Kevin O'Connor [Fri, 23 Oct 2009 01:48:39 +0000 (21:48 -0400)]
Pass 'drive_s' pointer instead of driveid.
Pass a pointer to the drive_s struct instead of a driveid array offset.
Don't allocate ata drive's until a real drive found.
Introduce getDrive() and allocDrive() functions.
Kevin O'Connor [Wed, 14 Oct 2009 23:28:04 +0000 (19:28 -0400)]
fix low bits in ROM and I/O sizing
This cleans up handling of low bits during BAR sizing,
to match PCI spec requirements, and to use symbolic
constants from pci_regs.h
Issues fixed:
For ROM BARs, bit 0 is writeable (enable bit), which we not
only don't want to set, but it will stick and make us think
it's an I/O port resource.
Further, PCI spec defines the following bits as reserved:
- bit 1 in I/O BAR
- bits 10:1 in ROM BAR
and we should be careful and preserve any values there,
and should ignore anything we read from these registers.
Bits 3:2 in I/O BAR might be writeable, so it
is wrong to mask them when calculating BAR size.
Spec references:
See 6.2.5.1 for I/O and memory, and 6.2.5.2 for ROM,
6.1 for reserved bit handling;
pages 225, 228 and 214 in PCI spec revision 3.0.
Kevin O'Connor [Mon, 12 Oct 2009 16:54:56 +0000 (12:54 -0400)]
Use pmm_malloc() for internal allocations.
Using pmm_malloc permits free() to be used on internal allocations.
Rework low zone expanding so that it doesn't cause gcc v4.3 error.
Allow pmm_malloc to use ZoneTmpLow if ZoneTmpHigh isn't available.
Kevin O'Connor [Mon, 12 Oct 2009 14:34:51 +0000 (10:34 -0400)]
enable io/memory unconditionally
VGA adapters need to claim memory and i/o
transactions even if they do not have any
i/o or memory bars. E.g. PCI spec, page 297,
gives an example of such a device:
Programming interface 0000 0000b
VGA-compatible controller. Memory
addresses 0A 0000h through 0B
FFFFh. I/O addresses 3B0h to 3BBh
and 3C0h to 3DFh and all aliases of
these addresses.
While we could check for these devices and special-case them, it is
easier to fix this by enabling i/o and memory space unconditionally:
devices that do not support it will just ignore this setting.
Kevin O'Connor [Mon, 12 Oct 2009 14:09:15 +0000 (10:09 -0400)]
Add stubs for USB OHCI support.
Enable USB wrapper functions to call either ohci or ehci helpers.
Implement pci_config_maskw in place of pci_set_bus_master.
Introduce 'struct usb_pipe' in place of 'void *' for pipes.
Kevin O'Connor [Fri, 9 Oct 2009 02:09:02 +0000 (22:09 -0400)]
Add SRAT ACPI table support.
Take NUMA topology info from the QEMU firmware configuration interface
(number of nodes, node for each (V)CPU and amount of memory) and build
a SRAT table describing this topology for the guest OS. Handles more than
4 GB of RAM by including a hole for 32bit PCI memory mapping.
Kevin O'Connor [Fri, 9 Oct 2009 02:02:33 +0000 (22:02 -0400)]
Remove irq 9 from the pci interrupt link resources.
qemu can't share isa irqs (which is how the acpi sci interrupt is implemente
with the pci irqs, so remove the sci interrupt from the pci link interrupt
candidate list.
Kevin O'Connor [Fri, 9 Oct 2009 01:59:51 +0000 (21:59 -0400)]
Advertise pci irqs as active high in DSDT
Now that kvm emulates the ioapic polarity correctly, we must describe
the polarity correctly in the acpi tables. Otherwise pci interrupts won't
be delivered correctly.
Qemu implements PCI interrupts as active high, but OSes assume
that they are active low as per PCI spec. Qemu works without override
only because ioapic there doesn't implement polarity bit.
Kevin O'Connor [Thu, 8 Oct 2009 12:51:24 +0000 (08:51 -0400)]
Really fix linker issue with "debuginfo" variable.
The 16bit and 32bit code can't share debuginfo because it points to
code and must point to 16bit code in 16bit mode. Introduce two
separate variables - one for 32bit mode and one for 16bit mode.
Kevin O'Connor [Wed, 7 Oct 2009 23:44:23 +0000 (19:44 -0400)]
resolve memory device roll over reporting issues with >32G guests
The field within the Memory Device type 17 is only a word with the MSB being
used to report MB/KB. Thereby, a guest with 32G and greater would report
incorrect memory device information rolling over to 0.
This presents more than one memory device and associated memory structures
if the memory is larger than 16G
Kevin O'Connor [Tue, 29 Sep 2009 01:32:08 +0000 (21:32 -0400)]
Initial support for USB, UHCI, and USB Keyboards.
This adds preliminary support for USB controllers and keyboards.
Add support for UHCI controllers.
Add support for "HID" USB keyboards.
Also, fix bug in hexdump() - len need not be power of 4.
Kevin O'Connor [Sun, 20 Sep 2009 23:47:45 +0000 (19:47 -0400)]
Allow rom to grow beyond 64K.
If rom is over 64K then use part of e-segment for 32bit code.
Push 32bit code as high as it can go in the f-segment.
Do version building before layoutrom.py - this way layoutrom knows
full size of rom.
Make layoutrom.py build the full ld script - remove now unused ld
scripts that just imported the output of layoutrom.py.
Also, use "objdump" instead of "nm" - reduce toolchain requirements.
Enhance tools/checkrom.py so that it can pad bios.bin to size qemu is
happy with.
Also, add dependencies to build rules for local tools - if tool
changes automatically rerun it.
Make sure option roms don't overwrite the 32bit code (should the 32bit
code be in the e-segment).
Make sure shadow code works even if part of the code is in the
e-segment.
Kevin O'Connor [Sun, 20 Sep 2009 19:33:08 +0000 (15:33 -0400)]
Support sleeping until an irq fires, and use where applicable.
Add wait_irq() - it's more efficient than looping with cpu_relax().
Also, move kbd irq enables down - only kbd_command needs it.
Also, make some minor code layout improvements to kbd.c.
Kevin O'Connor [Sat, 12 Sep 2009 23:35:04 +0000 (19:35 -0400)]
Unify cd emulation access and main disk access code.
Add a new backend driver for cd emulation (DTYPE_CDEMU). This backend
driver now does the work of scheduling mis-sized reads.
Add mechanism for obtaining emulated drive geometry.
Extend disk_1308() to support cdrom emulation.
Use regular disk_13*() calls even for cdemu.
Also, unify the X_SECTOR_SIZE definitions.
Kevin O'Connor [Sat, 12 Sep 2009 17:20:14 +0000 (13:20 -0400)]
Add mechanism for describing internal drives in boot menu.
Allow each drive type to describe itself on the boot menu.
This fixes a bug where driveid was used in place of ataid in the boot
menu.
Also, expand descriptions for each drive type.
Don't overload the IPL type for BCVs - instead introduce new types.
Kevin O'Connor [Sun, 30 Aug 2009 23:31:22 +0000 (19:31 -0400)]
Ramdisk cleanups.
Use malloc code to allocate space for floppy image instead of
searching e820 map - this prevents a possible conflict with memory
allocated by pmm.
Check int 1587 return code.
Kevin O'Connor [Sun, 23 Aug 2009 15:56:55 +0000 (11:56 -0400)]
Add defines for floppy constants.
Add defines for several of the codes used in floppy.c.
Also, actually send the gap length in the floppy commands.
Also, adjust count in floppy_cmd instead of all callers.
Kevin O'Connor [Wed, 19 Aug 2009 02:38:49 +0000 (22:38 -0400)]
Merge cdrom 13xx handlers with main disk 13xx handlers.
Most of the cdrom_13 code did the same thing as disk_13 - so merge
them together.
Add definitions for 0xe0 (EXTSTART_CD) and 0x80 (EXTSTART_HD).
This changes the behavior for cdroms on 1315 and 1316 calls -
previously 1315 returned an error and set ah=2, now it sets ah=1
and doesn't return an error; previously 1316 returned ok, not it
returns EPARAM.
Kevin O'Connor [Mon, 17 Aug 2009 01:59:40 +0000 (21:59 -0400)]
Add support for using floppy images in CBFS.
Add new "ramdisk" type for disk accesses.
Extract out high-mem finding code from pmm into find_high_area().
Fix bug in GDB_BASE and GDT_LIMIT macros (wrong bit shifts).
Kevin O'Connor [Sun, 16 Aug 2009 16:21:44 +0000 (12:21 -0400)]
Enhance boot menu to allow selection of floppy to boot from.
The bootmenu can now reorder the first floppy drive.
Note - only the drive mapping is changed - floppy variables in the BDA
are still set depending on the real floppy index.
Kevin O'Connor [Sun, 16 Aug 2009 16:09:44 +0000 (12:09 -0400)]
Add floppy controllers to "drives" list also.
The Drives.drives list now contains floppies, harddrives, and cdroms.
Add mapping table for external/internal drive ids for floppies.
Rename CONFIG_FLOPPY_SUPPORT to CONFIG_FLOPPY (for consistency).
Be consistent with "driveid" and "floppyid" variable names.
Replace switch statements of drive parameters into a global array.
There are some externally visible changes with this patch:
- Some calls will now return EPARAM instead of ETIMEOUT (or ECHANGED)
- floppy_1301/1308 are now only available when regs->dl is valid
- floppy_1308/1315 return EPARAM on invalid drives
Kevin O'Connor [Wed, 12 Aug 2009 01:59:37 +0000 (21:59 -0400)]
Separate ATA code from generic disk code.
Move generic code from ata.c to new file block.c.
Rename atabits.h to ata.h and move ata header definitions from disk.h.
Rename ATA.channels to ATA_channels.
Rename ATA structure to Drives.
Support both CONFIG_DRIVES and CONFIG_ATA options.
Kevin O'Connor [Wed, 12 Aug 2009 00:43:38 +0000 (20:43 -0400)]
Densely populate ATA.devices list.
Populate ATA.devices in order of drives found; the array index no
longer correlates with the ATA.channels list.
Add cntl_id to device struct for finding the channel info.
Kevin O'Connor [Mon, 10 Aug 2009 04:06:37 +0000 (00:06 -0400)]
Misc ATA cleanups.
Remove some unused defines.
Rename ATA_TYPE_* to DTYPE_* and move from atabits.h to disk.h.
Rename ATA_TRANSLATION_* to TRANSLATION_* and move from atabits.h to disk.h.
Convert bios is-drive-ready call to use send_disk_op().
Only add atapi devices to the cdmap if they are cd/dvd roms.
Remove 'device' from struct ata_devices_s.
Kevin O'Connor [Sun, 9 Aug 2009 23:17:11 +0000 (19:17 -0400)]
Route disk_op commands by drive type - not by command.
Don't confuse op->command with low-level ata cmd->command.
Modify __send_disk_op to route commands by drive type.
Add commands for verify and seek operations.
Do ata resets via disk_op also.
Kevin O'Connor [Sun, 9 Aug 2009 22:06:40 +0000 (18:06 -0400)]
Remove unused cdemu ATA code.
Remove the low-level ATA code for short reads on cdroms now that cd
emulation is done at higher-level.
Also, remove sector_count from ebda - ata code updates op->count now.
Kevin O'Connor [Sun, 9 Aug 2009 21:25:19 +0000 (17:25 -0400)]
Implement cdrom disk emulation at high-level instead of in low-level ATA.
Add a 2K buffer to the ebda to allow for cdrom 512 byte vs 2048 byte
sector emulation.
For unaliagned cdemu reads, just make multiple cdrom reads instead of
using ata specific code for short reads.
Also, define cdemu virtual sectors using struct chs_s, and update
legacy_lba() func to take pointer to a chs_s struct.
Kevin O'Connor [Sun, 9 Aug 2009 17:08:21 +0000 (13:08 -0400)]
Separate cdemu disk accesses from regular disk accesses.
Breakup basic_access() into basic_access, cdemu_access, and legacy_lba.
Also, check for verify and seek calls in __send_disk_op instead of all
callers.
Also, send_disk_op returns sector count in dop.count instead of via ebda.
Kevin O'Connor [Sun, 9 Aug 2009 15:32:00 +0000 (11:32 -0400)]
Use regular CD reading to load initial CDROM boot image.
Use cdrom_read() instead of cdrom_read_512() for loading the initial
image. The initial image will always start on a 2K block, and it
should be okay to load extra sectors if the sector count doesn't
end on a 2K boundary.
Kevin O'Connor [Sun, 2 Aug 2009 16:33:58 +0000 (12:33 -0400)]
Fix some PMM corner-cases.
Make sure not to divide by zero when calculating usage percent in
debug messages.
Make sure high memory zones are aligned; make sure zone is big enough
even after loss due to alignment.
Be sure to reset ZoneTmpHigh and PMMAllocs on failure paths.
Fix corner case in pmm_free() where an allocation might not get freed
due to ordering of pmmalloc_s structs.
Kevin O'Connor [Sat, 1 Aug 2009 15:45:37 +0000 (11:45 -0400)]
Add Post Memory Manager (PMM) support.
Complete the initial implementation of PMM.
Default new PMM code to be enabled.
Move malloc code from memmap.c to pmm.c.
Define new malloc zones for PMM (ZoneTmpLow and ZoneTmpHigh).
Change default READ/WRITE_SEG macros to use 32bit pointers (the 16bit
PMM code use "big real" mode and requires 32bit accesses).
Allow pmm allocations to occur during bcv executions also.
Add low-memory clearing before boot.
Also, align the default f-seg memory.
Kevin O'Connor [Wed, 29 Jul 2009 23:20:03 +0000 (19:20 -0400)]
mptable and madt irq override
Implement irq override support for timer interrupts. This matches what
QEMU+BOCHS has been doing for the latest 8 months, and is also what
real hardware does.
Kevin O'Connor [Wed, 29 Jul 2009 00:46:03 +0000 (20:46 -0400)]
Fix coreboot bios table copying by delaying to after memory scan.
Delay coreboot bios table scan after memory scan and malloc setup is
complete.
Also, fix sign underflow in malloc memory available check.
Add check to ensure high bios area is big enough for zone.
Add more debug info to malloc setup/finalize.
Kevin O'Connor [Sun, 26 Jul 2009 23:16:09 +0000 (19:16 -0400)]
Add support for enabling ram in 0xc0000-0xf0000 area.
Enhance shadow ram support to enable read/write of option rom area.
This enables support for option roms that modify themselves.
Support copying the roms before enabling shadowing, to work around the
qemu implementation of ram shadowing.
Rename next_rom to RomEnd and export it.
Support locking ram being used for optionroms before booting.
Don't bother and'ing/or'ing 0x59 register - it's usage is well defined.