Dear Allen Martin,
On Fri, Oct 26, 2012 at 11:39:48AM -0700, Joe Hershberger wrote:
Hi Marek,
On Fri, Oct 26, 2012 at 5:22 AM, Marek Vasut marex@denx.de wrote:
Dear Joe Hershberger,
Hi Allen,
On Thu, Oct 25, 2012 at 4:19 PM, Allen Martin amartin@nvidia.com wrote:
On Thu, Oct 25, 2012 at 02:02:55PM -0700, Simon Glass wrote:
Hi,
On Thu, Oct 25, 2012 at 12:03 PM, Marek Vasut marex@denx.de wrote: > Dear Simon Glass, > >> Hi, >> >> On Mon, Oct 22, 2012 at 10:23 AM, Allen Martin >> amartin@nvidia.com
wrote:
>> > On Sat, Oct 20, 2012 at 01:19:00AM -0700, Marek Vasut wrote: >> >> Dear Allen Martin, >> >> >> >> [...] >> >> >> >> > Hi Marek, the change to return value here broke serial >> >> > output on tegra. What I see is that the serial device >> >> > name (s->name) is "eserial0" as set by serial_ns16550.c, >> >> > and the name passed in from the stdout environment is >> >> > "serial" so they don't match and it fails. This always >> >> > used to be ok because the return code didn't indicate >> >> > failure and iomux_doenv() would continue on happily, but >> >> > now it causes iomux_doenv() to fail and no printfs() work >> >> > after that. >> >> > >> >> > Not sure what the right fix is, should stdout really be set >> >> > to "eserial0"? It seems "serial" should mean "the default >> >> > serial device" which for the normal case is the one and >> >> > only device. >> >> >> >> Looking at the source, the obvious course of action is to fix >> >> iomux.c . >> > >> > I've been looking at this call to serial_assign() from iomux.c >> > and I'm not convinced this code does anything meaningful at >> > all. It passes the name of a struct stdio_dev device which >> > serial_assign() then tries to match against the registered >> > struct serial_devices, which will never match. >> > >> > What I don't understand is the case where you have a board >> > that actually has more than one physical serial port and how >> > the mapping from stdio_dev to serial_device happens. >> > >> > Also, looking at the code to cmd_nvedit, I think your change >> > also broke "setenv stdout" for boards that don't define >> > CONFIG_CONSOLE_MUX. We always have this on for tegra, so we >> > don't go down this code path, but it looks identical to the >> > code in iomux.c >> >> Sorry if I missed it - what was the resolution here? Should we >> revert that change? > > Definitelly not. We should fix the iomux.c , possibly by flipping > the inequation mark as a short term solution.
OK that's fine. Is someone working on a patch?
I'll send out my proposal for a patch. Unfortunately I don't have a board with multiple serial ports to correctly test CONFIG_SERIAL_MULTI
One of the boards I'm working on does this (has more than one). At least before the serial rework from Marek, the stdout was either the serial device directly (each serial device was added as a console device as well) so the serial setting was redundant. You could just set them directly to the serial port (which is more flexible).
I had two patches (not sent to ML before Marek made them highly conflicting)
I know, he's such a bastard, always breaking stuff and interfering with other people's work !
: :) :
that take the opposite approach you were, since it preserves the flexibility. It removed the "serial" setting to each of the std* variables and instead sets it to the default serial device. I'll remake that patch on top of the new serial landscape sometime soon.
Actually, I'd like to merge the serial stuff and stdio stuff into one. So "setenv stdX serial" would be replaced with "setenv stdX <serial_driver_name>". I think that's the approach to take. But it'd break many boards.
I think that's fine. We should be able to make it work both ways. In other words, as long as there is only one serial device registered, then "serial" will be accepted. The actual name of the driver is also always accepted.
Or, ever better we could just say "serial" always maps to default_serial_console() and that should map exactly to existing functionality regardless if there is one or more serial devices. Then any board that wants more precise control can use the name of the serial driver instead.
So default_serial_console() will return a char * instead of driver pointer ... good.
Best regards, Marek Vasut