2022-07-16 00:19:53

by Jae Hyun Yoo

[permalink] [raw]
Subject: [PATCH v3 1/2] mtd: spi-nor: fix select_uniform_erase to skip 0 erase size

From: Michael Walle <[email protected]>

4bait will set the erase size to 0 if there is no corresponding
opcode for the 4byte erase. Fix spi_nor_select_uniform_erase to skip
the 0 erase size to avoid mtd device registration failure cases.

Reported-by: Jae Hyun Yoo <[email protected]>
Signed-off-by: Michael Walle <[email protected]>
Signed-off-by: Jae Hyun Yoo <[email protected]>
---
Changes in v3:
* Added newly in v3.

drivers/mtd/spi-nor/core.c | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c
index 502967c76c5f..666ef6ce9549 100644
--- a/drivers/mtd/spi-nor/core.c
+++ b/drivers/mtd/spi-nor/core.c
@@ -2113,6 +2113,10 @@ spi_nor_select_uniform_erase(struct spi_nor_erase_map *map,

tested_erase = &map->erase_type[i];

+ /* Skip masked erase types. */
+ if (!tested_erase->size)
+ continue;
+
/*
* If the current erase size is the one, stop here:
* we have found the right uniform Sector Erase command.
--
2.25.1


2022-07-16 00:41:04

by Jae Hyun Yoo

[permalink] [raw]
Subject: [PATCH v3 2/2] mtd: spi-nor: winbond: add support for W25Q512NW-IQ

Add support for Winbond W25Q512NW-IQ/IN

Signed-off-by: Jae Hyun Yoo <[email protected]>
Link: https://www.winbond.com/resource-files/W25Q512NW%20RevB%2007192021.pdf
Reviewed-by: Cédric Le Goater <[email protected]>
---
Changes in v2:
* Change sector size and number of sector to zero so that it can use parsed
info from sfdp. (Michael)
* Added md5sum of sfdp block. (Michael)

Changes in v2:
* Refined commit message. (Michael)

Test result on AST2600 SoC's SPI controller:
$ cat /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/jedec_id
ef6020

$ cat /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/manufacturer
winbond

$ cat /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/partname
w25q512nwq

$ hexdump /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
0000000 4653 5044 0106 ff01 0600 1001 0080 ff00
0000010 0084 0201 00d0 ff00 ffff ffff ffff ffff
0000020 6f00 7074 7420 7365 ff74 ffff ffff ffff
*
0000040 ffff ffff ffff ffff ffff ffff ffff ffff
*
0000080 20e5 fffb ffff 1fff eb44 6b08 3b08 bb42
0000090 fffe ffff ffff 0000 ffff eb40 200c 520f
00000a0 d810 0000 0233 00a6 e781 d914 63e9 3376
00000b0 757a 757a bdf7 5cd5 f719 ff5d 70e9 a5f9
00000c0 ffff ffff ffff ffff ffff ffff ffff ffff
00000d0 0aff fff0 ff21 ffdc
00000d8

$ md5sum /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
e1484fe7c993adaee5ea0d6246f52817

$ flash_otp_info -u /dev/mtd0
Number of OTP user blocks on /dev/mtd0: 3
block 0: offset = 0x0000 size = 256 bytes [unlocked]
block 1: offset = 0x0100 size = 256 bytes [unlocked]
block 2: offset = 0x0200 size = 256 bytes [unlocked]

$ flash_otp_dump -u /dev/mtd0 0x2d0
OTP user data for /dev/mtd0
0x02d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
0x02e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
0x02f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

$ echo -n otp test | flash_otp_write -u /dev/mtd0 0x2d0
Writing OTP user data on /dev/mtd0 at offset 0x2d0
Wrote 8 bytes of OTP user data

$ flash_otp_dump -u /dev/mtd0 0x2d0
OTP user data for /dev/mtd0
0x02d0: 6f 74 70 20 74 65 73 74 ff ff ff ff ff ff ff ff
0x02e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
0x02f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

$ flash_otp_erase -u /dev/mtd0 0x200 0x100

$ flash_otp_dump -u /dev/mtd0 0x2d0
OTP user data for /dev/mtd0
0x02d0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
0x02e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
0x02f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

drivers/mtd/spi-nor/winbond.c | 3 +++
1 file changed, 3 insertions(+)

diff --git a/drivers/mtd/spi-nor/winbond.c b/drivers/mtd/spi-nor/winbond.c
index ffaa24055259..ca39acf4112c 100644
--- a/drivers/mtd/spi-nor/winbond.c
+++ b/drivers/mtd/spi-nor/winbond.c
@@ -133,6 +133,9 @@ static const struct flash_info winbond_nor_parts[] = {
{ "w25m512jv", INFO(0xef7119, 0, 64 * 1024, 1024)
NO_SFDP_FLAGS(SECT_4K | SPI_NOR_QUAD_READ |
SPI_NOR_DUAL_READ) },
+ { "w25q512nwq", INFO(0xef6020, 0, 0, 0)
+ PARSE_SFDP
+ OTP_INFO(256, 3, 0x1000, 0x1000) },
{ "w25q512nwm", INFO(0xef8020, 0, 64 * 1024, 1024)
PARSE_SFDP
OTP_INFO(256, 3, 0x1000, 0x1000) },
--
2.25.1

2022-07-16 09:33:34

by Michael Walle

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] mtd: spi-nor: winbond: add support for W25Q512NW-IQ

Hi,

Am 2022-07-16 02:06, schrieb Jae Hyun Yoo:
> Add support for Winbond W25Q512NW-IQ/IN
>
> Signed-off-by: Jae Hyun Yoo <[email protected]>
> Link:
> https://www.winbond.com/resource-files/W25Q512NW%20RevB%2007192021.pdf
> Reviewed-by: Cédric Le Goater <[email protected]>

Thanks again for finding the real issue here!

Reviewed-by: Michael Walle <[email protected]>

But also see below.

> $ hexdump
> /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
> 0000000 4653 5044 0106 ff01 0600 1001 0080 ff00
> 0000010 0084 0201 00d0 ff00 ffff ffff ffff ffff
> 0000020 6f00 7074 7420 7365 ff74 ffff ffff ffff

Did you noticed your SFDP data changed compared to:
https://lore.kernel.org/linux-mtd/[email protected]/

Which makes me wonder if you ever had an offset 0
with the OTP_INFO(). For Winbond the first OTP
region is actually the SFDP.. So you might
accidentally modified it.

-michael

2022-07-16 13:54:04

by Jae Hyun Yoo

[permalink] [raw]
Subject: Re: [PATCH v3 2/2] mtd: spi-nor: winbond: add support for W25Q512NW-IQ

Hi,

On 7/16/2022 2:11 AM, Michael Walle wrote:
> Hi,
>
> Am 2022-07-16 02:06, schrieb Jae Hyun Yoo:
>> Add support for Winbond W25Q512NW-IQ/IN
>>
>> Signed-off-by: Jae Hyun Yoo <[email protected]>
>> Link:
>> https://www.winbond.com/resource-files/W25Q512NW%20RevB%2007192021.pdf
>> Reviewed-by: Cédric Le Goater <[email protected]>
>
> Thanks again for finding the real issue here!
>
> Reviewed-by: Michael Walle <[email protected]>

Thanks for your review!

> But also see below.
>
>> $ hexdump
>> /sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
>>
>> 0000000 4653 5044 0106 ff01 0600 1001 0080 ff00
>> 0000010 0084 0201 00d0 ff00 ffff ffff ffff ffff
>> 0000020 6f00 7074 7420 7365 ff74 ffff ffff ffff
>
> Did you noticed your SFDP data changed compared to:
> https://lore.kernel.org/linux-mtd/[email protected]/
>
>
> Which makes me wonder if you ever had an offset 0
> with the OTP_INFO(). For Winbond the first OTP
> region is actually the SFDP.. So you might
> accidentally modified it.

Indeed. Looks like the region got modified while I was testing OTP.
Please see below. I took the sfdp dump from a new clean part.

$ hexdump
/sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
0000000 4653 5044 0106 ff01 0600 1001 0080 ff00
0000010 0084 0201 00d0 ff00 ffff ffff ffff ffff
0000020 ffff ffff ffff ffff ffff ffff ffff ffff
*
0000080 20e5 fffb ffff 1fff eb44 6b08 3b08 bb42
0000090 fffe ffff ffff 0000 ffff eb40 200c 520f
00000a0 d810 0000 0233 00a6 e781 d914 63e9 3376
00000b0 757a 757a bdf7 5cd5 f719 ff5d 70e9 a5f9
00000c0 ffff ffff ffff ffff ffff ffff ffff ffff
00000d0 0aff fff0 ff21 ffdc
00000d8

$ md5sum
/sys/bus/platform/devices/1e620000.spi/spi_master/spi0/spi0.1/spi-nor/sfdp
347f0027ec5a09e8e78d5685485cceed

Thanks,

Jae

2022-11-21 15:50:48

by Tudor Ambarus

[permalink] [raw]
Subject: (subset)Re: [PATCH v3 1/2] mtd: spi-nor: fix select_uniform_erase to skip 0 erase size

On Fri, 15 Jul 2022 17:06:42 -0700, Jae Hyun Yoo wrote:
> From: Michael Walle <[email protected]>
>
> 4bait will set the erase size to 0 if there is no corresponding
> opcode for the 4byte erase. Fix spi_nor_select_uniform_erase to skip
> the 0 erase size to avoid mtd device registration failure cases.
>
>
> [...]

Applied [2/2] to spi-nor/next, thanks!
I haven't applied 1/2 as Michael sent a similar patch which
I had already taken at the time I got to your patch set, see:
https://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux.git/commit/?id=39eece67a3cf
Sorry!

[2/2] mtd: spi-nor: winbond: add support for W25Q512NW-IQ
https://git.kernel.org/mtd/c/ef434f08b056

Best regards,
--
Tudor Ambarus <[email protected]>