Switch programs activation to whole-set based gating #11736

ryoqun · 2020-08-20T15:05:49Z

Problem

get_entered_epoch_callback isn't called when restoring from snapshots, which is too confusing and error-prone.
Also, we can't simply call it immediately after snapshot restoration because it expects to be called exactly once at each epoch boundary

get_programs and get_builtins returns delta set. i.e. add these new additions of programs to the current available set at the given epoch. This works nicely in the ideal world, where we're running the validator since genesis without ever restarting a perfect bug-free validator.

In reality, we must rely on snapshots 99.999% of time. When restoring from snapshots, the delta set doesn't work quite: we don't persist the current available set (namely bank.message_processor is effectively serde(skip)).

Summary of Changes

So, just reflect the reality by making these functions snapshot-friendly by returning whole-set of available programs at the given epoch. And make it callable from finish_init(), which is called after snapshot restoration.

Also, fix a bunch of other dangerous code along the way.

Also, this is intended to be back-port friendly; so the fix is intentionally not exhaustive. Still, get_entered_epoch_callback is a bit error-prone. Specifically, it must be idempotent. (We could solve this by artificially introducing some intermediate struct like ScheduledBankFeatures or the like instead of mind-opening way of passing &mut Bank).

Fixes #

ryoqun · 2020-08-20T15:08:18Z

genesis-programs/src/lib.rs

                // The epoch of std::u64::MAX is a placeholder and is expected
                // to be reduced in a future network update.
-                Some(vec![
+                programs.extend(vec![
                    Program::BuiltinLoader(solana_bpf_loader_program!()),
                    Program::Native(solana_vest_program!()),


According to this code, mainnet-beta has yet to enable smart-contracts... ;)

Yeah, @mvines indicated that it was enabled on a branch at some point for mainnet and therefore now part of the ledger

ryoqun · 2020-08-20T15:09:13Z

runtime/src/bank.rs

@@ -584,6 +574,7 @@ impl Bank {
        if !new.fix_recent_blockhashes_sysvar_delay() {
            new.update_recent_blockhashes();
        }
+        dbg!(&new.message_processor);


t-nelson · 2020-08-20T15:22:22Z

runtime/src/bank.rs

@@ -3178,6 +3165,20 @@ impl Bank {
        consumed_budget.saturating_sub(budget_recovery_delta)
    }

+    // This is called from snapshot restore and for each epoch boundary
+    // The entire code path herein must be idempotent
+    pub fn refresh_programs_and_inflation(&mut self) {


NIT: I have a feeling this name won't age well. Maybe something like apply_feature_activations() ?

ryoqun · 2020-08-20T15:43:06Z

out of scope of this pr todos:

Make bank.message_processor serde(skip), not individual fields. (Well, this is already addressed by Fold bank serialisation into serde snapshot #10581).
assert!s bank.message_processor after snapshot retore.

jackcmay · 2020-08-20T15:44:51Z

genesis-programs/src/lib.rs

-                Some(vec![Program::BuiltinLoader(solana_bpf_loader_program!())])
-            } else {
-                None
+                programs.extend(vec![Program::BuiltinLoader(solana_bpf_loader_program!())]);


Is this added to the list twice?

jackcmay · 2020-08-20T15:50:18Z

genesis-programs/src/lib.rs

-        OperatingMode::Preview => {
-            if epoch == std::u64::MAX {
+        OperatingMode::Stable => {
+            // at which epoch, bpf_loader_program is enabled??


@mvines Do you know at what epoch the bpf loader was added?

Also, @ryoqun At whatever epoch that is ^^ we should be adding solana_bpf_loader_deprecated_program because on tip that is the original loader.

Looks like the original loader was enabled in epoch 34

solana/genesis-programs/src/lib.rs

Line 74 in 86419df

if epoch == 34 {

@mvines Does this jive with what you would expect?

@ryoqun I pushed changes to this to solana/fragile-programs-gating

jackcmay · 2020-08-20T15:52:07Z

genesis-programs/src/lib.rs

+    native_programs
+}
+
+fn recheck_cross_program_support(bank: &mut Bank) {


Why not leave this logic in the bank?

jackcmay · 2020-08-20T16:06:04Z

@ryoqun I'm not too familiar with the snapshot test coverage, can we add tests to recreate the scenario we ran into last night?

ryoqun · 2020-08-20T16:19:35Z

Make bank.message_processor serde(skip), not individual fields. (Well, this is already addressed by Fold bank serialisation into serde snapshot #10581).

Specifically, newer snapshot restoration code is now using MessageProcessor::default() here: https://github.com/solana-labs/solana/blob/master/runtime/src/serde_snapshot/future.rs#L188

jackcmay · 2020-08-20T20:00:21Z

runtime/src/bank.rs

+        }
+
+        for program in get_builtins(self.operating_mode(), self.epoch()) {
+            self.add_builtin(&program.name, program.id, program.entrypoint);


Adding the same programs at every epoch is going to fail because the account already exists. Looks like the bank hash ci failure is related, looking into it.

jackcmay · 2020-08-20T21:36:02Z

@ryoqun Closing to switch to the branch in the solana repo:

#11750

ryoqun · 2020-08-28T01:07:26Z

genesis-programs/src/lib.rs

-        if OperatingMode::Stable == operating_mode {
-            bank.set_cross_program_support(bank.epoch() >= 63);
-        } else {
-            bank.set_cross_program_support(true);
-        }


ryoqun added 2 commits August 20, 2020 23:45

Implement Debug for MessageProcessor

833dfb8

Switch from delta-based gating to whole-set gating

f096ffb

ryoqun commented Aug 20, 2020

View reviewed changes

Remove dbg!

daeec6e

t-nelson reviewed Aug 20, 2020

View reviewed changes

ryoqun added 2 commits August 21, 2020 00:25

Fix clippy

d087942

Clippy

5eae5f7

jackcmay reviewed Aug 20, 2020

View reviewed changes

Add test

5a94742

jackcmay reviewed Aug 20, 2020

View reviewed changes

jackcmay closed this Aug 20, 2020

jackcmay mentioned this pull request Aug 25, 2020

Switch programs activation to whole-set based gating #11750

Merged

ryoqun commented Aug 28, 2020

View reviewed changes

ryoqun mentioned this pull request Aug 31, 2020

Add tests for the Debug and activation Vecs #11926

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Switch programs activation to whole-set based gating #11736

Switch programs activation to whole-set based gating #11736

ryoqun commented Aug 20, 2020

ryoqun Aug 20, 2020

jackcmay Aug 20, 2020

ryoqun Aug 20, 2020

t-nelson Aug 20, 2020

ryoqun commented Aug 20, 2020 •

edited

Loading

jackcmay Aug 20, 2020

jackcmay Aug 20, 2020

jackcmay Aug 20, 2020

jackcmay Aug 20, 2020

jackcmay Aug 20, 2020

jackcmay commented Aug 20, 2020

ryoqun commented Aug 20, 2020

jackcmay Aug 20, 2020

jackcmay commented Aug 20, 2020

ryoqun Aug 28, 2020

Switch programs activation to whole-set based gating #11736

Switch programs activation to whole-set based gating #11736

Conversation

ryoqun commented Aug 20, 2020

Problem

Summary of Changes

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ryoqun commented Aug 20, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jackcmay commented Aug 20, 2020

ryoqun commented Aug 20, 2020

Choose a reason for hiding this comment

jackcmay commented Aug 20, 2020

Choose a reason for hiding this comment

ryoqun commented Aug 20, 2020 •

edited

Loading