[thin_check (rust)] Keep track of data block ref counts

as we walk the mapping tree.
2020-08-08 09:54:16 +01:00 · 2020-08-08 09:54:16 +01:00 · ec8f7b7fa8
parent 8f76371bb2
commit ec8f7b7fa8
3 changed files with 206 additions and 28 deletions
--- a/Cargo.lock
+++ b/Cargo.lock
@ -85,9 +85,9 @@ checksum = "4785bdd1c96b2a846b2bd7cc02e86b6b3dbf14e7e53446c4f54c92a361040822"

 [[package]]
 name = "clap"
-version = "2.33.1"
+version = "2.33.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bdfa80d47f954d53a35a64987ca1422f495b8d6483c0fe9f7117b36c2a792129"
+checksum = "10040cdf04294b565d9e0319955430099ec3813a64c952b86a41200ad714ae48"
 dependencies = [
 "ansi_term",
 "atty",
@ -305,9 +305,9 @@ dependencies = [

 [[package]]
 name = "libc"
-version = "0.2.73"
+version = "0.2.74"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bd7d4bd64732af4bf3a67f367c27df8520ad7e230c5817b8ff485864d80242b9"
+checksum = "a2f02823cf78b754822df5f7f268fb59822e7296276d3e069d8e8cb26a14bd10"

 [[package]]
 name = "log"
@ -359,9 +359,9 @@ dependencies = [

 [[package]]
 name = "num-derive"
-version = "0.3.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0c8b15b261814f992e33760b1fca9fe8b693d8a65299f20c9901688636cfb746"
+checksum = "e0396233fb2d5b0ae3f05ff6aba9a09185f7f6e70f87fb01147d545f85364665"
 dependencies = [
 "proc-macro2",
 "quote",
@ -607,9 +607,9 @@ checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"

 [[package]]
 name = "syn"
-version = "1.0.36"
+version = "1.0.38"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4cdb98bcb1f9d81d07b536179c269ea15999b5d14ea958196413869445bb5250"
+checksum = "e69abc24912995b3038597a7a593be5053eb0fb44f3cc5beec0deb421790c1f4"
 dependencies = [
 "proc-macro2",
 "quote",
--- a/src/pdata/space_map.rs
+++ b/src/pdata/space_map.rs
@ -1,5 +1,7 @@
 use anyhow::{anyhow, Result};
+use fixedbitset::FixedBitSet;
 use nom::{number::complete::*, IResult};
+use std::collections::HashMap;

 use crate::block_manager::*;
 use crate::pdata::btree::Unpack;
@ -16,9 +18,7 @@ pub struct SMRoot {

 pub fn unpack_root(data: &[u8]) -> Result<SMRoot> {
    match SMRoot::unpack(data) {
-        Err(_e) => {
-            Err(anyhow!("couldn't parse SMRoot"))
-        },
+        Err(_e) => Err(anyhow!("couldn't parse SMRoot")),
        Ok((_i, v)) => Ok(v),
    }
 }
@ -27,19 +27,22 @@ impl Unpack for SMRoot {
    fn disk_size() -> u32 {
        32
    }
-    
+
    fn unpack(data: &[u8]) -> IResult<&[u8], SMRoot> {
        let (i, nr_blocks) = le_u64(data)?;
        let (i, nr_allocated) = le_u64(i)?;
        let (i, bitmap_root) = le_u64(i)?;
        let (i, ref_count_root) = le_u64(i)?;

-        Ok  ((i, SMRoot {
-            nr_blocks,
-            nr_allocated,
-            bitmap_root,
-            ref_count_root,
-        }))
+        Ok((
+            i,
+            SMRoot {
+                nr_blocks,
+                nr_allocated,
+                bitmap_root,
+                ref_count_root,
+            },
+        ))
    }
 }

@ -62,7 +65,14 @@ impl Unpack for IndexEntry {
        let (i, nr_free) = le_u32(i)?;
        let (i, none_free_before) = le_u32(i)?;

-        Ok((i, IndexEntry {blocknr, nr_free, none_free_before}))
+        Ok((
+            i,
+            IndexEntry {
+                blocknr,
+                nr_free,
+                none_free_before,
+            },
+        ))
    }
 }

@ -85,11 +95,18 @@ impl Unpack for BitmapHeader {
        let (i, not_used) = le_u32(i)?;
        let (i, blocknr) = le_u64(i)?;

-        Ok((i, BitmapHeader {csum, not_used, blocknr} ))
+        Ok((
+            i,
+            BitmapHeader {
+                csum,
+                not_used,
+                blocknr,
+            },
+        ))
    }
 }

-#[derive(Debug)]
+#[derive(Clone, Debug, PartialEq)]
 pub enum BitmapEntry {
    Small(u8),
    Overflow,
@ -109,7 +126,7 @@ impl Unpack for Bitmap {
    fn unpack(data: &[u8]) -> IResult<&[u8], Self> {
        let (mut i, header) = BitmapHeader::unpack(data)?;

-	let mut entries = Vec::new();
+        let mut entries = Vec::new();
        let nr_words = (BLOCK_SIZE - BitmapHeader::disk_size() as usize) / 8;
        for _w in 0..nr_words {
            let (tmp, mut word) = le_u64(i)?;
@ -128,7 +145,92 @@ impl Unpack for Bitmap {
            i = tmp;
        }

-        Ok((i, Bitmap {header, entries}))
+        Ok((i, Bitmap { header, entries }))
+    }
+}
+
+//------------------------------------------
+
+pub struct CoreSpaceMap {
+    nr_entries: u64,
+    bits: FixedBitSet,
+    overflow: HashMap<u64, u32>,
+}
+
+impl CoreSpaceMap {
+    pub fn new(nr_entries: u64) -> CoreSpaceMap {
+        let bits = FixedBitSet::with_capacity(nr_entries as usize * 2);
+        CoreSpaceMap {
+            nr_entries,
+            bits,
+            overflow: HashMap::new(),
+        }
+    }
+
+    fn get_bits(&self, b: u64) -> Result<u32> {
+        if b >= self.nr_entries {
+            return Err(anyhow!("space map index out of bounds"));
+        }
+
+        let low_bit = self.bits.contains((b * 2) as usize);
+        let high_bit = self.bits.contains((b * 2 + 1) as usize);
+        let mut result = 0u32;
+
+        if high_bit {
+            result += 2;
+        }
+        if low_bit {
+            result += 1;
+        }
+
+        Ok(result)
+    }
+
+    fn set_bits(&mut self, b: u64, count: u32) -> Result<()> {
+        if count > 3 {
+            return Err(anyhow!("internal error: bits entry should be <= 3"));
+        }
+
+        self.bits.set((b * 2) as usize, (count & 0x1) != 0);
+        self.bits.set((b * 2 + 1) as usize, (count & 0x2) != 0);
+        Ok(())
+    }
+
+    pub fn get(&self, b: u64) -> Result<u32> {
+        let result = self.get_bits(b)?;
+        if result < 3 {
+            Ok(result)
+        } else {
+            match self.overflow.get(&b) {
+                None => Err(anyhow!(
+                    "internal error: missing overflow entry in space map"
+                )),
+                Some(result) => Ok(*result),
+            }
+        }
+    }
+
+    pub fn inc(&mut self, b: u64) -> Result<()> {
+        let old = self.get(b)?;
+
+        if old < 3 {
+            // bump up the bits
+            self.set_bits(b, old + 1)?;
+
+            if old == 2 {
+                // insert overflow entry
+                self.overflow.insert(b, 1);
+            }
+        } else if let Some(count) = self.overflow.get_mut(&b) {
+            // increment the overflow
+            *count += 1;
+        } else {
+            return Err(anyhow!(
+                "internal error: missing overflow entry in space map"
+            ));
+        }
+
+        Ok(())
    }
 }

--- a/src/thin/check.rs
+++ b/src/thin/check.rs
@ -1,7 +1,7 @@
 use anyhow::{anyhow, Result};
 use fixedbitset::FixedBitSet;
 use nom::{number::complete::*, IResult};
-use std::collections::HashMap;
+use std::collections::{HashMap, BTreeMap};
 use std::path::Path;
 use std::sync::{Arc, Mutex};
 use std::time::Instant;
@ -66,10 +66,21 @@ impl Unpack for BlockTime {
    }
 }

-struct BottomLevelVisitor {}
+struct BottomLevelVisitor {
+    data_sm: Arc<Mutex<CoreSpaceMap>>,
+}

 impl NodeVisitor<BlockTime> for BottomLevelVisitor {
-    fn visit(&mut self, _w: &BTreeWalker, _b: &Block, _node: &Node<BlockTime>) -> Result<()> {
+    fn visit(&mut self, _w: &BTreeWalker, _b: &Block, node: &Node<BlockTime>) -> Result<()> {
+        // FIXME: do other checks
+
+        if let Node::Leaf {header: _h, keys: _k, values} = node {
+            let mut data_sm = self.data_sm.lock().unwrap();
+            for bt in values {
+               data_sm.inc(bt.block)?;
+            }
+        }
+
        Ok(())
    }
 }
@ -197,6 +208,46 @@ impl<V: Unpack + Clone> NodeVisitor<V> for ValueCollector<V> {

 //------------------------------------------

+struct RangeBuilder {
+    run: Option<(u64, BitmapEntry)>,
+    runs: Vec<(u64, BitmapEntry)>
+}
+
+impl RangeBuilder {
+    fn new() -> RangeBuilder {
+        RangeBuilder {
+            run: None,
+            runs: Vec::new(),
+        }
+    }
+
+    fn push(&mut self, e: &BitmapEntry) {
+        match &self.run {
+            Some((len, e2)) if *e == *e2 => {
+                self.run = Some((*len + 1, e2.clone()));
+            },
+            Some((len, e2)) => {
+                self.runs.push((*len, e2.clone()));
+                self.run = Some((1, e.clone()));
+            },
+            None => {
+                self.run = Some((1, e.clone()));
+            }
+        }
+    }
+
+    fn complete(&mut self) {
+        match &self.run {
+            Some((len, e)) => {
+                self.runs.push((*len, e.clone()));
+            },
+            None => {}
+        }
+    }
+}
+
+//------------------------------------------
+
 pub fn check(dev: &Path) -> Result<()> {
    let engine = Arc::new(AsyncIoEngine::new(dev, 256)?);

@ -230,10 +281,14 @@ pub fn check(dev: &Path) -> Result<()> {
            engine.get_nr_blocks() as usize,
        )));

+        let root = unpack::<SMRoot>(&sb.data_sm_root[0..])?;
+        let data_sm = Arc::new(Mutex::new(CoreSpaceMap::new(root.nr_blocks)));
+
        for (thin_id, root) in roots {
            let mut w = BTreeWalker::new_with_seen(engine.clone(), seen.clone(), false);
+            let data_sm = data_sm.clone();
            pool.execute(move || {
-                let mut v = BottomLevelVisitor {};
+                let mut v = BottomLevelVisitor {data_sm};
                let result = w.walk(&mut v, root).expect("walk failed"); // FIXME: return error
                eprintln!("checked thin_dev {} -> {:?}", thin_id, result);
            });
@ -266,6 +321,8 @@ pub fn check(dev: &Path) -> Result<()> {
        let _result = w.walk(&mut v, root.bitmap_root);
        eprintln!("{} index entries", v.entries.len());

+        let mut builder = RangeBuilder::new();
+
        for i in v.entries {
            let mut b = Block::new(i.blocknr);
            engine.read(&mut b)?;
@ -274,7 +331,26 @@ pub fn check(dev: &Path) -> Result<()> {
                return Err(anyhow!("Index entry points to block ({}) that isn't a bitmap", b.loc));
            }

-            let _bitmap = unpack::<Bitmap>(b.get_data())?;
+            let bitmap = unpack::<Bitmap>(b.get_data())?;
+            for e in bitmap.entries {
+                builder.push(&e);
+            }
+        }
+        builder.complete();
+        eprintln!("{} ranges", builder.runs.len());
+
+        let mut counts = BTreeMap::new();
+        for (len, _v) in builder.runs {
+            if let Some(c) = counts.get(&len) {
+                let new_c = *c + 1;
+                counts.insert(len, new_c);
+            } else {
+                counts.insert(len, 1);
+            }
+        }
+
+        for (len, c) in counts {
+            eprintln!("{}: {}", len, c);
        }
    }