQE recovery [ 1.347344] I/O error, dev mmcblk0, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2 [ 1.355843] Buffer I/O error on dev mmcblk0, logical block 0, async page read
Few question :
(1) As far as I know, most of eMMC "CQE recovery" error already solved by SDK 10.01.10.04, isn't?
Because I think SDK 10.01.10.04 should be update these patch about EXT_EP-12086
https://sir.ext.ti.com/jira/browse/EXT_EP-12086?jql=text%20~%20%22running%20CQE%20recovery%22
(2) How to track uboot / linux code that we ever make any modification for eMMC issues? because I would like check one by one ...
(3) What's difference (update) between SDK 11 and SDK 10.01.10.04 about eMMC code update? I also track this thread, it seems still have eMMC issues under tracking w/ SDK11
We need some debug direction, need your suggestion.
Thank You Very Much
Gibbs
