Skip to contentSkip to conversation

configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

PATCH New Bartosz Bilas 5 participants 14 messages Nov 3, 2019 – Jan 3, 2020 Arm STM32

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 configs/stm32mp157c_dk2_defconfig | 1 +
 1 file changed, 1 insertion(+)
configs/stm32mp157c_dk2_defconfig +1 -0
index a1e61d752d..ba044e5e55 100644
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
 BR2_TARGET_ROOTFS_EXT2=y
 BR2_TARGET_ROOTFS_EXT2_4=y
 BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"
 # BR2_TARGET_ROOTFS_TAR is not set
 BR2_TARGET_UBOOT=y
 BR2_TARGET_UBOOT_BUILD_SYSTEM_KCONFIG=y
Signature · 1 line
2.23.0
#1 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

Since we know the proper solution, remove that faulty patch.

---
 ...-fix-checking-the-presence-of-an-env.patch | 37 -------------------
 configs/stm32mp157c_dk2_defconfig             |  1 -
 2 files changed, 38 deletions(-)
 delete mode 100644 board/stmicroelectronics/stm32mp157c-dk2/patches/uboot/0001-stm32mp1-configs-fix-checking-the-presence-of-an-env.patch
board/stmicroelectronics/stm32mp157c-dk2/patches/uboot/0001-stm32mp1-configs-fix-checking-the-presence-of-an-env.patch deleted +0 -37
deleted file mode 100644
index 695c8bbc42..0000000000
@@ -1,37 +0,0 @@
-From a7204e5ae1e5f87ad449c7425cf7614205e60734 Mon Sep 17 00:00:00 2001
-From: Bartosz Bilas <[email protected]>
-Date: Sun, 27 Oct 2019 09:01:12 +0100
-Subject: [PATCH] stm32mp1: configs: fix checking the presence of an
- environment
-
-Execute env check command within extra env settings section instead of
-bootcmd whereby we are able to mount rootfs partition from sd card properly.
-
-Signed-off-by: Bartosz Bilas <[email protected]>
----
- include/configs/stm32mp1.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/include/configs/stm32mp1.h b/include/configs/stm32mp1.h
-index 988992b336..cadc0358fd 100644
---- a/include/configs/stm32mp1.h
-+++ b/include/configs/stm32mp1.h
-@@ -115,7 +115,6 @@
- 	"if test ${boot_device} = serial || test ${boot_device} = usb;" \
- 	"then stm32prog ${boot_device} ${boot_instance}; " \
- 	"else " \
--		"run env_check;" \
- 		"if test ${boot_device} = mmc;" \
- 		"then env set boot_targets \"mmc${boot_instance}\"; fi;" \
- 		"if test ${boot_device} = nand;" \
-@@ -160,6 +159,7 @@
- 	"initrd_high=0xffffffff\0" \
- 	"altbootcmd=run bootcmd\0" \
- 	"env_default=1\0" \
-+	"run env_check;" \
- 	"env_check=if test $env_default -eq 1;"\
- 		" then env set env_default 0;env save;fi\0" \
- 	STM32MP_BOOTCMD \
--- 
-2.23.0
-
configs/stm32mp157c_dk2_defconfig +0 -1
index ba044e5e55..7d3cca9987 100644
@@ -1,7 +1,6 @@
 BR2_arm=y
 BR2_cortex_a7=y
 BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_5_3=y
-BR2_GLOBAL_PATCH_DIR="board/stmicroelectronics/stm32mp157c-dk2/patches"
 BR2_ROOTFS_OVERLAY="board/stmicroelectronics/stm32mp157c-dk2/overlay/"
 BR2_ROOTFS_POST_IMAGE_SCRIPT="support/scripts/genimage.sh"
 BR2_ROOTFS_POST_SCRIPT_ARGS="-c board/stmicroelectronics/stm32mp157c-dk2/genimage.cfg"
Signature · 1 line
2.23.0
#2 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

Quoted 3 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Bartosz" == Bartosz Bilas <[email protected]> writes:

Since we know the proper solution, remove that faulty patch.

Ehh, can you be a bit more specific? Do you mean that the change of ext4fs parameters fix the issue that this is trying to work around?

Signature · 1 line
Bye, Peter Korsgaard
#3 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

Hello Peter,

Quoted 5 lines · Peter Korsgaard
Quoted 2 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Bartosz" == Bartosz Bilas <[email protected]> writes:

Since we know the proper solution, remove that faulty patch.

Ehh, can you be a bit more specific? Do you mean that the change of ext4fs parameters fix the issue that this is trying to work around?

I'm sorry for that - it was so obvious for me during patches generation. Yes, I do - disabling new features that are enabled by default in new mkfs version fix the issue with mounting rootfs partition.

Best
Bartek

#4 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

In reply to Bartosz Bilas
Quoted 8 lines · On Sun, 3 Nov 2019 12:50:17 +0100 Bartosz Bilas <[email protected]>…

Since we know the proper solution, remove that faulty patch.

---
 ...-fix-checking-the-presence-of-an-env.patch | 37 -------------------
 configs/stm32mp157c_dk2_defconfig             |  1 -
 2 files changed, 38 deletions(-)
 delete mode 100644 board/stmicroelectronics/stm32mp157c-dk2/patches/uboot/0001-stm32mp1-configs-fix-checking-the-presence-of-an-env.patch

Applied to master after improving the commit log. Thanks!

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#5 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

Hi Bartosz,

Is possible to bump linux kernel to version 5.4.7? The 5.3.x series has been EOL, the "longterm maintenance" series will be acceptable.

Thanks.

Quoted 15 lines · On Thu, Jan 2, 2020, 00:21 Thomas Petazzoni <[email protected]>…
Quoted 8 lines · On Sun, 3 Nov 2019 12:50:17 +0100 Bartosz Bilas <[email protected]>…

Since we know the proper solution, remove that faulty patch.

---
 ...-fix-checking-the-presence-of-an-env.patch | 37 -------------------
 configs/stm32mp157c_dk2_defconfig             |  1 -
 2 files changed, 38 deletions(-)
 delete mode 100644

board/stmicroelectronics/stm32mp157c-dk2/patches/uboot/0001-stm32mp1-configs-fix-checking-the-presence-of-an-env.patch

Applied to master after improving the commit log. Thanks!

Quoted 10 lines

Thomas
--
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20200103/a2e2ff23/attachment.html>

#6 lore raw

PATCH 2/2 board/stm32mp157c-dk2: remove unnecessary workaround

Hello Scott,

sure. I'll check if the board boots with the newer kernel and if so I'll send a patch at the weekend.

Best
Bartek

Quoted 39 lines · Scott Fan

Hi Bartosz,

Is possible to bump linux kernel to version 5.4.7? The 5.3.x series has been EOL, ?the "longterm maintenance" series will be acceptable.

Thanks.

On Thu, Jan 2, 2020, 00:21 Thomas Petazzoni <thomas.petazzoni at bootlin.com <mailto:[email protected]>> wrote:

    On Sun,? 3 Nov 2019 12:50:17 +0100
    Bartosz Bilas <[email protected]
    <mailto:[email protected]>> wrote:

    > Since we know the proper solution, remove that faulty patch.
    >
    > Signed-off-by: Bartosz Bilas <[email protected]
    <mailto:[email protected]>>
    > ---
    >? ...-fix-checking-the-presence-of-an-env.patch | 37
    -------------------
    >? configs/stm32mp157c_dk2_defconfig? ? ? ? ? ? ?|? 1 -
    >? 2 files changed, 38 deletions(-)
    >? delete mode 100644
    board/stmicroelectronics/stm32mp157c-dk2/patches/uboot/0001-stm32mp1-configs-fix-checking-the-presence-of-an-env.patch

    Applied to master after improving the commit log. Thanks!

    Thomas
    -- 
    Thomas Petazzoni, CTO, Bootlin
    Embedded Linux and Kernel engineering
    https://bootlin.com
    _______________________________________________
    buildroot mailing list
    buildroot at busybox.net <mailto:[email protected]>
    http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 18 lines
*BARTOSZ BI?AS*
Embedded Software Developer 	GRINN
skype:
TEL.+48 71 716 40 99
WEB. grinn-global.com
	Grinn sp. z o.o.
Strzegomska 140a
54-429 Wroc?aw
	KRS:0000230049
NIP:8992730302
REGON:020047322 	S?d Rejonowy we Wroc?awiu
VI Wydzia? Gospodarczy
Kapita? zak?adowy: 125.000,00 z?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20200103/07578b28/attachment.html>
#7 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

In reply to Bartosz Bilas
Quoted 16 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Bartosz" == Bartosz Bilas <[email protected]> writes:

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 configs/stm32mp157c_dk2_defconfig | 1 +
 1 file changed, 1 insertion(+)
configs/stm32mp157c_dk2_defconfig +0 -0
index a1e61d752d..ba044e5e55 100644

configs/stm32mp157c_dk2_defconfig BR2_LINUX_KERNEL_INSTALL_TARGET=y · line 19

Quoted 4 lines
configs/stm32mp157c_dk2_defconfig +1 -0
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
 BR2_TARGET_ROOTFS_EXT2=y
 BR2_TARGET_ROOTFS_EXT2_4=y
 BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Is this a general issue with the ext4 support in u-boot? Should we do this generally by default, similar to how we do it for 64bit?

Signature · 1 line
Bye, Peter Korsgaard
#8 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

Hello Peter,

Quoted 10 lines · Peter Korsgaard
Quoted 9 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Bartosz" == Bartosz Bilas <[email protected]> writes:

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 configs/stm32mp157c_dk2_defconfig | 1 +
 1 file changed, 1 insertion(+)
Quoted 13 lines

configs/stm32mp157c_dk2_defconfig BR2_LINUX_KERNEL_INSTALL_TARGET=y · line 19

Quoted 9 lines
configs/stm32mp157c_dk2_defconfig +1 -0
index a1e61d752d..ba044e5e55 100644
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
 BR2_TARGET_ROOTFS_EXT2=y
 BR2_TARGET_ROOTFS_EXT2_4=y
 BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Is this a general issue with the ext4 support in u-boot? Should we do this generally by default, similar to how we do it for 64bit?

probably in case when device saves env in file on ext4 it is, but it should be checked on the other boards in order to inquiry if I'm right. As I know the requests done by U-Boot are acceptable if the ext4 file system is generated with only the features supported by U-Boot. Unfortunately I don't have any board except that which allows me to do so.

Best
Bartek

#9 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

In reply to Bartosz Bilas

Hello Bartosz,

Quoted 14 lines · On Sun, 3 Nov 2019 12:50:16 +0100 Bartosz Bilas <[email protected]>…

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 configs/stm32mp157c_dk2_defconfig | 1 +
 1 file changed, 1 insertion(+)
configs/stm32mp157c_dk2_defconfig +0 -0
index a1e61d752d..ba044e5e55 100644

configs/stm32mp157c_dk2_defconfig BR2_LINUX_KERNEL_INSTALL_TARGET=y · line 19

Quoted 4 lines
configs/stm32mp157c_dk2_defconfig +1 -0
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
 BR2_TARGET_ROOTFS_EXT2=y
 BR2_TARGET_ROOTFS_EXT2_4=y
 BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Are you sure it is necessary to disable those three options?

Indeed, U-Boot ext4 write support clearly excludes filesystems that have the metadata_csum option enabled:

https://gitlab.denx.de/u-boot/u-boot/blob/master/fs/ext4/ext4_write.c#L880

However, there's nothing about the 64bit and dir_index options. How did you conclude that they were causing problems? Did you try with just -O ^metadata_csum ?

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#10 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

Hello Thomas,

Quoted 25 lines · Thomas Petazzoni

Hello Bartosz,

Quoted 14 lines · On Sun, 3 Nov 2019 12:50:16 +0100 Bartosz Bilas <[email protected]>…

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
  configs/stm32mp157c_dk2_defconfig | 1 +
  1 file changed, 1 insertion(+)
configs/stm32mp157c_dk2_defconfig +0 -0
index a1e61d752d..ba044e5e55 100644

configs/stm32mp157c_dk2_defconfig BR2_LINUX_KERNEL_INSTALL_TARGET=y · line 19

Quoted 4 lines
configs/stm32mp157c_dk2_defconfig +1 -0
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
  BR2_TARGET_ROOTFS_EXT2=y
  BR2_TARGET_ROOTFS_EXT2_4=y
  BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Are you sure it is necessary to disable those three options?

Quoted 7 lines

Indeed, U-Boot ext4 write support clearly excludes filesystems that have the metadata_csum option enabled:

   https://gitlab.denx.de/u-boot/u-boot/blob/master/fs/ext4/ext4_write.c#L880

However, there's nothing about the 64bit and dir_index options. How did you conclude that they were causing problems? Did you try with just -O ^metadata_csum ?

I was talking with the maintainer of that board on U-Boot mailing list
[1] and he said that I should disable those 2 options (dir_index and metada_csum) to have U-Boot working properly. I'm not sure about 64bit because I saw that was set by default in buildroot. I didn't check that with metadata_csum option only but I'll do that and let you know if it works.

[1] https://lists.denx.de/pipermail/u-boot/2019-October/388618.html

Best
Bartek

Quoted 4 lines

Thanks,

Thomas

#11 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

Quoted 37 lines · Bartosz Bilas

Hello Thomas,

Quoted 26 lines · Thomas Petazzoni

Hello Bartosz,

Quoted 15 lines · On Sun,? 3 Nov 2019 12:50:16 +0100 Bartosz Bilas <[email protected]>…

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
? configs/stm32mp157c_dk2_defconfig | 1 +
? 1 file changed, 1 insertion(+)

configs/stm32mp157c_dk2_defconfig +0 -0
index a1e61d752d..ba044e5e55 100644
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
Quoted 4 lines

? BR2_TARGET_ROOTFS_EXT2=y
? BR2_TARGET_ROOTFS_EXT2_4=y
? BR2_TARGET_ROOTFS_EXT2_SIZE="120M" +BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Are you sure it is necessary to disable those three options?

Quoted 7 lines

Indeed, U-Boot ext4 write support clearly excludes filesystems that have the metadata_csum option enabled:

?? https://gitlab.denx.de/u-boot/u-boot/blob/master/fs/ext4/ext4_write.c#L880
However, there's nothing about the 64bit and dir_index options. How did you conclude that they were causing problems? Did you try with just -O ^metadata_csum ?

Quoted 5 lines

I was talking with the maintainer of that board on U-Boot mailing list [1] and he said that I should disable those 2 options (dir_index and metada_csum) to have U-Boot working properly. I'm not sure about 64bit because I saw that was set by default in buildroot. I didn't check that with metadata_csum option only but I'll do that and let you know if it works.

As explained in the help text of the option:

The default is "-O ^64bit", i.e. disable 64-bit filesystem support. This default value has been chosen because U-Boot versions before 2017.02 don't support this filesystem
option: using it may make the filesystem unreadable by U-Boot.

Since this defconfig uses a more recent U-Boot, it's not necessary.

That said, if those two options should be disabled for U-Boot to work, it's
probably a good idea to add them to the defaults (i.e. all three options, like you put in the defconfig here).

Regards,
Arnout

Regards,
Arnout

#12 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

Hello guys,

Quoted 51 lines · Arnout Vandecappelle
Quoted 36 lines · Bartosz Bilas

Hello Thomas,

Quoted 26 lines · Thomas Petazzoni

Hello Bartosz,

Quoted 15 lines · On Sun,? 3 Nov 2019 12:50:16 +0100 Bartosz Bilas <[email protected]>…

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 ? configs/stm32mp157c_dk2_defconfig | 1 +

? 1 file changed, 1 insertion(+)

configs/stm32mp157c_dk2_defconfig +0 -0
index a1e61d752d..ba044e5e55 100644

configs/stm32mp157c_dk2_defconfig BR2_LINUX_KERNEL_INSTALL_TARGET=y · line 19

Quoted 4 lines
configs/stm32mp157c_dk2_defconfig +1 -0
@@ -16,6 +16,7 @@ BR2_LINUX_KERNEL_INSTALL_TARGET=y
 ? BR2_TARGET_ROOTFS_EXT2=y
 ? BR2_TARGET_ROOTFS_EXT2_4=y
 ? BR2_TARGET_ROOTFS_EXT2_SIZE="120M"
+BR2_TARGET_ROOTFS_EXT2_MKFS_OPTIONS="-O ^64bit,^metadata_csum,^dir_index"

Are you sure it is necessary to disable those three options?

Quoted 7 lines

Indeed, U-Boot ext4 write support clearly excludes filesystems that have the metadata_csum option enabled:

?? https://gitlab.denx.de/u-boot/u-boot/blob/master/fs/ext4/ext4_write.c#L880
However, there's nothing about the 64bit and dir_index options. How did you conclude that they were causing problems? Did you try with just -O ^metadata_csum ?

Quoted 5 lines

I was talking with the maintainer of that board on U-Boot mailing list [1] and he said that I should disable those 2 options (dir_index and metada_csum) to have U-Boot working properly. I'm not sure about 64bit because I saw that was set by default in buildroot. I didn't check that with metadata_csum option only but I'll do that and let you know if it works.

As explained in the help text of the option:

The default is "-O ^64bit", i.e. disable 64-bit filesystem support. This default value has been chosen because U-Boot versions before 2017.02 don't support this filesystem
option: using it may make the filesystem unreadable by U-Boot.

Quoted 5 lines

Since this defconfig uses a more recent U-Boot, it's not necessary.

That said, if those two options should be disabled for U-Boot to work, it's
probably a good idea to add them to the defaults (i.e. all three options, like you put in the defconfig here).

I've checked if it worked with disabled metadata_csum option only and it turns out that Thomas is right. Kernel is able to mount rootfs partition properly even if I do any environment operation in U-Boot. So it means that the default 64bit option is not necessary as well.

Quoted 8 lines

Regards,
Arnout

Regards,
Arnout

Best
Bartek

#13 lore raw

PATCH 1/2 configs/stm32mp157c_dk2: disable mkfs metadata_csum and dir_index options

In reply to Bartosz Bilas
Quoted 8 lines · On Sun, 3 Nov 2019 12:50:16 +0100 Bartosz Bilas <[email protected]>…

To solve issue with non-possibilities to mount rootfs partition we should disable a new mkfs features such as metadata_csum and dir_index because there is incompatibility with these options.

---
 configs/stm32mp157c_dk2_defconfig | 1 +
 1 file changed, 1 insertion(+)

Applied to master after changing the patch to only disable the metadata_csum option, as we discussed in the thread following this patch.

Thanks!

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com