[Fix|NominationPools] Only allow apply slash to be executed if the slash amount is atleast ED

[Ank4n]

This change prevents pools::apply_slash from being executed when the pending slash amount of the member is lower than the ED.

The issue came to light with the failing benchmark test in Kusama. The problem arises from the inexact conversion between points and balance. Specifically, when points are converted to balance and then back to points, rounding can introduce a small discrepancy between the input and the resulting value. This issue surfaced in Kusama due to its ED being different from Westend and Polkadot (1 UNIT/300), making the rounding issue noticeable.

This fix is also significant because applying a slash is feeless and permissionless. Allowing super small slash amounts to be applied without a fee is undesirable. With this change, such small slashes will still be applied but only when member funds are withdrawn.

[rockbmb]

Why do we want to to avoid slashing below the ED here (pub fn apply_slash)?

[gui1117]

This fix is also significant because applying a slash is feeless and permissionless. Allowing super small slash amounts to be applied without a fee is undesirable. With this change, such small slashes will still be applied but only when member funds are withdrawn.

I don't think this is an issue, slash are applied for rare circumstances no?

[rockbmb]

My previous comment only made sense when taken with another made to this line, which asked: why don't we want to enforce the ED checks here?

Sorry, I had the GH open for a while, it might have been lost when I refreshed the browser.

[Ank4n]

At this point, a gas fee is charged for slashing. Additionally, we need to clean up all bookkeeping before a member withdraws.

[Ank4n]

This fix is also significant because applying a slash is feeless and permissionless. Allowing super small slash amounts to be applied without a fee is undesirable. With this change, such small slashes will still be applied but only when member funds are withdrawn.

I don't think this is an issue, slash are applied for rare circumstances no?

Single slash can affect unbounded number of pool members. The slashes are applied per member and if that is too small, is it worth to provide free compute for it?

[gpestana]

nit but in order to prevent spreading do_apply_slash(..., bool) across the code which is not very readable, how about creating an enum that helps with readability? something along the lines of

enum SlashConfig {
 EnforceMinimum,
 DoNotEnforceMinimuim
}

Not sure about the enum and variant names, but in general

// reads better
do_apply_slash(..., SlashConfig::EnforceMinimum);

// than this
do_apply_slash(..., true);

[paritytech-cmd-bot-polkadot-sdk]

Created backport PR for stable2407:

• [stable2407] Backport #6540 #6591 with remaining conflicts!

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin backport-6540-to-stable2407
git worktree add --checkout .worktree/backport-6540-to-stable2407 backport-6540-to-stable2407
cd .worktree/backport-6540-to-stable2407
git reset --hard HEAD^
git cherry-pick -x bf20a9ee18f7215210bbbabf79e955c8c35b3360
git push --force-with-lease

[paritytech-cmd-bot-polkadot-sdk]

Created backport PR for stable2409:

• [stable2409] Backport #6540 #6592 with remaining conflicts!

Please cherry-pick the changes locally and resolve any conflicts.

git fetch origin backport-6540-to-stable2409
git worktree add --checkout .worktree/backport-6540-to-stable2409 backport-6540-to-stable2409
cd .worktree/backport-6540-to-stable2409
git reset --hard HEAD^
git cherry-pick -x bf20a9ee18f7215210bbbabf79e955c8c35b3360
git push --force-with-lease