blob: 94d2465cb5cec7e6261dff04895f27d3f8b6fd58 [file] [log] [blame]
Kevin O'Connor65e63422008-07-19 14:12:32 -04001Clean up timer code. Don't use PORT_DIAG as delay mechanism.
2
Kevin O'Connorf292f3b2008-06-28 12:19:52 -04003Disable a20 on jump to 16bit mode.
Kevin O'Connor22dfb852008-06-07 15:24:19 -04004
Kevin O'Connorc7812932008-06-08 23:08:12 -04005The keyboard driver should detect ctrl+alt+del and reboot the
6machine.
7
Kevin O'Connorf292f3b2008-06-28 12:19:52 -04008Fix keyboard LEDs.
9
Kevin O'Connorf13b0082008-08-17 11:26:42 -040010Improve option rom handling. A real bios can copy the option roms
Kevin O'Connor99373862008-10-25 23:10:27 -040011from PCI space. Also, there is no handling for BCVs today. Verify
12option roms wont stomp on seabios stack and bss.
Kevin O'Connorf13b0082008-08-17 11:26:42 -040013
14Clean up and clarify usage/scope of .bss and .data variables.
15
Kevin O'Connor2ad37442008-05-06 19:49:01 -040016Review changes committed to coreboot, virtualbox, qemu, kvm, and bochs
17cvs tip.
Kevin O'Connor99373862008-10-25 23:10:27 -040018 * bochs cvs 20081016: -- better post resume support
19 * coreboot (r3348): 20060708 -- no noteworthy enhancements
20 * qemu (r5536): 20080603 -- has cmos smp count
Kevin O'Connor970a0322008-10-26 12:01:21 -040021 * kvm (1a33292): 20071016 -- acpi dsdt changes, e820 vmx
Kevin O'Connor99373862008-10-25 23:10:27 -040022 pages, setup mtrrs, clear tsc, acpi intsrcovr table, different pci
23 irqs, smbios always uses 16 cpus, no acpi ssdt, fadt->gpe0_blk set
24 * virtualbox (r13560): 20061231 -- lots of mouse changes, logo,
25 scsi/etherboot hooks, int 1589, floppy data rate?, int19 calls
26 post
Kevin O'Connor2ad37442008-05-06 19:49:01 -040027
Kevin O'Connor5c4a8c62008-05-12 23:50:16 -040028Look into ways to reduce stack usage. Alter ISRs so that they do not
29enable irqs. Only call out to other bios functions after minimizing
30stack usage.
Kevin O'Connor2ad37442008-05-06 19:49:01 -040031
Kevin O'Connor7a558e42008-03-11 20:38:33 -040032Audit all sti/cli calls. Audit all call16 calls to make sure flags is
33setup properly with respect to irqs.
Kevin O'Connor44c631d2008-03-02 11:24:36 -050034
Kevin O'Connored128492008-03-11 11:14:59 -040035Audit statements where a 32bit intermediary changes meaning of a 16bit
36comparison.
37
Kevin O'Connor63dbcfb2008-03-01 22:17:07 -050038Code assumes ebda segment is static - it doesn't read 0x40e.
39
Kevin O'Connorb8aacb02008-03-01 14:56:07 -050040The __call16 code does a long jump to the interrupt trampolines - this
41is unnecessary.
Kevin O'Connorf076a3e2008-02-25 22:25:15 -050042
Kevin O'Connorf076a3e2008-02-25 22:25:15 -050043Cleanup setting of ES on GET/SET_BDA
44
Kevin O'Connor99373862008-10-25 23:10:27 -040045Audit code for 16bit protected mode accesses. Support 1ab1 from 16bit
46protected mode.
Kevin O'Connor65e63422008-07-19 14:12:32 -040047
Kevin O'Connor2ad37442008-05-06 19:49:01 -040048Possibly implement 32bit pcibios support.
Kevin O'Connorf076a3e2008-02-25 22:25:15 -050049
50Allow one to select adding 32 bit code to 0xf000 or in a separate
51location.
52
Kevin O'Connor65e63422008-07-19 14:12:32 -040053See if it is possible to handle interrupts while in 32bit mode.
54
Kevin O'Connorf13b0082008-08-17 11:26:42 -040055Add a kconfig style configuration program instead of requiring users
56to modify config.h.
57
Kevin O'Connor22dfb852008-06-07 15:24:19 -040058Look at integrating the lgpl vgabios into tree.
Kevin O'Connor65e63422008-07-19 14:12:32 -040059
60Look at usb booting specs. Look at possibly supporting usb
61keyboard/mice.
62
63Add a graphical boot splash screen?