https://bugzilla.kernel.org/show_bug.cgi?id=219098 Bug ID: 219098 Summary: Termios man page description of semantics of MIN, TIME Product: Documentation Version: unspecified Hardware: All OS: Linux Status: NEW Severity: normal Priority: P3 Component: man-pages Assignee: documentation_man-pages@xxxxxxxxxxxxxxxxxxxx Reporter: stewart.weiss@xxxxxxx Regression: No The current version of the termios(3) man page states, "MIN > 0, TIME == 0 (blocking read) read(2) blocks until MIN bytes are available, and returns up to the number of bytes requested. ". This used to be true but it no longer true. The case it does not cover is when the number of bytes requested is less than MIN. In this case read(2) is unblocked as soon as the number of bytes in the tty's line discipline is at least the number of bytes requested, even if it is less than MIN. The n_tty_read() function's while loop breaks when this is true. If I c_cc[VMIN] = 3, c_cc[VTIME] = 0 and disable icanon in the driver and then run a program such as: int main( int argc, char *argv[] ) { char inbuf; while( read(STDIN_FILENO, &inbuf, 1) > 0 ) { if ( -1 == write( 1, &inbuf, 1 )) exit(1); } exit(EXIT_SUCCESS); } the call to read() will return after each character. Shouldn't the correct description be: "MIN > 0, TIME == 0 (blocking read) If MIN is less than or equal to the number of bytes requested, read(2) blocks until MIN bytes are available, and returns up to the number of bytes requested. If MIN is greater than the number of bytes requested, read(2) is unblocked as soon as the number of bytes requested is available, regardless of the value of MIN. " My apologies if this is the wrong category for this bug report. I should add that this behavior was not true back in 2012, when I would teach it to my students. Then, it did behave as the current man page describes. ". -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
