[En-Nut-Discussion] UsartRead and timeout

Uwe Bonnes bon at elektron.ikp.physik.tu-darmstadt.de
Tue Apr 25 11:25:10 CEST 2023


okay, the implementation does not contradict the specs. And I already
implemented reading in a loop.

Thanks for the feedback

>>>>> "Ole" == Ole Reinhardt <ole.reinhardt at embedded-it.de> writes:

    Ole> Hi Uwe, yes I would expect this behavior to be correct.

    Ole> The UsartRead() routine will normally be called by read()
    Ole> from the c-library.

    Ole> read() is defined to return the currently available number of
    Ole> bytes, which may be lower than the requested amount.

    Ole> See e.g. https://man7.org/linux/man-pages/man2/read.2.html

    Ole> When writing a parser, that reads data from a device, I would
    Ole> normally use a loop, that calls read() until the expected
    Ole> amount of data is received. If you do not know the number of
    Ole> bytes that shall be received, just implement an timeout,
    Ole> until which the first character shall be received and then an
    Ole> interbyte timeout, until which the next character) shall be
    Ole> received.

    Ole> I would use the select() routine to implement these timeouts.

    Ole> Best regards,

    Ole> Ole Reinhardt

    Ole> Am 24.04.23 um 19:28 schrieb Uwe Bonnes:
    >> Hello,
    >> ../../dev/usart.c in UsartRead reads: 228 /* 229 * Wait until
    >> at least one character is buffered or until a read 230 *
    >> timeout occured.  231 */
    >> Is this really the intended behaviour? I would expect to read
    >> like
    >> * Wait until the requested number of character is buffered or
    >> until a read * timeout occured.
    >> This would be in line with e.g.
    >> https://learn.microsoft.com/en-us/windows-hardware/drivers/serports/setting-read-and-write-timeouts-for-a-serial-device"
    >> "Serial.sys continues to transfer bytes until the requested
    >> number of bytes are transferred or a time-out event occurs."
    >> Returning after the first received byte makes it hard to wait
    >> for the full multi-byte answer of an external device.
    >> What is your opinion?
    >> Thanks

    Ole> _______________________________________________
    Ole> http://lists.egnite.de/mailman/listinfo/en-nut-discussion

Uwe Bonnes                bon at elektron.ikp.physik.tu-darmstadt.de

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
--------- Tel. 06151 1623569 ------- Fax. 06151 1623305 ---------

More information about the En-Nut-Discussion mailing list