rust/hg-core/src/revlog/nodemap_docket.rs
author Simon Sapin <simon.sapin@octobus.net>
Fri, 10 Sep 2021 09:53:09 +0200
changeset 47955 e834b79def74
parent 46630 842f2372ced6
child 47963 001d747c2baf
permissions -rw-r--r--
rust: Switch to the memmap2-rs crate https://github.com/RazrFalcon/memmap2-rs This is a fork of the original memmap crate which appears to be unmaintained: https://github.com/danburkert/memmap-rs/issues/90 This fork is the most popular according to https://crates.io/keywords/mmap Differential Revision: https://phab.mercurial-scm.org/D11397
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
     1
use crate::errors::{HgError, HgResultExt};
46630
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
     2
use crate::requirements;
46390
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
     3
use bytes_cast::{unaligned, BytesCast};
47955
e834b79def74 rust: Switch to the memmap2-rs crate
Simon Sapin <simon.sapin@octobus.net>
parents: 46630
diff changeset
     4
use memmap2::Mmap;
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
     5
use std::path::{Path, PathBuf};
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
     6
46167
8a4914397d02 rust: introduce Repo and Vfs types for filesystem abstraction
Simon Sapin <simon.sapin@octobus.net>
parents: 46090
diff changeset
     7
use super::revlog::RevlogError;
8a4914397d02 rust: introduce Repo and Vfs types for filesystem abstraction
Simon Sapin <simon.sapin@octobus.net>
parents: 46090
diff changeset
     8
use crate::repo::Repo;
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
     9
use crate::utils::strip_suffix;
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    10
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    11
const ONDISK_VERSION: u8 = 1;
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    12
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    13
pub(super) struct NodeMapDocket {
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    14
    pub data_length: usize,
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    15
    // TODO: keep here more of the data from `parse()` when we need it
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    16
}
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    17
46390
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    18
#[derive(BytesCast)]
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    19
#[repr(C)]
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    20
struct DocketHeader {
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    21
    uid_size: u8,
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    22
    _tip_rev: unaligned::U64Be,
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    23
    data_length: unaligned::U64Be,
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    24
    _data_unused: unaligned::U64Be,
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    25
    tip_node_size: unaligned::U64Be,
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    26
}
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    27
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    28
impl NodeMapDocket {
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    29
    /// Return `Ok(None)` when the caller should proceed without a persistent
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    30
    /// nodemap:
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    31
    ///
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    32
    /// * This revlog does not have a `.n` docket file (it is not generated for
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    33
    ///   small revlogs), or
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    34
    /// * The docket has an unsupported version number (repositories created by
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    35
    ///   later hg, maybe that should be a requirement instead?), or
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    36
    /// * The docket file points to a missing (likely deleted) data file (this
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    37
    ///   can happen in a rare race condition).
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    38
    pub fn read_from_file(
46167
8a4914397d02 rust: introduce Repo and Vfs types for filesystem abstraction
Simon Sapin <simon.sapin@octobus.net>
parents: 46090
diff changeset
    39
        repo: &Repo,
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    40
        index_path: &Path,
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    41
    ) -> Result<Option<(Self, Mmap)>, RevlogError> {
46630
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    42
        if !repo
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    43
            .requirements()
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    44
            .contains(requirements::NODEMAP_REQUIREMENT)
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    45
        {
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    46
            // If .hg/requires does not opt it, don’t try to open a nodemap
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    47
            return Ok(None);
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    48
        }
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
    49
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    50
        let docket_path = index_path.with_extension("n");
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    51
        let docket_bytes = if let Some(bytes) =
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    52
            repo.store_vfs().read(&docket_path).io_not_found_as_none()?
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    53
        {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    54
            bytes
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    55
        } else {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    56
            return Ok(None);
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    57
        };
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    58
46390
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    59
        let input = if let Some((&ONDISK_VERSION, rest)) =
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    60
            docket_bytes.split_first()
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    61
        {
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    62
            rest
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    63
        } else {
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    64
            return Ok(None);
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    65
        };
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    66
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    67
        /// Treat any error as a parse error
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    68
        fn parse<T, E>(result: Result<T, E>) -> Result<T, RevlogError> {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    69
            result.map_err(|_| {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    70
                HgError::corrupted("nodemap docket parse error").into()
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    71
            })
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    72
        }
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    73
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    74
        let (header, rest) = parse(DocketHeader::from_bytes(input))?;
46390
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    75
        let uid_size = header.uid_size as usize;
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    76
        // TODO: do we care about overflow for 4 GB+ nodemap files on 32-bit
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    77
        // systems?
46390
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    78
        let tip_node_size = header.tip_node_size.get() as usize;
0800aa42bb4c rust: use the bytes-cast crate to parse persistent nodemaps
Simon Sapin <simon.sapin@octobus.net>
parents: 46167
diff changeset
    79
        let data_length = header.data_length.get() as usize;
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    80
        let (uid, rest) = parse(u8::slice_from_bytes(rest, uid_size))?;
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    81
        let (_tip_node, _rest) =
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    82
            parse(u8::slice_from_bytes(rest, tip_node_size))?;
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    83
        let uid = parse(std::str::from_utf8(uid))?;
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    84
        let docket = NodeMapDocket { data_length };
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    85
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    86
        let data_path = rawdata_path(&docket_path, uid);
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    87
        // TODO: use `vfs.read()` here when the `persistent-nodemap.mmap`
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    88
        // config is false?
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    89
        if let Some(mmap) = repo
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    90
            .store_vfs()
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    91
            .mmap_open(&data_path)
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    92
            .io_not_found_as_none()?
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    93
        {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    94
            if mmap.len() >= data_length {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    95
                Ok(Some((docket, mmap)))
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    96
            } else {
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    97
                Err(HgError::corrupted("persistent nodemap too short").into())
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
    98
            }
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
    99
        } else {
46630
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
   100
            // Even if .hg/requires opted in, some revlogs are deemed small
842f2372ced6 rhg: Don’t attempt to read persistent nodemap without .hg/requires opt-in
Simon Sapin <simon.sapin@octobus.net>
parents: 46443
diff changeset
   101
            // enough to not need a persistent nodemap.
46443
43d63979a75e rust: use HgError in RevlogError and Vfs
Simon Sapin <simon.sapin@octobus.net>
parents: 46390
diff changeset
   102
            Ok(None)
46090
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   103
        }
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   104
    }
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   105
}
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   106
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   107
fn rawdata_path(docket_path: &Path, uid: &str) -> PathBuf {
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   108
    let docket_name = docket_path
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   109
        .file_name()
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   110
        .expect("expected a base name")
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   111
        .to_str()
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   112
        .expect("expected an ASCII file name in the store");
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   113
    let prefix = strip_suffix(docket_name, ".n.a")
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   114
        .or_else(|| strip_suffix(docket_name, ".n"))
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   115
        .expect("expected docket path in .n or .n.a");
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   116
    let name = format!("{}-{}.nd", prefix, uid);
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   117
    docket_path
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   118
        .parent()
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   119
        .expect("expected a non-root path")
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   120
        .join(name)
9eb07ab3f2d4 rhg: use persistent nodemap when available
Simon Sapin <simon-commits@exyr.org>
parents:
diff changeset
   121
}