[pbs-devel] [RFC v2 proxmox-backup 12/23] fix #3174: archiver/extractor: impl appendix ref

[Christian Ebner <c.ebner@proxmox.com>]

Implements the functionality to create and extract appendix references
via the pbs client.

This reuses the pxar encoders functionality to write appendix reference
entries and adds the implementation to store and access them in the
catalog.

Signed-off-by: Christian Ebner <c.ebner@proxmox.com>
---
Changes since version 1:
- Use custom types for appendix ref archive offsets
- Include catalog implementation in this patch

 pbs-client/src/catalog_shell.rs |   2 +-
 pbs-client/src/pxar/create.rs   |  12 ++++
 pbs-client/src/pxar/extract.rs  |  16 +++++
 pbs-client/src/pxar/tools.rs    |   8 +++
 pbs-datastore/src/catalog.rs    | 103 ++++++++++++++++++++++++++++++++
 5 files changed, 140 insertions(+), 1 deletion(-)

diff --git a/pbs-client/src/catalog_shell.rs b/pbs-client/src/catalog_shell.rs
index f53b3cc5..99416d2f 100644
--- a/pbs-client/src/catalog_shell.rs
+++ b/pbs-client/src/catalog_shell.rs
@@ -1147,7 +1147,7 @@ impl<'a> ExtractorState<'a> {
             (_, DirEntryAttribute::Directory { .. }) => {
                 self.handle_new_directory(entry, match_result?).await?;
             }
-            (true, DirEntryAttribute::File { .. }) => {
+            (true, DirEntryAttribute::File { .. } | DirEntryAttribute::AppendixRef { .. }) => {
                 self.dir_stack.push(PathStackEntry::new(entry));
                 let file = Shell::walk_pxar_archive(self.accessor, &mut self.dir_stack).await?;
                 self.extract_file(file).await?;
diff --git a/pbs-client/src/pxar/create.rs b/pbs-client/src/pxar/create.rs
index a2338218..611d7421 100644
--- a/pbs-client/src/pxar/create.rs
+++ b/pbs-client/src/pxar/create.rs
@@ -735,6 +735,18 @@ impl Archiver {
         Ok(out.file_offset())
     }
 
+    async fn add_appendix_ref<T: SeqWrite + Send>(
+        &mut self,
+        encoder: &mut Encoder<'_, T>,
+        file_name: &Path,
+        appendix_offset: pxar::encoder::AppendixRefOffset,
+        file_size: u64,
+    ) -> Result<(), Error> {
+        Ok(encoder
+            .add_appendix_ref(file_name, appendix_offset, file_size)
+            .await?)
+    }
+
     async fn add_symlink<T: SeqWrite + Send>(
         &mut self,
         encoder: &mut Encoder<'_, T>,
diff --git a/pbs-client/src/pxar/extract.rs b/pbs-client/src/pxar/extract.rs
index f78e06c2..d2d42749 100644
--- a/pbs-client/src/pxar/extract.rs
+++ b/pbs-client/src/pxar/extract.rs
@@ -74,6 +74,7 @@ struct ExtractorIterState {
     err_path_stack: Vec<OsString>,
     current_match: bool,
     end_reached: bool,
+    appendix_list: Vec<(PathBuf, u64, u64)>,
 }
 
 /// An [`Iterator`] that encapsulates the process of extraction in [extract_archive].
@@ -98,6 +99,7 @@ impl ExtractorIterState {
             err_path_stack: Vec::new(),
             current_match: options.extract_match_default,
             end_reached: false,
+            appendix_list: Vec::new(),
         }
     }
 }
@@ -373,6 +375,20 @@ where
                 }
                 .context(PxarExtractContext::ExtractFile)
             }
+            (
+                true,
+                EntryKind::AppendixRef {
+                    appendix_offset,
+                    file_size,
+                },
+            ) => {
+                self.state.appendix_list.push((
+                    entry.path().to_path_buf(),
+                    *appendix_offset,
+                    *file_size,
+                ));
+                Ok(())
+            }
             (false, _) => Ok(()), // skip this
         };
 
diff --git a/pbs-client/src/pxar/tools.rs b/pbs-client/src/pxar/tools.rs
index 0cfbaf5b..aac5a1e7 100644
--- a/pbs-client/src/pxar/tools.rs
+++ b/pbs-client/src/pxar/tools.rs
@@ -156,6 +156,14 @@ pub fn format_multi_line_entry(entry: &Entry) -> String {
 
     let (size, link, type_name) = match entry.kind() {
         EntryKind::File { size, .. } => (format!("{}", *size), String::new(), "file"),
+        EntryKind::AppendixRef {
+            appendix_offset,
+            file_size,
+        } => (
+            format!("{} {}", appendix_offset, file_size),
+            String::new(),
+            "appendix ref",
+        ),
         EntryKind::Symlink(link) => (
             "0".to_string(),
             format!(" -> {:?}", link.as_os_str()),
diff --git a/pbs-datastore/src/catalog.rs b/pbs-datastore/src/catalog.rs
index c4d1a4de..da68dac9 100644
--- a/pbs-datastore/src/catalog.rs
+++ b/pbs-datastore/src/catalog.rs
@@ -28,6 +28,14 @@ pub trait BackupCatalogWriter {
         ctime: i64,
         file_offset: pxar::encoder::LinkOffset,
     ) -> Result<(), Error>;
+    fn add_appendix_ref(
+        &mut self,
+        name: &CStr,
+        size: u64,
+        mtime: i64,
+        ctime: i64,
+        appendix_ref_offset: pxar::encoder::AppendixRefOffset,
+    ) -> Result<(), Error>;
     fn add_symlink(&mut self, name: &CStr) -> Result<(), Error>;
     fn add_hardlink(&mut self, name: &CStr) -> Result<(), Error>;
     fn add_block_device(&mut self, name: &CStr) -> Result<(), Error>;
@@ -41,6 +49,7 @@ pub trait BackupCatalogWriter {
 pub enum CatalogEntryType {
     Directory = b'd',
     File = b'f',
+    AppendixRef = b'r',
     Symlink = b'l',
     Hardlink = b'h',
     BlockDevice = b'b',
@@ -56,6 +65,7 @@ impl TryFrom<u8> for CatalogEntryType {
         Ok(match value {
             b'd' => CatalogEntryType::Directory,
             b'f' => CatalogEntryType::File,
+            b'r' => CatalogEntryType::AppendixRef,
             b'l' => CatalogEntryType::Symlink,
             b'h' => CatalogEntryType::Hardlink,
             b'b' => CatalogEntryType::BlockDevice,
@@ -72,6 +82,7 @@ impl From<&DirEntryAttribute> for CatalogEntryType {
         match value {
             DirEntryAttribute::Directory { .. } => CatalogEntryType::Directory,
             DirEntryAttribute::File { .. } => CatalogEntryType::File,
+            DirEntryAttribute::AppendixRef { .. } => CatalogEntryType::AppendixRef,
             DirEntryAttribute::Symlink => CatalogEntryType::Symlink,
             DirEntryAttribute::Hardlink => CatalogEntryType::Hardlink,
             DirEntryAttribute::BlockDevice => CatalogEntryType::BlockDevice,
@@ -98,9 +109,22 @@ impl FileOffset {
         self.offset
     }
 }
+
+#[derive(Clone, Copy, Debug, Eq, PartialEq, Ord, PartialOrd)]
+pub struct AppendixRefOffset {
+    offset: u64,
+}
+
+impl AppendixRefOffset {
+    pub fn raw(&self) -> u64 {
+        self.offset
+    }
+}
+
 #[derive(Clone, Copy, Debug, Eq, PartialEq, Ord, PartialOrd)]
 pub enum Offset {
     FileOffset { offset: u64 },
+    AppendixRefOffset { offset: u64 },
 }
 
 /// Represents a named directory entry
@@ -130,6 +154,12 @@ pub enum DirEntryAttribute {
         mtime: i64,
         extension: Option<CatalogV2Extension>,
     },
+    AppendixRef {
+        size: u64,
+        mtime: i64,
+        ctime: i64,
+        appendix_ref_offset: AppendixRefOffset,
+    },
     Symlink,
     Hardlink,
     BlockDevice,
@@ -195,6 +225,17 @@ impl DirEntry {
                 name,
                 attr: DirEntryAttribute::Socket,
             },
+            (CatalogEntryType::AppendixRef, Some(Offset::AppendixRefOffset { offset })) => {
+                DirEntry {
+                    name,
+                    attr: DirEntryAttribute::AppendixRef {
+                        size,
+                        mtime,
+                        ctime,
+                        appendix_ref_offset: AppendixRefOffset { offset },
+                    },
+                }
+            }
             _ => panic!("unexpected parameters '{etype}' and '{offset:?}'"),
         }
     }
@@ -204,6 +245,7 @@ impl DirEntry {
         Some(match self.attr {
             DirEntryAttribute::Directory { .. } => pxar::mode::IFDIR,
             DirEntryAttribute::File { .. } => pxar::mode::IFREG,
+            DirEntryAttribute::AppendixRef { .. } => pxar::mode::IFREG,
             DirEntryAttribute::Symlink => pxar::mode::IFLNK,
             DirEntryAttribute::Hardlink => return None,
             DirEntryAttribute::BlockDevice => pxar::mode::IFBLK,
@@ -271,6 +313,24 @@ impl DirInfo {
                     catalog_encode_u64(writer, file_offset.raw())?;
                 }
             }
+            DirEntry {
+                name,
+                attr:
+                    DirEntryAttribute::AppendixRef {
+                        size,
+                        mtime,
+                        ctime,
+                        appendix_ref_offset,
+                    },
+            } => {
+                writer.write_all(&[CatalogEntryType::AppendixRef as u8])?;
+                catalog_encode_u64(writer, name.len() as u64)?;
+                writer.write_all(name)?;
+                catalog_encode_u64(writer, *size)?;
+                catalog_encode_i64(writer, *mtime)?;
+                catalog_encode_i64(writer, *ctime)?;
+                catalog_encode_u64(writer, appendix_ref_offset.raw())?;
+            }
             DirEntry {
                 name,
                 attr: DirEntryAttribute::Symlink,
@@ -390,6 +450,21 @@ impl DirInfo {
                     };
                     callback(etype, name, 0, size, mtime, ctime, offset)?
                 }
+                CatalogEntryType::AppendixRef => {
+                    let size = catalog_decode_u64(&mut cursor)?;
+                    let mtime = catalog_decode_i64(&mut cursor)?;
+                    let ctime = catalog_decode_i64(&mut cursor)?;
+                    let offset = catalog_decode_u64(&mut cursor)?;
+                    callback(
+                        etype,
+                        name,
+                        0,
+                        size,
+                        mtime,
+                        ctime,
+                        Some(Offset::AppendixRefOffset { offset }),
+                    )?
+                }
                 _ => callback(etype, name, 0, 0, 0, 0, None)?,
             };
             if !cont {
@@ -517,6 +592,34 @@ impl<W: Write> BackupCatalogWriter for CatalogWriter<W> {
         Ok(())
     }
 
+    fn add_appendix_ref(
+        &mut self,
+        name: &CStr,
+        size: u64,
+        mtime: i64,
+        ctime: i64,
+        appendix_ref_offset: pxar::encoder::AppendixRefOffset,
+    ) -> Result<(), Error> {
+        let dir = self
+            .dirstack
+            .last_mut()
+            .ok_or_else(|| format_err!("outside root"))?;
+        let name = name.to_bytes().to_vec();
+        let appendix_ref_offset = AppendixRefOffset {
+            offset: appendix_ref_offset.raw(),
+        };
+        dir.entries.push(DirEntry {
+            name,
+            attr: DirEntryAttribute::AppendixRef {
+                size,
+                mtime,
+                ctime,
+                appendix_ref_offset,
+            },
+        });
+        Ok(())
+    }
+
     fn add_symlink(&mut self, name: &CStr) -> Result<(), Error> {
         let dir = self
             .dirstack
-- 
2.39.2