I cannot give a general definition of "embedded systems", but for sure I can explain my concern : my project (= hobby) is to run ELKS on this SBC : http://www.advantech.com/products/1-2jkp2y/snmp-1000-b/mod_b01abe8d-c8aa-4530-8e3c-9fcd2623d9d6 It was primarily intended for remote PC HW control / monitoring (Telnet, HTTP, SNMP, etc) but it can easily be hacked as it is based on a 80188 / PC architecture, and has a removable Flash that can be reprogrammed easily with SW protocol. Moreover it has an ETH interface, plenty of I/O pins, and I can find several of them in my company garbage containers :-). In the Flash, the manufacturer (Advantech) implemented a firmware clearly derived from a PC BIOS, but with uggly shortcuts (like redirecting the BIOS 10h to serial, or reusing some standard INT numbers without compliance to the IBM BIOS reference, etc). You can have a look to the EMU86 INT emulation to get an idea of how far this BIOS is away from the standard : https://github.com/jbruchon/dev86/blob/master/emu86/emu-int.c This is why, in such situation, if Jody's adds invocation of the BIOS 15h in ELKS, it should be surrounded by a configuration option, otherwise I would have to add an INT 15h handler in the Flash to return CF=1. If even if I would like to implement the power control, I see no HW facility on the board to cut its own power supply (= passive voltage adapter) without adding an external circuit. MFLD 2016-05-26 11:40 GMT+02:00 Kirn Gill <segin2005@xxxxxxxxx>: > I'm not 100% on the linguistics, but can you please explain what you meant > about embedded systems? Do you mean ones without a BIOS? > > Wouldn't it make more sense to tie this feature to a build target (e.g. PC > and not embedded) instead of it's own config switch? > > I suppose this will make for a good learning experience for me. > > On May 26, 2016 04:34, "Marc-François LUCCA-DANIAU" <mfld.fr@xxxxxxxxx> > wrote: >> >> IMHO, this feature makes sense for emulators intended for end-user, >> that try to mimic a real PC with a full-featured BIOS,but raises an >> issue for the embedded devices, because it forces to implement BIOS >> 15h, at least a stub to return with CF set. >> >> For that reason, I would enhance your change by adding something like >> an "BIOS power management" option in the ELKS configuration >> (CONFIG_BIOS_POWER ?). >> >> MFLD >> >> >> 2016-05-26 3:51 GMT+02:00 Jody Bruchon <jody@xxxxxxxxxxxxxxx>: >> > I have just committed a 'poweroff' command to ELKS. Piggybacking on the >> > existing sys_reboot() syscall, the commit uses Advanced Power Management >> > (APM) BIOS calls to power down the computer when the poweroff command is >> > issued. If you're running in an emulator like QEMU, Bochs, or >> > Virtualbox, or >> > if you're running ELKS on any APM-enabled platform such as most 486 or >> > higher laptops and any older ATX PC, this is a much more convenient way >> > to >> > shut down. >> > >> > I find this to be a nicer way of exiting the QEMU curses interface than >> > ESC, >> > 2, q, [enter] in any case, and most ELKS testing seems to be done on >> > more >> > modern systems and emulators, so it makes sense. I would like someone to >> > test it on a non-APM system to make sure that a failure due to no APM >> > support falls through to an error message. >> > >> > Also, I am thinking of renaming sys_reboot() to sys_shutdown() to better >> > reflect its actual purpose. Any objections? >> > >> > -Jody >> > -- >> > To unsubscribe from this list: send the line "unsubscribe linux-8086" in >> > the body of a message to majordomo@xxxxxxxxxxxxxxx >> > More majordomo info at http://vger.kernel.org/majordomo-info.html >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-8086" in >> the body of a message to majordomo@xxxxxxxxxxxxxxx >> More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-8086" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
