Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[kconfig] [ARMHF] [marvell armada cpu support] #172

Closed

Conversation

antony-rheneus
Copy link
Contributor

Deleted Arm related unused patches
kconfig_inclusions:
Enable Optoe
Gpio based I2c Mux
flash mtd parts support for flash partitions
kconfig_exclusions:
Removed uncommon soc support to reduce kernel size
Removed sound card
Reoved debug slab

Signed-off-by: Antony Rheneus arheneus@marvell.com

Deleted Arm related unused patches
kconfig_inclusions:
  Enable Optoe
  Gpio based I2c Mux
  flash mtd parts support for flash partitions
kconfig_exclusions:
    Removed uncommon soc support to reduce kernel size
    Removed sound card
    Reoved debug slab

Signed-off-by: Antony Rheneus <arheneus@marvell.com>
@paulmenzel
Copy link
Contributor

Please splite this up into one commit per list item in the merge/pull request description, and describe in the commit message, why that’s needed/useful.

@antony-rheneus
Copy link
Contributor Author

Please splite this up into one commit per list item in the merge/pull request description, and describe in the commit message, why that’s needed/useful.

@paulmenzel, Do you want each Kernel config should be raised as a separate PRs?

@paulmenzel
Copy link
Contributor

Please splite this up into one commit per list item in the merge/pull request description, and describe in the commit message, why that’s needed/useful.

@paulmenzel, Do you want each Kernel config should be raised as a separate PRs?

No, not for me. But one commit per change with a commit message would be nice.

@antony-rheneus
Copy link
Contributor Author

Please splite this up into one commit per list item in the merge/pull request description, and describe in the commit message, why that’s needed/useful.

@paulmenzel, Do you want each Kernel config should be raised as a separate PRs?

No, not for me. But one commit per change with a commit message would be nice.

Added commits as per your request,

@paulmenzel
Copy link
Contributor

Thank you for splitting this up. Please use git rebase to get rid of the first two commits, as the revert isn’t needed for review or the history.

Additionally, are you sure you do not break existing devices with these config changes? Keep in mind, the same Linux kernel image is used for all devices.

Lastly, for review and understanding your changes, commit messages are really necessary to explain the motivation and change. I believe the link I added in my first comment explains this well.

For example, the two commits:

  1. use gzip compression for uboot extraction
  2. kernel binary in gzipped for uboot to extract

It reads like they try to achive the same goal.

But, what is the current default, and what is the problem when loading the Linux kernel from U-Boot (what version)? Shouldn’t XZ be used, because the Linux kernel is smaller, and loading a smaller image is better than the added decompression time? Why Gzip? Big distributions moved to LZ4 or Zstd because that give the fastest loading speed. Without documenting your goal and reasoning, nobody knows why these changes are done.

@antony-rheneus
Copy link
Contributor Author

Thank you for the updates, but I am sorry, I am still not able to review this properly. The summary of the whole merge/pull request should be updated with goal, often the commits are missing the reason, so nobody knows, why that change is done, and some of the earlier comments were still not addressed.

For example, the whole series should be rebased to be easily reviewable, and to be able to have a booting platform with each commit. That means, for example, that commit Separate out SOC specific configuration based on platform architecture should be first in the series and only change the scripts.

Lastly, there are also typos in the commit message summaries. For example unsedunused.

@paulmenzel , I have rebased. Let me know if it is now easily reviewable?

@lguohan
Copy link
Contributor

lguohan commented Nov 30, 2020

in my opinion, the comment added for each commit is not really helpful. The comment needs explain the reason why, not what. People can look at the commit and understand what you did, but it is hard for people to figure out why. that is basically @paulmenzel 's point. I do not think rebase address his concern.

@antony-rheneus
Copy link
Contributor Author

in my opinion, the comment added for each commit is not really helpful. The comment needs explain the reason why, not what. People can look at the commit and understand what you did, but it is hard for people to figure out why. that is basically @paulmenzel 's point. I do not think rebase address his concern.

I have opened a new PR 176, please review it. As I could not rebase muliple times over this PR, as it doubled the commits from 24 to 48 and ended up in conflicts for each commit.
#176

@antony-rheneus
Copy link
Contributor Author

If you agree on review the PR 176, I will close this PR 172.

@paulmenzel
Copy link
Contributor

(You could have force pushed the rebased/new branch into this branch.)

The commits do not double when rebasing. What command did you use for rebasing? Maybe we can help you.

#176 is definitely an improvement. Thank you for the work. But it’s still lacking the git commit message summaries. Please read How to Write a Git Commit Message again. If there are open questions, please tell us. I’d really prefer to keep the discussion in this merge/pull request. But if you prefer #176, then we can move there too.

@antony-rheneus
Copy link
Contributor Author

in my opinion, the comment added for each commit is not really helpful. The comment needs explain the reason why, not what. People can look at the commit and understand what you did, but it is hard for people to figure out why. that is basically @paulmenzel 's point. I do not think rebase address his concern.

(You could have force pushed the rebased/new branch into this branch.)

The commits do not double when rebasing. What command did you use for rebasing? Maybe we can help you.

#176 is definitely an improvement. Thank you for the work. But it’s still lacking the git commit message summaries. Please read How to Write a Git Commit Message again. If there are open questions, please tell us. I’d really prefer to keep the discussion in this merge/pull request. But if you prefer #176, then we can move there too.

I used "git rebase --interactive HEAD~24", as totally I had 24 commits for each config.
If I were writing a feature, I could write a better summary or description. However Im just tapping out the existing config as per the default configuration required for a CPU Arch.

Basically the config present in "arch/arm/configs/mvebu_v7_defconfig" should get into marvell-armhf configuration. Which is not so when debian build generates, hence I had to port configs from the mvebu_v7_defconfig, else kernel would not boot at all.

Another example I can refer, OPTOE driver, which I want to enable or disable, I can only brief enabling OPTOE driver for SFF optics, rest woudl be just non-terse explanation from the OPTOE feature.

Better quote comment for my each commit description for what explanation you want to know specific to Armada A385 soc, so that I can get more or better explanation from soc team to add it.

@antony-rheneus
Copy link
Contributor Author

antony-rheneus commented Dec 1, 2020

@paulmenzel

For me problem is not the commits, but the conflict occurs even for "reword" commit description

Command used for git reabse:-
git rebase --interactive HEAD~24

Git log of 46 commits:-
`$ git logline -n 46

  • 8c21154 - (HEAD -> arm_update_master) Separate out SOC specific configuration based on platform architecture (7 hours ago)
  • f5272b5 - no debug slab (7 hours ago)
  • 0b57a01 - uefi based kernel module load security (7 hours ago)
  • 3f87c83 - remove unsed sensor and sound card (7 hours ago)
  • ae9b060 - no ntp external PPS needed (7 hours ago)
  • 2b5c141 - removed unsed old 3com as limited memory is available to load kernelfrom uboot (7 hours ago)
  • 06d94bf - no qemu firmware needed (7 hours ago)
  • 40ab251 - EFI is not used (7 hours ago)
  • f884729 - timer is set 100Hz (7 hours ago)
  • ef68190 - removed unsed arch as limited memory is available to load kernelfrom uboot (7 hours ago)
  • 43cb516 - removed unsed arch as limited memory is available to load kernelfrom uboot (7 hours ago)
  • d3957aa - slab allocator is used (7 hours ago)
  • fe5940f - tickless idle system (7 hours ago)
  • cf9735d - use gzip compression for uboot extraction (7 hours ago)
  • b2e4d9f - MTD parts for flash partition layout (7 hours ago)
  • f827261 - peripheral drm quirks, usb support, ioport map (7 hours ago)
  • 1e5f6b3 - Enable gpio i2c mux, opte and sff eeprom module (7 hours ago)
  • 5df7918 - timer to 100Hz (7 hours ago)
  • 926bb26 - default gpio pins to 0 (7 hours ago)
  • 349d089 - regular slab allocator (7 hours ago)
  • 4f0bbe4 - Disable heap randomization (7 hours ago)
  • 650d76d - Periodic timer ticks (7 hours ago)
  • 76f9556 - kernel binary in gzipped for uboot to extract (7 hours ago)
  • 95367d7 - [ARMHF] kernel config reverted to address comment to add per change per commit (7 hours ago)
  • d73c4cc - [kconfig] [ARMHF] [marvell armada cpu support] (7 hours ago)
  • e6ff106 - remove unused sensor and sound card (7 hours ago)
  • 0715da9 - no ntp external PPS needed (7 hours ago)
  • e176474 - removed unused old 3com as limited memory is available to load kernel from uboot (7 hours ago)
  • 2776504 - no qemu firmware needed (7 hours ago)
  • 8579128 - EFI based firmware boot is not enabled in armarda a385 (7 hours ago)
  • b477f3f - timer is set 100Hz for ticks in armada soc (7 hours ago)
  • 77c818a - removed unused arch as limited memory is available to load kernel from uboot (7 hours ago)
  • 812e1f4 - removed unused arch as limited memory is available to load kernel from uboot (7 hours ago)
  • c46084b - MTD parts for flash partition layout (7 hours ago)
  • 2156cb4 - peripheral drm quirks, usb support, ioport map (7 hours ago)
  • 77a0273 - Enable gpio i2c mux, opte and sff eeprom module (7 hours ago)
  • f715915 - timer set to 100Hz for armada soc (7 hours ago)
  • deb1c60 - default gpio pins to 0, number of gpio pins will be based out from DTB as per platform config (7 hours ago)
  • 5d5d609 - slab allocator is used and not SLUB as SLAB is well stable and tested for 32 bit arch (7 hours ago)
  • 4605677 - Use well stabled regular slab allocator instead of SLUB (7 hours ago)
  • 273a1ee - Disable heap randomization while calculating heap brk values to avoid any mem faults in 32 bit arch (7 hours ago)
  • ffa2f97 - Use HZ_PERIODIC Periodic timer ticks and not dynamic ticks for armada soc (7 hours ago)
  • da693fb - kernel binary in gzipped for uboot to extract (3 days ago)
  • 6a727bd - [kconfig] [ARMHF] [marvell armada cpu support] (3 days ago)
  • f4699e3 - Separate out SOC specific configuration based on platform architecture (3 days ago)
    `

@antony-rheneus
Copy link
Contributor Author

All the comments were addressed in PR #176

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants