diff options
author | Huang Shijie <b32955@freescale.com> | 2012-08-08 10:37:59 +0800 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2012-08-16 12:15:33 -0700 |
commit | 00592021010ad86d3b26bac7034034f6af145a2c (patch) | |
tree | ad64391844ab9a7cbfaf4a9b693ccd73586f75be /lib/mpi | |
parent | 7d9739cd6b03575bc88c6ce3c60853dbc50a6ae2 (diff) | |
download | linux-3.10-00592021010ad86d3b26bac7034034f6af145a2c.tar.gz linux-3.10-00592021010ad86d3b26bac7034034f6af145a2c.tar.bz2 linux-3.10-00592021010ad86d3b26bac7034034f6af145a2c.zip |
serial: mxs-auart: fix the wrong RTS hardware flow control
Without checking if the auart supports the hardware flow control or not,
the old mxs_auart_set_mctrl() asserted the RTS pin blindly.
This will causes the auart receives wrong data in the following case:
The far-end has already started the write operation, and wait for
the auart asserts the RTS pin. Then the auart starts the read operation,
but mxs_auart_set_mctrl() may be called before we set the RTSCTS in the
mxs_auart_settermios(). So the RTS pin is asserted in a wrong situation,
and we get the wrong data in the end.
This bug has been catched when I connect the mx23(DTE) to the mx53(DCE).
This patch also replaces the AUART_CTRL2_RTS with AUART_CTRL2_RTSEN.
We should use the real the hardware flow control, not the software-controled
hardware flow control.
Signed-off-by: Huang Shijie <b32955@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'lib/mpi')
0 files changed, 0 insertions, 0 deletions