Hi Donald, Donald Russell <russell....@gmail.com> writes:
> I have RHEL 7.9 setup (on zVM) so I can use the 3270 console instead of > 3215 and it works quite nicely provided the 3270 geometry is 24 X 80. > > If the screen geometry is other than that, it’s not usable…. Input commands > are not parsed properly and fail, and of course the output display is a > mess due to the wrong buffer address calculations for sba orders etc. > > I detect in near real time when the console reconnects … is there a (linux) > command I can issue that will “shoulder tap” the 3270 device driver to > reread the device characteristics, specifically screen geometry? Presumably > by issuing diagnose 8c or issue the query to the 3270 device “tell me about > yourself”. I checked the kernel source, and it looks like Linux has different ways of figuring out the geometry: - on VM, it uses a diag120 call. From the device type, it determines the geometry. So it can only handle the fixed resolution of the -1, -2 and so on models. I haven't found any reference to diag 8c. - if it's !VM (lpar), it uses a read partition query to figure out the geometry. Interestingly though, this also seems to work on VM when i force the 3270 driver to use the "lpar mode". I have to ask the VM people why that differentiation exists - maybe VM couldn't handle the query in former versions. But maybe it's just my limited knowledge of 3270 support in VM and Linux, i have never seen a real 3270 :-) > I have a udev rule that is trigggered (indirectly) by a reconnect event so > that’s the point where I want to tell the driver “check the 3270 device > characteristics”. I haven't checked whether something like 'stty [rows|columns] X < /dev/3270/ttyX' works. But with the networking 3270 clients we have nowadays, resolutions are changing all the time and i guess this wouldn't help much even if it works. > P.S. I’m going to do this on RHEL 8 soon and maybe 9. I haven’t yet > checked to see if the driver has been updated in those versions. I tested it with upstream 6.1-rc and was able to reproduce the behaviour you described. So i don't think you need to test it with RHEL8/9. ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit http://www2.marist.edu/htbin/wlvindex?LINUX-390
