<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>On 14/02/2020 16:29, Joel Sherrill wrote:<br>
    </p>
    <blockquote type="cite"
cite="mid:CAF9ehCWx1LpTt2EUX=yptVZW4C37krXWnsRXpy3Dp4TUpvtn+A@mail.gmail.com">
      <div dir="ltr"><br>
      </div>
      <br>
      <div class="gmail_quote">
        <div dir="ltr" class="gmail_attr">On Fri, Feb 14, 2020 at 9:16
          AM Sebastian Huber <<a
            href="mailto:sebastian.huber@embedded-brains.de"
            moz-do-not-send="true">sebastian.huber@embedded-brains.de</a>>
          wrote:<br>
        </div>
        <blockquote class="gmail_quote" style="margin:0px 0px 0px
          0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
          <div>
            <p>On 14/02/2020 15:45, Joel Sherrill wrote:<br>
            </p>
            <blockquote type="cite">
              <div dir="ltr" class="gmail_attr">On Fri, Feb 14, 2020 at
                7:08 AM Sebastian Huber <<a
                  href="mailto:sebastian.huber@embedded-brains.de"
                  target="_blank" moz-do-not-send="true">sebastian.huber@embedded-brains.de</a>>
                wrote:<br>
              </div>
              <blockquote class="gmail_quote" style="margin:0px 0px 0px
                0.8ex;border-left:1px solid
                rgb(204,204,204);padding-left:1ex">On 14/02/2020 13:57,
                Joel Sherrill wrote:<br>
                <br>
                > Hi<br>
                ><br>
                > I haven't investigated or checked on other bsps but
                the xilinx qemu <br>
                > bsp has no output with the hello world in examples.
                It uses the devfs. <br>
                > Changing it to the default filesystem results in it
                working.<br>
                ><br>
                > I am suspicious of something with all the termios
                and new console <br>
                > changes breaking it.<br>
                The devfs works only with legacy IO drivers. The Termios
                drivers depend <br>
                on IMFS. I would like to remove the devfs in the future.
                The IMFS is <br>
                highly configurable, there is no need for an additional
                in-memory file <br>
                system.<br>
              </blockquote>
              <div><br>
              </div>
              <div>Do not remove it. Please fix it.</div>
            </blockquote>
            <p>The IMFS should be improved, so that we can remove the
              devfs. We already have a bunch of configuration options to
              customize it:</p>
            <p><a
href="https://docs.rtems.org/branches/master/c-user/configuring_a_system.html#file-system-configuration-parameters"
                target="_blank" moz-do-not-send="true">https://docs.rtems.org/branches/master/c-user/configuring_a_system.html#file-system-configuration-parameters</a></p>
          </div>
        </blockquote>
        <div><br>
        </div>
        <div>You should look at the requirements for the older ESA RTEMS
          versions. They </div>
        <div>had very specific requirements which ended up with the
          filesystem code </div>
        <div>largerly removed AFAIK.</div>
      </div>
    </blockquote>
    <p>You can get rid of the file system with the
      CONFIGURE_APPLICATION_DISABLE_FILESYSTEM configuration option.<br>
    </p>
    <blockquote type="cite"
cite="mid:CAF9ehCWx1LpTt2EUX=yptVZW4C37krXWnsRXpy3Dp4TUpvtn+A@mail.gmail.com">
      <div class="gmail_quote">
        <div><br>
        </div>
        <div>And you did break existing code in the public repository.
          If we need more tests </div>
        <div>to ensure it works, then so be it. It is a feature in the
          current code base and it </div>
        <div>just broke.</div>
        <div><br>
        </div>
        <div>Until the IMFS can be proved to be configurable to meet the
          requirements that </div>
        <div>resulted in the filesystem code being removed/minimized in
          the old ESA  versions<br>
        </div>
        <div>and other folks who needed the simplest possible
          infrastructure for devices so</div>
        <div>our normal frameworks would work, it needs to stay and
          work.</div>
      </div>
    </blockquote>
    If you want a minimal setup, then why do you want to use the POSIX
    open/read/write API to access devices? The POSIX functions pull in
    errno for example which pulls in the re-entrancy support which pulls
    in a hell of dependencies. Why do you want to use a console driver
    with Termios in a minimal setup?<br>
    <blockquote type="cite"
cite="mid:CAF9ehCWx1LpTt2EUX=yptVZW4C37krXWnsRXpy3Dp4TUpvtn+A@mail.gmail.com">
      <div class="gmail_quote">
        <div><br>
        </div>
        <div>You can have a goal which you may or may nor achieve but
          you can't leave a </div>
        <div>trail of code that is broken just because you don't like
          it.</div>
      </div>
    </blockquote>
    To fix this break, I suggest to map CONFIGURE_FILESYSTEM_DEVFS to a
    suitable IMFS configuration similar to
    CONFIGURE_USE_MINIIMFS_AS_BASE_FILESYSTEM.<br>
  </body>
</html>