[PATCH] mtd: spi-nor: Fix the qspi parallel memory issues

[Abbarapu, Venkatesh]

Hi Dominik,
I have tried on the zc706 board and didn’t see any issues with the parallel configuration.
This is the upstream commit f96e5133184aac377b54bd61f8abf78710f2ee56 + patch attached
Zynq> sf probe 0 0 0
SF: Detected n25q512ax3 with page size 512 Bytes, erase size 128 KiB, total 128 MiB
Verified with the below tests and didn’t see any issue.
Log:
U-Boot 2025.01-rc3-00046-gf96e5133184a-dirty (Dec 04 2024 - 19:50:42 +0530)

CPU:   Zynq 7z045
Silicon: v3.1
Model: Xilinx ZC706 board
DRAM:  ECC disabled 1 GiB
jedec_spi_nor flash at 0: has no valid 'reg' property (0)
Core:  31 devices, 19 uclasses, devicetree: board
Flash: 0 Bytes
NAND:  0 MiB
MMC:   mmc at e0100000: 0
Loading Environment from nowhere... OK
In:    serial at e0001000
Out:   serial at e0001000
Err:   serial at e0001000
Net:
ZYNQ GEM: e000b000, mdio bus e000b000, phyaddr 7, interface rgmii-id

Warning: ethernet at e000b000 (eth0) using random MAC address - b6:69:a8:60:e2:de
eth0: ethernet at e000b000
Hit any key to stop autoboot:  0
Zynq> sf probe 0 0 0
SF: Detected n25q512ax3 with page size 512 Bytes, erase size 128 KiB, total 128 MiB
Zynq> time sf erase 0x0 0x2000000;mw.b 0x8000 aabbccdd 0x2000000;time sf write 0x8000 0x0 0x2000000;mw.b 0x8008000 0x0 0x2000000;time sf read 0x8008000 0x0 0x2000000;cmp.b 0x8000 0x8008000 0x2000000
SF: 33554432 bytes @ 0x0 Erased: OK

time: 3 minutes, 37.604 seconds
device 0 offset 0x0, size 0x2000000
SF: 33554432 bytes @ 0x0 Written: OK

time: 47.074 seconds
device 0 offset 0x0, size 0x2000000
SF: 33554432 bytes @ 0x0 Read: OK

time: 5.411 seconds
Total of 33554432 byte(s) were the same
Zynq> time sf erase 0x0 0x4000000; mw.b 0x10000000 dd 0x4000000; time sf write 0x10000000 0x0 0x4000000; mw.b 0x20000000 0x0 0x4000000; time sf read 0x20000000 0x0 0x4000000; cmp.b 0x10000000 0x20000000 0x4000000
SF: 67108864 bytes @ 0x0 Erased: OK

time: 6 minutes, 45.941 seconds
device 0 offset 0x0, size 0x4000000
SF: 67108864 bytes @ 0x0 Written: OK

time: 1 minutes, 34.443 seconds
device 0 offset 0x0, size 0x4000000
SF: 67108864 bytes @ 0x0 Read: OK

time: 10.821 seconds
Total of 67108864 byte(s) were the same
Zynq> time sf erase 0x0 0x8000000;mw.b 0x8000 aabbccdd 0x8000000;time sf write 0x8000 0x0 0x8000000;mw.b 0x8008000 0x0 0x8000000;time sf read 0x8008000 0x0 0x8000000;cmp.b 0x8000 0x8008000 0x8000000
SF: 134217728 bytes @ 0x0 Erased: OK

time: 12 minutes, 50.601 seconds
device 0 whole chip
SF: 134217728 bytes @ 0x0 Written: OK

time: 3 minutes, 6.852 seconds
device 0 whole chip
SF: 134217728 bytes @ 0x0 Read: OK

time: 21.643 seconds
Total of 134217728 byte(s) were the same
Zynq>

Let me know if you have any questions.

Thanks
Venkatesh

From: Dominik Wernberger <dominik.wernberger at rohde-schwarz.com>
Sent: Tuesday, December 3, 2024 5:57 PM
To: Abbarapu, Venkatesh <venkatesh.abbarapu at amd.com>
Cc: git (AMD-Xilinx) <git at amd.com>; jagan at amarulasolutions.com; Simek, Michal <michal.simek at amd.com>; trini at konsulko.com; u-boot at lists.denx.de; vigneshr at ti.com
Subject: RE: [PATCH] mtd: spi-nor: Fix the qspi parallel memory issues

Hi Venkatesh,

Thanks for such a fast response.

Running ‘sf probe’ gives:
Flash: SF: Detected n25q256a with page size 512 Bytes, erase size 128 KiB, total 128 MiB

To be correct we use 2x MT25QL256ABA1EW9 [1, 2] which is basically the same as the n25q256a reported by sf.
In defconfig I’ve specified CONFIG_ENV_SECT_SIZE=0x00040000 which is divided by 2 in parallel memories
logic [3], resulting in the 128 KiB per Flash chip.

If you need any further information, please let me know.

[1] https://www.micron.com/products/storage/nor-flash/serial-nor/part-catalog/part-detail/mt25ql256aba1ew9-0sit
[2] https://xonstorage.blob.core.windows.net/pdf/micron_mt25ql256aba1ew90sit_apr22_xonlink.pdf
[3] https://lore.kernel.org/all/20240926045508.1492638-3-venkatesh.abbarapu@amd.com/

BR
Dominik

From: Abbarapu, Venkatesh <venkatesh.abbarapu at amd.com<mailto:venkatesh.abbarapu at amd.com>>
Sent: Tuesday, December 3, 2024 1:01 PM
To: Wernberger Dominik (1GS4) <dominik.wernberger at rohde-schwarz.com<mailto:dominik.wernberger at rohde-schwarz.com>>
Cc: git (AMD-Xilinx) <git at amd.com<mailto:git at amd.com>>; jagan at amarulasolutions.com<mailto:jagan at amarulasolutions.com>; Simek, Michal <michal.simek at amd.com<mailto:michal.simek at amd.com>>; trini at konsulko.com<mailto:trini at konsulko.com>; u-boot at lists.denx.de<mailto:u-boot at lists.denx.de>; vigneshr at ti.com<mailto:vigneshr at ti.com>
Subject: *EXT* RE: [PATCH] mtd: spi-nor: Fix the qspi parallel memory issues

Hi Dominik,
Could you please let me know the details of the flash? I will check from my side.

Thanks
Venkatesh
From: Dominik Wernberger <dominik.wernberger at rohde-schwarz.com<mailto:dominik.wernberger at rohde-schwarz.com>>
Sent: Tuesday, December 3, 2024 5:21 PM
To: Abbarapu, Venkatesh <venkatesh.abbarapu at amd.com<mailto:venkatesh.abbarapu at amd.com>>
Cc: git (AMD-Xilinx) <git at amd.com<mailto:git at amd.com>>; jagan at amarulasolutions.com<mailto:jagan at amarulasolutions.com>; Simek, Michal <michal.simek at amd.com<mailto:michal.simek at amd.com>>; trini at konsulko.com<mailto:trini at konsulko.com>; u-boot at lists.denx.de<mailto:u-boot at lists.denx.de>; vigneshr at ti.com<mailto:vigneshr at ti.com>
Subject: Re: [PATCH] mtd: spi-nor: Fix the qspi parallel memory issues

Hello Venkatesh,

Is there any news on the status of this patch? I'm working with a Zynq 7z020 system where 2 Flashes
operate in parallel QSPI mode. This patch is essential for the correct Flash erase in U-Boot v2025.01-rc3.

Without this patch Flash erase fails on every second multiple of Flash erase section size (FESS). Failing
means that the Flash erase is not initiated or successful. I.e., the Flash section is not altered.

For example, lets assume a FESS with 0x20000 (128 kiB).

sf test 00000 20000 # Works
sf test 20000 20000 # Fails
sf test 40000 20000 # Works
sf test 60000 20000 # Fails
# ...

Note that executing

sf erase 20000 20000
SF: 131072 bytes @ 0x20000 Erased: OK

shows OK even though the Flash sector were not erased.

In case you have any new/rebased patches you want me to test, I would be happy to do so. I'm looking
forward getting this patch into mainline U-Boot.

Best Regards,

Dominik


Content provided within this e-mail including any attachments, is for the use of the intended recipients and may contain Rohde & Schwarz company restricted information. Any unauthorized use, disclosure, or distribution of this communication in whole or in part is strictly prohibited. If you are not the intended recipient, please notify the sender by reply email or by telephone and delete the communication in its entirety.


Content provided within this e-mail including any attachments, is for the use of the intended recipients and may contain Rohde & Schwarz company restricted information. Any unauthorized use, disclosure, or distribution of this communication in whole or in part is strictly prohibited. If you are not the intended recipient, please notify the sender by reply email or by telephone and delete the communication in its entirety.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-mtd-spi-nor-Fix-the-qspi-parallel-memory-issues.patch
Type: application/octet-stream
Size: 3398 bytes
Desc: 0001-mtd-spi-nor-Fix-the-qspi-parallel-memory-issues.patch
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20241205/6039d2f2/attachment.obj>