diff --git a/src/tape/mod.rs b/src/tape/mod.rs
index b5858a6f..36e2efc5 100644
--- a/src/tape/mod.rs
+++ b/src/tape/mod.rs
@@ -1 +1,4 @@
 pub mod file_formats;
+
+mod tape_write;
+pub use tape_write::*;
diff --git a/src/tape/tape_write.rs b/src/tape/tape_write.rs
new file mode 100644
index 00000000..7e354c88
--- /dev/null
+++ b/src/tape/tape_write.rs
@@ -0,0 +1,105 @@
+use std::io::Write;
+
+use endian_trait::Endian;
+
+use proxmox::sys::error::SysError;
+
+use crate::tape::file_formats::MediaContentHeader;
+
+/// Write trait for tape devices
+///
+/// The 'write_all' function returns if the drive reached the Logical
+/// End Of Media (early warning).
+///
+/// It is mandatory to call 'finish' before closing the stream to mark it
+/// as correctly written.
+///
+/// Please note that there is no flush method. Tapes flush there internal
+/// buffer when they write an EOF marker.
+pub trait TapeWrite {
+    /// writes all data, returns true on LEOM
+    fn write_all(&mut self, data: &[u8]) -> Result<bool, std::io::Error>;
+
+    /// Returns how many bytes (raw data on tape) have been written
+    fn bytes_written(&self) -> usize;
+
+    /// flush last block, write file end mark
+    ///
+    /// The incomplete flag is used to mark multivolume stream.
+    fn finish(&mut self, incomplete: bool) -> Result<bool, std::io::Error>;
+
+    /// Returns true if the writer already detected the logical end of media
+    fn logical_end_of_media(&self) -> bool;
+
+    /// writes header and data, returns true on LEOM
+    fn write_header(
+        &mut self,
+        header: &MediaContentHeader,
+        data: &[u8],
+    ) -> Result<bool, std::io::Error> {
+        if header.size as usize != data.len() {
+            proxmox::io_bail!("write_header with wrong size - internal error");
+        }
+        let header = header.to_le();
+
+        let res = self.write_all(unsafe { std::slice::from_raw_parts(
+            &header as *const MediaContentHeader as *const u8,
+            std::mem::size_of::<MediaContentHeader>(),
+        )})?;
+
+        if data.is_empty() { return Ok(res); }
+
+        self.write_all(data)
+    }
+}
+
+/// Write a single block to a tape device
+///
+/// Assumes that 'writer' is a linux tape device.
+///
+/// EOM Behaviour on Linux: When the end of medium early warning is
+/// encountered, the current write is finished and the number of bytes
+/// is returned. The next write returns -1 and errno is set to
+/// ENOSPC. To enable writing a trailer, the next write is allowed to
+/// proceed and, if successful, the number of bytes is returned. After
+/// this, -1 and the number of bytes are alternately returned until
+/// the physical end of medium (or some other error) is encountered.
+///
+/// See: https://github.com/torvalds/linux/blob/master/Documentation/scsi/st.rst
+///
+/// On sucess, this returns if we en countered a EOM condition.
+pub fn tape_device_write_block<W: Write>(
+    writer: &mut W,
+    data: &[u8],
+) -> Result<bool, std::io::Error> {
+
+    let mut leof = false;
+
+    loop {
+        match writer.write(data) {
+            Ok(count) if count == data.len() => return Ok(leof),
+            Ok(count) if count > 0 => {
+                proxmox::io_bail!(
+                    "short block write ({} < {}). Tape drive uses wrong block size.",
+                    count, data.len());
+            }
+            Ok(_) => { // count is 0 here, assume EOT
+                return Err(std::io::Error::from_raw_os_error(nix::errno::Errno::ENOSPC as i32));
+            }
+            // handle interrupted system call
+            Err(err) if err.kind() == std::io::ErrorKind::Interrupted => {
+                continue;
+            }
+            // detect and handle LEOM (early warning)
+            Err(err) if err.is_errno(nix::errno::Errno::ENOSPC) => {
+                if leof {
+                    return Err(err);
+                } else {
+                    leof = true;
+                    continue; // next write will succeed
+                }
+            }
+            Err(err) => return Err(err),
+        }
+    }
+}