Re: [U-Boot] [linux-sunxi] [UBOOT NAND] [PATCH 0/4] Add nand reading support to u-boot
Hi,
On 02-01-15 09:53, Daniel Kochmański wrote:
This is a series of patches to enable nand read functionality on sunxi devices. It uses DMA and is able to read from syndrome partitions and normal ones. Additionaly mksunxiboot tool is patched to be able to format resulting binary to be able to read from BROM, so NAND SPL builds are now possible. a20_nandread command is added for conveniance.
Afaik DMA controller doesn't vary between a10 and a20, so this should work on a10 devices, but due to no a10 hardware it wasn't tested.
First of all many thanks for working on this!
I see that this is against linux-sunxi/u-boot-sunxi.git, that is not (really) being actively developed anymore, see:
https://www.marshut.net/kuispp/linux-sunxi-u-boot-sunxi-is-no-longer-support...
Can you please rebase on top of upstream u-boot, on top of the next branch: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next ?
Your patches all seem to only have a subject and not a proper commit message, please add a commit message to all of them describing the changes in somewhat more detail then you do in the subject, you could e.g. re-use bits of your coverletter. Also for upstream u-boot we require a Signed-off-by as part of the commit message, so please add a line like this one to all your commit messages:
Signed-off-by: Daniel Kochmański dkochmanski@antmicro.com
You can make git do this automatically by specifying the -s option when committing, also please send further versions of this patch-set to the upstream u-boot list.
Besides all the above, which is easily fixed, unfortunately we also have the problem (or luxury) that Yassin Jaffer (added to the CC) has also been working on nand support, re-using the upstream kernel code Boris Brezillon has been working on, see:
https://github.com/yassinjaffer/u-boot/commits/sunxi-nand
I believe that Yassin's version is better then yours because it re-uses existing nand infrastructure rather then more or less starting from scratch.
But Yassin has not been very active on this, and has been reluctant to submit his work upstream because it has not been tested a lot yet.
The best way forward may very well be you picking up Yassin's work (in coordination with Yassin), and submitting that to upstream u-boot.
Yassin, I notice that the patches in your git repo do not have a Signed-off-by, can you please reply to this mail with your Signed-off-by to indicate that it is ok for us to add your Signed-off-by to your patches ?
Regards,
Hans
Daniel Kochmański (4): sunxi: nand: add minimal nand driver (reading flash only). sunxi: mksunxiboot: increase block size, so checksum is correct on nand sunxi: nand: add nand spl boot option sunxi: nand: add a20_nandread command
arch/arm/cpu/armv7/sunxi/board.c | 4 + board/sunxi/Makefile | 1 + board/sunxi/nand.c | 252 +++++++++++++++++++++++++++++++++++++++ common/Makefile | 1 + common/cmd_a20_nandread.c | 26 ++++ common/spl/spl_nand.c | 23 +++- include/configs/sunxi-common.h | 25 ++-- tools/mksunxiboot.c | 2 +- 8 files changed, 322 insertions(+), 12 deletions(-) create mode 100644 board/sunxi/nand.c create mode 100644 common/cmd_a20_nandread.c
Hi Hans
I was working on utilizing the DM framework for the nand driver, but this is going to take sometime , anyway I will try to submit the driver based on bbrezillon work. I'm quite happy if Daniel wants to pickup my work, besides Daniel patches are for the SPL nand driver.
Regards
On Tuesday, January 6, 2015 12:08:48 AM UTC+11, Hans de Goede wrote:
Hi,
On 02-01-15 09:53, Daniel Kochmański wrote:
This is a series of patches to enable nand read functionality on sunxi devices. It uses DMA and is able to read from syndrome partitions and normal ones. Additionaly mksunxiboot tool is patched to be able to format resulting binary to be able to read from BROM, so NAND SPL builds are now possible. a20_nandread command is added for conveniance.
Afaik DMA controller doesn't vary between a10 and a20, so this should work on a10 devices, but due to no a10 hardware it wasn't tested.
First of all many thanks for working on this!
I see that this is against linux-sunxi/u-boot-sunxi.git, that is not (really) being actively developed anymore, see:
https://www.marshut.net/kuispp/linux-sunxi-u-boot-sunxi-is-no-longer-support...
Can you please rebase on top of upstream u-boot, on top of the next branch: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next ?
Your patches all seem to only have a subject and not a proper commit message, please add a commit message to all of them describing the changes in somewhat more detail then you do in the subject, you could e.g. re-use bits of your coverletter. Also for upstream u-boot we require a Signed-off-by as part of the commit message, so please add a line like this one to all your commit messages:
Signed-off-by: Daniel Kochmański <dkoch...@antmicro.com javascript:>
You can make git do this automatically by specifying the -s option when committing, also please send further versions of this patch-set to the upstream u-boot list.
Besides all the above, which is easily fixed, unfortunately we also have the problem (or luxury) that Yassin Jaffer (added to the CC) has also been working on nand support, re-using the upstream kernel code Boris Brezillon has been working on, see:
https://github.com/yassinjaffer/u-boot/commits/sunxi-nand
I believe that Yassin's version is better then yours because it re-uses existing nand infrastructure rather then more or less starting from scratch.
But Yassin has not been very active on this, and has been reluctant to submit his work upstream because it has not been tested a lot yet.
The best way forward may very well be you picking up Yassin's work (in coordination with Yassin), and submitting that to upstream u-boot.
Yassin, I notice that the patches in your git repo do not have a Signed-off-by, can you please reply to this mail with your Signed-off-by to indicate that it is ok for us to add your Signed-off-by to your patches ?
Regards,
Hans
Daniel Kochmański (4): sunxi: nand: add minimal nand driver (reading flash only). sunxi: mksunxiboot: increase block size, so checksum is correct on nand sunxi: nand: add nand spl boot option sunxi: nand: add a20_nandread command
arch/arm/cpu/armv7/sunxi/board.c | 4 + board/sunxi/Makefile | 1 + board/sunxi/nand.c | 252
+++++++++++++++++++++++++++++++++++++++
common/Makefile | 1 + common/cmd_a20_nandread.c | 26 ++++ common/spl/spl_nand.c | 23 +++- include/configs/sunxi-common.h | 25 ++-- tools/mksunxiboot.c | 2 +- 8 files changed, 322 insertions(+), 12 deletions(-) create mode 100644 board/sunxi/nand.c create mode 100644 common/cmd_a20_nandread.c
Hi,
On 06-01-15 12:57, Yassin Jaffer wrote:
Hi Hans
I was working on utilizing the DM framework for the nand driver, but this is going to take sometime , anyway I will try to submit the driver based on bbrezillon work.
Ok, thanks. I'm looking forward to the posting of the next version of your series.
I'm quite happy if Daniel wants to pickup my work,
It looks to me like Daniel may have more time to work on this atm then you, so that might be a good idea. Either way please coordinate with each other to avoid double work. You can use private mail to Daniel for the coordination stuff if you prefer, it would be nice if you could send a short summary of who plans to do what once that is clear.
besides Daniel patches are for the SPL nand driver.
We should be able to use your adaption of Boris' work in the SPL too, for mmc we are also using the same driver in both SPL and regular u-boot, and I assume that Daniel's patches are also capable of loading say a kernel from nand, right Daniel ?
Regards,
Hans
Regards
On Tuesday, January 6, 2015 12:08:48 AM UTC+11, Hans de Goede wrote:
Hi,
On 02-01-15 09:53, Daniel Kochmański wrote:
This is a series of patches to enable nand read functionality on sunxi devices. It uses DMA and is able to read from syndrome partitions and normal ones. Additionaly mksunxiboot tool is patched to be able to format resulting binary to be able to read from BROM, so NAND SPL builds are now possible. a20_nandread command is added for conveniance.
Afaik DMA controller doesn't vary between a10 and a20, so this should work on a10 devices, but due to no a10 hardware it wasn't tested.
First of all many thanks for working on this!
I see that this is against linux-sunxi/u-boot-sunxi.git, that is not (really) being actively developed anymore, see:
https://www.marshut.net/kuispp/linux-sunxi-u-boot-sunxi-is-no-longer-support...
Can you please rebase on top of upstream u-boot, on top of the next branch: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next ?
Your patches all seem to only have a subject and not a proper commit message, please add a commit message to all of them describing the changes in somewhat more detail then you do in the subject, you could e.g. re-use bits of your coverletter. Also for upstream u-boot we require a Signed-off-by as part of the commit message, so please add a line like this one to all your commit messages:
Signed-off-by: Daniel Kochmański <dkoch...@antmicro.com javascript:>
You can make git do this automatically by specifying the -s option when committing, also please send further versions of this patch-set to the upstream u-boot list.
Besides all the above, which is easily fixed, unfortunately we also have the problem (or luxury) that Yassin Jaffer (added to the CC) has also been working on nand support, re-using the upstream kernel code Boris Brezillon has been working on, see:
https://github.com/yassinjaffer/u-boot/commits/sunxi-nand
I believe that Yassin's version is better then yours because it re-uses existing nand infrastructure rather then more or less starting from scratch.
But Yassin has not been very active on this, and has been reluctant to submit his work upstream because it has not been tested a lot yet.
The best way forward may very well be you picking up Yassin's work (in coordination with Yassin), and submitting that to upstream u-boot.
Yassin, I notice that the patches in your git repo do not have a Signed-off-by, can you please reply to this mail with your Signed-off-by to indicate that it is ok for us to add your Signed-off-by to your patches ?
Regards,
Hans
Daniel Kochmański (4): sunxi: nand: add minimal nand driver (reading flash only). sunxi: mksunxiboot: increase block size, so checksum is correct on nand sunxi: nand: add nand spl boot option sunxi: nand: add a20_nandread command
arch/arm/cpu/armv7/sunxi/board.c | 4 + board/sunxi/Makefile | 1 + board/sunxi/nand.c | 252
+++++++++++++++++++++++++++++++++++++++
common/Makefile | 1 + common/cmd_a20_nandread.c | 26 ++++ common/spl/spl_nand.c | 23 +++- include/configs/sunxi-common.h | 25 ++-- tools/mksunxiboot.c | 2 +- 8 files changed, 322 insertions(+), 12 deletions(-) create mode 100644 board/sunxi/nand.c create mode 100644 common/cmd_a20_nandread.c
Hi,
Hans de Goede writes:
Hi,
On 06-01-15 12:57, Yassin Jaffer wrote:
Hi Hans
I was working on utilizing the DM framework for the nand driver, but this is going to take sometime , anyway I will try to submit the driver based on bbrezillon work.
Ok, thanks. I'm looking forward to the posting of the next version of your series.
I'm quite happy if Daniel wants to pickup my work,
It looks to me like Daniel may have more time to work on this atm then you, so that might be a good idea. Either way please coordinate with each other to avoid double work. You can use private mail to Daniel for the coordination stuff if you prefer, it would be nice if you could send a short summary of who plans to do what once that is clear.
I'd be glad to, but I have limited time to devote for this project, and I might me not be as that helpful as you expect me to.
besides Daniel patches are for the SPL nand driver.
We should be able to use your adaption of Boris' work in the SPL too, for mmc we are also using the same driver in both SPL and regular u-boot, and I assume that Daniel's patches are also capable of loading say a kernel from nand, right Daniel ?
I'm not sure if I remember correctly, but I had problems with adaptation of lichee driver for nand, because it was too big to fit in SPL (BROM expects boot0 to fit at first 8K).
Yes, these patches are capable of this. I haven't included patch which added MBOOT functionality, since it seems unrelated to nand support. Putting it short, it was downloading multiimage from mtd2 (which had 16M in my config), and extracted from it kernel, devicetree, ramdisk and u-boot script which was performing boot.
Regards,
Hans
Regards
On Tuesday, January 6, 2015 12:08:48 AM UTC+11, Hans de Goede wrote:
Hi,
On 02-01-15 09:53, Daniel Kochmański wrote:
This is a series of patches to enable nand read functionality on sunxi devices. It uses DMA and is able to read from syndrome partitions and normal ones. Additionaly mksunxiboot tool is patched to be able to format resulting binary to be able to read from BROM, so NAND SPL builds are now possible. a20_nandread command is added for conveniance.
Afaik DMA controller doesn't vary between a10 and a20, so this should work on a10 devices, but due to no a10 hardware it wasn't tested.
First of all many thanks for working on this!
I see that this is against linux-sunxi/u-boot-sunxi.git, that is not (really) being actively developed anymore, see:
https://www.marshut.net/kuispp/linux-sunxi-u-boot-sunxi-is-no-longer-support...
Can you please rebase on top of upstream u-boot, on top of the next branch: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next ?
I'll certainly do. In fact, it's done, but not tested yet.
Your patches all seem to only have a subject and not a proper commit message, please add a commit message to all of them describing the changes in somewhat more detail then you do in the subject, you could e.g. re-use bits of your coverletter. Also for upstream u-boot we require a Signed-off-by as part of the commit message, so please add a line like this one to all your commit messages:
Signed-off-by: Daniel Kochmański <dkoch...@antmicro.com javascript:>
You can make git do this automatically by specifying the -s option when committing, also please send further versions of this patch-set to the upstream u-boot list.
Ok, thanks for the tip.
Besides all the above, which is easily fixed, unfortunately we also have the problem (or luxury) that Yassin Jaffer (added to the CC) has also been working on nand support, re-using the upstream kernel code Boris Brezillon has been working on, see:
I wasn't aware of his work untill now. I'll give a closer look to it.
Best Regards, Daniel
I believe that Yassin's version is better then yours because it re-uses existing nand infrastructure rather then more or less starting from scratch.
But Yassin has not been very active on this, and has been reluctant to submit his work upstream because it has not been tested a lot yet.
The best way forward may very well be you picking up Yassin's work (in coordination with Yassin), and submitting that to upstream u-boot.
Yassin, I notice that the patches in your git repo do not have a Signed-off-by, can you please reply to this mail with your Signed-off-by to indicate that it is ok for us to add your Signed-off-by to your patches ?
Regards,
Hans
Daniel Kochmański (4): sunxi: nand: add minimal nand driver (reading flash only). sunxi: mksunxiboot: increase block size, so checksum is correct on nand sunxi: nand: add nand spl boot option sunxi: nand: add a20_nandread command
arch/arm/cpu/armv7/sunxi/board.c | 4 + board/sunxi/Makefile | 1 + board/sunxi/nand.c | 252
+++++++++++++++++++++++++++++++++++++++
common/Makefile | 1 + common/cmd_a20_nandread.c | 26 ++++ common/spl/spl_nand.c | 23 +++- include/configs/sunxi-common.h | 25 ++-- tools/mksunxiboot.c | 2 +- 8 files changed, 322 insertions(+), 12 deletions(-) create mode 100644 board/sunxi/nand.c create mode 100644 common/cmd_a20_nandread.c
Hi,
On 07-01-15 08:15, Daniel Kochmański wrote:
Hi,
Hans de Goede writes:
Hi,
On 06-01-15 12:57, Yassin Jaffer wrote:
Hi Hans
I was working on utilizing the DM framework for the nand driver, but this is going to take sometime , anyway I will try to submit the driver based on bbrezillon work.
Ok, thanks. I'm looking forward to the posting of the next version of your series.
I'm quite happy if Daniel wants to pickup my work,
It looks to me like Daniel may have more time to work on this atm then you, so that might be a good idea. Either way please coordinate with each other to avoid double work. You can use private mail to Daniel for the coordination stuff if you prefer, it would be nice if you could send a short summary of who plans to do what once that is clear.
I'd be glad to, but I have limited time to devote for this project, and I might me not be as that helpful as you expect me to.
besides Daniel patches are for the SPL nand driver.
We should be able to use your adaption of Boris' work in the SPL too, for mmc we are also using the same driver in both SPL and regular u-boot, and I assume that Daniel's patches are also capable of loading say a kernel from nand, right Daniel ?
I'm not sure if I remember correctly, but I had problems with adaptation of lichee driver for nand, because it was too big to fit in SPL (BROM expects boot0 to fit at first 8K).
Yes, these patches are capable of this. I haven't included patch which added MBOOT functionality, since it seems unrelated to nand support. Putting it short, it was downloading multiimage from mtd2 (which had 16M in my config), and extracted from it kernel, devicetree, ramdisk and u-boot script which was performing boot.
Regards,
Hans
Regards
On Tuesday, January 6, 2015 12:08:48 AM UTC+11, Hans de Goede wrote:
Hi,
On 02-01-15 09:53, Daniel Kochmański wrote:
This is a series of patches to enable nand read functionality on sunxi devices. It uses DMA and is able to read from syndrome partitions and normal ones. Additionaly mksunxiboot tool is patched to be able to format resulting binary to be able to read from BROM, so NAND SPL builds are now possible. a20_nandread command is added for conveniance.
Afaik DMA controller doesn't vary between a10 and a20, so this should work on a10 devices, but due to no a10 hardware it wasn't tested.
First of all many thanks for working on this!
I see that this is against linux-sunxi/u-boot-sunxi.git, that is not (really) being actively developed anymore, see:
https://www.marshut.net/kuispp/linux-sunxi-u-boot-sunxi-is-no-longer-support...
Can you please rebase on top of upstream u-boot, on top of the next branch: http://git.denx.de/?p=u-boot/u-boot-sunxi.git;a=shortlog;h=refs/heads/next ?
I'll certainly do. In fact, it's done, but not tested yet.
Your patches all seem to only have a subject and not a proper commit message, please add a commit message to all of them describing the changes in somewhat more detail then you do in the subject, you could e.g. re-use bits of your coverletter. Also for upstream u-boot we require a Signed-off-by as part of the commit message, so please add a line like this one to all your commit messages:
Signed-off-by: Daniel Kochmański <dkoch...@antmicro.com javascript:>
You can make git do this automatically by specifying the -s option when committing, also please send further versions of this patch-set to the upstream u-boot list.
Ok, thanks for the tip.
Besides all the above, which is easily fixed, unfortunately we also have the problem (or luxury) that Yassin Jaffer (added to the CC) has also been working on nand support, re-using the upstream kernel code Boris Brezillon has been working on, see:
I wasn't aware of his work untill now. I'll give a closer look to it.
Great, I'm looking forward to seeing a new patchs-series from you, adding nand support to u-boot would be great.
Regards,
Hans
participants (3)
-
Daniel Kochmański -
Hans de Goede -
Yassin Jaffer