From: nospam@needed.invalid   
      
   On Mon, 9/15/2025 10:55 PM, Lawrence D’Oliveiro wrote:   
   > On Mon, 15 Sep 2025 22:34:30 -0400, Paul wrote:   
   >   
   >> On Mon, 9/15/2025 7:49 PM, Lawrence D’Oliveiro wrote:   
   >>>   
   >>> On Mon, 15 Sep 2025 13:33:17 -0700, bilsch01 wrote:   
   >>>   
   >>>> I want to create a bootable Ubuntu installer thumb drive using Win11.   
   >>>   
   >>> You could install WSL2 and take advantage of its Linux functionality,   
   >>> like the “dd” command.   
   >>>   
   >>>   
   >> Yikes.   
   >>   
   >> You can do better than that.   
   >   
   > You can on a native Linux system, but we’re trying to get to that.   
   >   
   >> That gives you "dd.exe", to be run in a Windows Administrator terminal   
   >> session. The tool was previously "rawrite" and the name changed to "dd"   
   >> at some point. It is an acquired taste (does not detect the end of the   
   >> source disk properly for one kind of media, not a big deal).   
   >   
   > A load of crap, in other words.   
   >   
   >> The WSL2 does not have a working /dev layer. It's not designed to hack   
   >> around or subvert Windows itself. The people who worked on WSL,   
   >> evidence points to them being pretty skilled at what they did.   
   >   
   > But maybe not skilled enough. Does “lsblk” work? Does the content of   
   > /proc/self/mountinfo show anything sensible?   
   >   
   >> The Linux "dd" might work on those, but I don't think there is much   
   >> profit to be had, by doing so.   
   >   
   > So much for being “skilled”, eh?   
   >   
   >> And if we look at mounts...   
   >>   
   >> $ ls /mnt   
   >> c  d  e  f  g  h  i  k  s  wsl  wslg   
   >>   
   >> it's a pretty fuzzy collection of stuff you could just as easily attack   
   >> from the Windows side.   
   >   
   > That’s not how you normally query what’s mounted. Try the “df”   
   command.   
   >   
   >> Using the Linux "dd" from there, is not a big help.   
   >   
   > You might be right.   
   >   
   >> However, if you use Windows VirtualBox and had the PUEL hardware   
   >> passthru installed (as a home user, not for commercial use), then you   
   >> could connect a USB stick to the PC, and via passthru, a Linux LiveDVD   
   >> in VirtualBox could be used to write to the USB stick. Similarly, you   
   >> could run Linux VirtualBox, and using the passthru, do the same thing.   
   >> The passthru does not work for everything (you can't boot from a USB   
   >> stick via passthru), but for a limited set of tasks, you can succeed via   
   >> passthru. Maybe even a USB DVD writer could be run that way. A regular   
   >> DVD writer over a SATA cable, does not normally work,   
   >> but you might succeed over USB.   
   >   
   > Windows just doesn’t want to make it easy, does it?   
   >   
      
   There is one in Coreutils.   
      
      https://gnuwin32.sourceforge.net/packages/coreutils.htm   
      
   Needs the bin package and the dep package (two DLLs).   
      
   Wed, 04/20/2005  02:41 PM            87,552 dd.exe   
   Tue, 03/16/2004  04:37 PM           898,048 libiconv2.dll   
   Sat, 10/09/2004  12:25 PM           101,888 libintl3.dll   
      
   *******   
   D:\>.\dd --help                     # Coreutils...   
   Usage: .\dd [OPERAND]...   
     or:  .\dd OPTION   
   Copy a file, converting and formatting according to the operands.   
      
     bs=BYTES        force ibs=BYTES and obs=BYTES   
     cbs=BYTES       convert BYTES bytes at a time   
     conv=CONVS      convert the file as per the comma separated symbol list   
     count=BLOCKS    copy only BLOCKS input blocks   
     ibs=BYTES       read BYTES bytes at a time   
     if=FILE         read from FILE instead of stdin   
     iflag=FLAGS     read as per the comma separated symbol list   
     obs=BYTES       write BYTES bytes at a time   
     of=FILE         write to FILE instead of stdout   
     oflag=FLAGS     write as per the comma separated symbol list   
     seek=BLOCKS     skip BLOCKS obs-sized blocks at start of output   
     skip=BLOCKS     skip BLOCKS ibs-sized blocks at start of input   
     status=noxfer   suppress transfer statistics   
      
   BLOCKS and BYTES may be followed by the following multiplicative suffixes:   
   xM M, c 1, w 2, b 512, kB 1000, K 1024, MB 1000*1000, M 1024*1024,   
   GB 1000*1000*1000, G 1024*1024*1024, and so on for T, P, E, Z, Y.   
      
   Each CONV symbol may be:   
      
     ascii     from EBCDIC to ASCII   
     ebcdic    from ASCII to EBCDIC   
     ibm       from ASCII to alternate EBCDIC   
     block     pad newline-terminated records with spaces to cbs-size   
     unblock   replace trailing spaces in cbs-size records with newline   
     lcase     change upper case to lower case   
     nocreat   do not create the output file   
     excl      fail if the output file already exists   
     notrunc   do not truncate the output file   
     ucase     change lower case to upper case   
     swab      swap every pair of input bytes   
     noerror   continue after read errors   
     sync      pad every input block with NULs to ibs-size; when used   
                 with block or unblock, pad with spaces rather than NULs   
     fdatasync physically write output file data before finishing   
     fsync     likewise, but also write metadata   
      
   Each FLAG symbol may be:   
      
     append    append mode (makes sense only for output)   
     sync      likewise, but also for metadata   
     nonblock  use non-blocking I/O   
      
   Sending a SIGUSR1 signal to a running `dd' process makes it   
   print I/O statistics to standard error, then to resume copying.   
      
     $ dd if=/dev/zero of=/dev/null& pid=$!   
     $ kill -USR1 $pid; sleep 1; kill $pid   
     18335302+0 records in   
     18335302+0 records out   
     9387674624 bytes (9.4 GB) copied, 34.6279 seconds, 271 MB/s   
      
   Options are:   
      
         --help     display this help and exit   
         --version  output version information and exit   
      
   Report bugs to .   
      
   D:\>   
   *******   
      
   The chrysocome.net one can make a file. A 1GB test file.   
      
       dd if=/dev/zero of=D:\test.bin bs=1048576 count=1000   
      
   The gnuwin32 Coreutils one can copy it. It's doing about 4GB a second,   
   but the utility is a little "math challenged" when working out the speed.   
      
   dd if=test.bin of=testcopy.bin bs=1048576   
   1000+0 records in   
   1000+0 records out   
   1048576000 bytes (1.0 GB) copied, 0.278 seconds, -   
   1798100249891578800000000000000000000000000000000000000000000000   
   0000000000000000000000000000000000000000000000000000000000000000   
   0000000000000000000000000000000000000000000000000000000000000000   
   00000000000000000000000000000000000000000000 GB/s   
      
   If we do the same copy test with the chrysocome one...   
      
   D:\TEMP>dd if=D:\test.bin of=D:\testcopy.bin bs=1048576   
   rawwrite dd for windows version 0.6beta3.   
   Written by John Newbigin    
   This program is covered by terms of the GPL Version 2.   
      
   1000+0 records in   
   1000+0 records out   
      
   *******   
      
   The letter D: is on /dev/sdc, so next we try for a physical layer test.   
      
   D:\>.\dd if=/dev/sdc of=- bs=1048576 | wc -c   
   .\dd: opening `/dev/sdc': No such file or directory   
      
   So it's not a cygwin type implementation.   
      
   We can try the namespace the chrysocome one knows.   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)