ftruncate of unlinked file returns -1 with errno ENOENT

[qbolec]

Windows Version

Microsoft Windows [Version 10.0.26200.8037]

WSL Version

2.6.3.0

Are you using WSL 1 or WSL 2?

• WSL 2
• WSL 1

Kernel Version

6.6.87.2-1

Distro Version

podman-machine-default and Ubuntu, too

Other Software

I can reproduce it on both podman-machine-default and wsl -d Ubuntu

Repro Steps

#include <cerrno>
#include <cstring>
#include <fcntl.h>
#include <iostream>
#include <sstream>
#include <string>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>

namespace {

void print_errno(const char *label) {
  std::cout << label << ": errno=" << errno << " (" << std::strerror(errno) << ")\n";
}

void print_stat(const std::string &path) {
  struct stat st {};
  if (stat(path.c_str(), &st) == 0) {
    std::cout << "stat(" << path << "): mode=" << std::oct << (st.st_mode & 0777) << std::dec
              << " size=" << st.st_size << " uid=" << st.st_uid << " gid=" << st.st_gid << "\n";
  } else {
    print_errno(("stat(" + path + ")").c_str());
  }
}

void read_back(int fd, const char *label) {
  if (lseek(fd, 0, SEEK_SET) == -1) {
    print_errno("lseek(SEEK_SET)");
    return;
  }

  char buf[64] = {};
  errno = 0;
  ssize_t n = read(fd, buf, sizeof(buf) - 1);
  std::cout << label << ": read rc=" << n << "\n";
  if (n < 0) {
    print_errno(label);
    return;
  }
  std::cout << label << ": data=('" << std::string(buf, static_cast<size_t>(n)) << "')\n";
}

}

int main(int argc, char **argv) {
  std::string path = argc > 1 ? argv[1] : "/workspace/mysql-bin/ftruncate-repro.tmp";

  std::cout << "Using path: " << path << "\n";
  print_stat(path);

  int fd = open(path.c_str(), O_CREAT | O_RDWR | O_TRUNC, 0666);
  if (fd < 0) {
    print_errno("open");
    return 1;
  }
  std::cout << "open: fd=" << fd << "\n";
  print_stat(path);

  const char payload[] = "hello";
  ssize_t written = write(fd, payload, sizeof(payload) - 1);
  if (written < 0) {
    print_errno("write");
    close(fd);
    return 2;
  }
  std::cout << "write: wrote=" << written << "\n";
  print_stat(path);

  int rc = 0;
  errno = 0;
  rc = unlink(path.c_str());
  std::cout << "unlink(path): rc=" << rc << "\n";
  if (rc != 0) {
    print_errno("unlink(path)");
  }
  print_stat(path);

  errno = 0;
  rc = ftruncate(fd, 0);
  std::cout << "ftruncate(fd, 0): rc=" << rc << "\n";
  if (rc != 0) {
    print_errno("ftruncate(fd, 0)");
  }
  print_stat(path);
  errno = 0;
  written = write(fd, "Z", 1);
  std::cout << "write after ftruncate(fd, 0): wrote=" << written << "\n";
  if (written < 0) {
    print_errno("write after ftruncate(fd, 0)");
  }
  read_back(fd, "read after ftruncate(fd, 0)");

  errno = 0;
  rc = ftruncate(fd, 12345);
  std::cout << "ftruncate(fd, 12345): rc=" << rc << "\n";
  if (rc != 0) {
    print_errno("ftruncate(fd, 12345)");
  }
  print_stat(path);
  errno = 0;
  if (lseek(fd, 0, SEEK_END) == -1) {
    print_errno("lseek(SEEK_END)");
  }
  written = write(fd, "Y", 1);
  std::cout << "write after ftruncate(fd, 12345): wrote=" << written << "\n";
  if (written < 0) {
    print_errno("write after ftruncate(fd, 12345)");
  }
  read_back(fd, "read after ftruncate(fd, 12345)");

  if (close(fd) != 0) {
    print_errno("close");
    return 3;
  }
  std::cout << "close: ok\n";
  print_stat(path);
  return 0;
}

compile it, then run within the wsl like this: ./ftruncate_repro ./ftruncate-unlinked-repro.tmp

Expected Behavior

I would expect that all the operations (seek, write, read, and ftruncate) to still work on a descriptor even after the file was unlinked.

Actual Behavior

While write and read continue to work even after unlink, the ftruncate, doesn't:

qbolec@DESKTOP-UCOA13U:/mnt/c/ade/mysql/worker-1/mysql-bin$ ./ftruncate_repro ./ftruncate-unlinked-repro.tmp
Using path: ./ftruncate-unlinked-repro.tmp
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
open: fd=3
stat(./ftruncate-unlinked-repro.tmp): mode=777 size=0 uid=1000 gid=1000
write: wrote=5
stat(./ftruncate-unlinked-repro.tmp): mode=777 size=5 uid=1000 gid=1000
unlink(path): rc=0
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
ftruncate(fd, 0): rc=-1
ftruncate(fd, 0): errno=2 (No such file or directory)
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
write after ftruncate(fd, 0): wrote=1
read after ftruncate(fd, 0): read rc=6
read after ftruncate(fd, 0): data=('helloZ')
ftruncate(fd, 12345): rc=-1
ftruncate(fd, 12345): errno=2 (No such file or directory)
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
write after ftruncate(fd, 12345): wrote=1
read after ftruncate(fd, 12345): read rc=7
read after ftruncate(fd, 12345): data=('helloZY')
close: ok
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)

Diagnostic Logs

No response

[github-actions]

Logs are required for review from WSL team

If this a feature request, please reply with '/feature'. If this is a question, reply with '/question'.
Otherwise, please attach logs by following the instructions below, your issue will not be reviewed unless they are added. These logs will help us understand what is going on in your machine.

How to collect WSL logs

Download and execute collect-wsl-logs.ps1 in an administrative powershell prompt:

Invoke-WebRequest -UseBasicParsing "https://raw.githubusercontent.com/microsoft/WSL/master/diagnostics/collect-wsl-logs.ps1" -OutFile collect-wsl-logs.ps1
Set-ExecutionPolicy Bypass -Scope Process -Force
.\collect-wsl-logs.ps1

The script will output the path of the log file once done.

If this is a networking issue, please use .\collect-wsl-logs.ps1 -LogProfile networking instead, following the instructions in Collect WSL logs for networking issues

Once completed please upload the output files to this GitHub issue.

See Collect WSL logs (recommended method).

If you choose to email these logs instead of attaching them to the bug, please send them to wsl-gh-logs@microsoft.com with the GitHub issue number in the subject, and include a link to your GitHub issue comment in the message body, and reply with '/emailed-logs'.

[qbolec]

/emailed-logs

[github-actions]

Diagnostic information

Found '/emailed-logs', adding tag 'emailed-logs'

[OneBlue]

@qbolec: This is standard linux behavior, which WSL2 doesn't control. Closing since this isn't a WSL issue.

[qbolec]

I don't think it is a standard linux behaviour. What makes you think so?
Here's the same program executed on Oracle Enterprise Linux 8, for example:

[jlopusza@compile-el8-01]/export/home/tmp/jlopusza/mysql: /opt/rh/gcc-toolset-14/root/bin/g++ -o ftruncate_repro ftruncate_repro.cpp
[jlopusza@compile-el8-01]/export/home/tmp/jlopusza/mysql: ./ftruncate_repro ./ftruncate-unlinked-repro.tmp
Using path: ./ftruncate-unlinked-repro.tmp
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
open: fd=3
stat(./ftruncate-unlinked-repro.tmp): mode=644 size=0 uid=244454 gid=31415
write: wrote=5
stat(./ftruncate-unlinked-repro.tmp): mode=644 size=5 uid=244454 gid=31415
unlink(path): rc=0
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
ftruncate(fd, 0): rc=0
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
write after ftruncate(fd, 0): wrote=1
read after ftruncate(fd, 0): read rc=6
read after ftruncate(fd, 0): data=('Z')
ftruncate(fd, 12345): rc=0
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)
write after ftruncate(fd, 12345): wrote=1
read after ftruncate(fd, 12345): read rc=63
read after ftruncate(fd, 12345): data=('Z')
close: ok
stat(./ftruncate-unlinked-repro.tmp): errno=2 (No such file or directory)

Notice, how in this case ftruncate returns 0.