Paolo Bonzini [Wed, 6 May 2015 10:38:29 +0000 (12:38 +0200)]
smm: ignore bits 16,18-31 of SMM revision ID
Bits 16-31 of the SMM revision ID are feature bits. We only need to
check that SMBASE relocation is supported, but do not care about other
features. In particular, this allows the SMM I/O instruction restart
feature to be present.
Unlike planned earlier virtio-vga will be compatible with
the qemu stdvga, with the framebuffer in bar 0. Drop the
virtio-vga chunk, update comment accordingly.
Kevin O'Connor [Wed, 3 Jun 2015 23:17:12 +0000 (19:17 -0400)]
build: Support "make VERSION=xyz" to override the default build version
Add a build option to explicitly set the version information compiled
into the seabios and seavgabios binaries. This may assist in
reproducible builds or to better link builds to distribution packages.
If the new "VERSION=" parameter is not provided then the default build
version remains unchanged.
At least AHCI on Intel on some machines claim to support only 6 ports
while in fact higher port numbers are present and work.
This is needed to access CD-ROM on PackardBell MS2290.
Both GRUB and Linux have a similar workaround.
Kevin O'Connor [Thu, 9 Apr 2015 21:15:23 +0000 (17:15 -0400)]
vgabios: Add config option for assembler fixups
Add a kconfig build option (CONFIG_VGA_FIXUP_ASM) to allow users to
build the vgabios without the complex assembler fixups that work
around emulator bugs.
In 0673b787 the QEMU SMP init code was changed to run in 32bit mode.
Unfortunately, the transition32 assembler function is not
multi-processor safe, because it modifies the global RTC index
register. This race condition led to sporadic failures when emulating
machines with a large number of processors.
This patch changes the entry_smp code to use a variant of transition32
that does not touch the RTC registers.
Kevin O'Connor [Wed, 31 Dec 2014 16:05:39 +0000 (11:05 -0500)]
xhci: Merge some xhci_xfer_x() functions into xhci_send_pipe()
The xhci_xfer_setup, xhci_xfer_data, xhci_xfer_status, and
xhci_xfer_normal functions are very similar - enhance xhci_xfer_queue
to reduce the boiler plate in the above functions. Merge the
resulting setup, data, and status code into the only function that
uses them - xhci_send_pipe().
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 07:07:37 +0000 (02:07 -0500)]
xhci: Merge xhci_send_control with xhci_send_bulk
Merge both the control and bulk pipe sending functions into one new
function: xhci_send_pipe(). This makes the xhci interface similar to
the other usb drivers.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 06:56:53 +0000 (01:56 -0500)]
uhci: Merge uhci_send_control with uhci_send_bulk
Merge both the control and bulk pipe sending functions into one new
function: uhci_send_pipe(). The two existing functions were similar,
and by merging them the resulting code supports more flexible control
transfers.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 06:55:48 +0000 (01:55 -0500)]
ohci: Merge ohci_send_control with ohci_send_bulk
Merge both the control and bulk pipe sending functions into one new
function: ohci_send_pipe(). The two existing functions were similar,
and by merging them the resulting code supports more flexible control
transfers.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 02:16:04 +0000 (21:16 -0500)]
ehci: Merge ehci_send_control with ehci_send_bulk
Merge both the control and bulk pipe sending functions into one new
function: ehci_send_pipe(). The two existing functions were similar,
and by merging them the resulting code supports more flexible control
transfers.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Thu, 1 Jan 2015 17:51:04 +0000 (12:51 -0500)]
ehci: Simplify fillTDbuffer() and rename
Simplify the calculation of the maximum transfer size per qtd,
simplify the fillTDbuffer() function so that it only fills the buffer
pointers, and rename fillTDbuffer() to ehci_fill_tdbuf().
Also, don't modify 'data' or 'datasize' so that usb_xfer_time() can
use 'datasize' at the end of the function.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 01:50:44 +0000 (20:50 -0500)]
ehci: No need to support td array wrapping
The maximum bulk transfer is 64K and 4 QTDs can always transfer 64K.
So, there is no need to support a transfer with more than 4 QTDs.
Build the entire transaction and then submit it in one operation to
simplify the code.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 6 Jan 2015 14:36:41 +0000 (09:36 -0500)]
vgabios: Support emulated text in gfx_read_char()
When emulating text mode on "coreboot framebuffer" SeaVGABIOS, return
a foreground and background attribute from gfx_read_char() and prefer
returning a space character (instead of null) on blank cells.
This also returns the foreground color (instead of always returning
zero) for regular graphics mode gfx_read_char() calls. This seems
fine as tests show other vgabios implementations also return various
values here.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Paolo Bonzini [Fri, 2 Jan 2015 17:32:25 +0000 (12:32 -0500)]
vgabios: implement read char in graphics mode
GWBasic relies on this, so implement it to enable some serious retrocomputing.
There is no better way to do it than trying to match all characters one by one
against the current font.
This makes it possible to actually do something in SCREEN 1 and SCREEN 2
(without it, you can use graphics in the programs but not in direct mode).
I couldn't find documentation for what to return as the attribute, but
experimenting with DOSBox suggests 0 (and GWBasic accepts it).
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Paolo Bonzini [Thu, 1 Jan 2015 20:27:46 +0000 (21:27 +0100)]
vgabios: fix graphics operation with Bochs VGA in non-DISPI modes
For legacy VGA modes that do not set the VBE_DISPI_ENABLED bit,
bochsvga_get_linelength returns 0. Thus all characters are squashed
into the first scanline. Fix this by falling back to stdvga for
the legacy modes.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Kevin O'Connor [Wed, 31 Dec 2014 06:42:38 +0000 (01:42 -0500)]
uhci: Increase bulk transfer STACKTDS to 16
Increase the number of simultaneous transfer descriptors that the
driver will build for uhci. The old value of 4 was a leftover from
when SeaBIOS had a tiny 512 byte extra stack - now that there is a 2K
extra stack there is plenty of space for additional descriptors.
Using a value of 16 should allow for an entire 1ms frame of bulk
transfer content to be setup in advance (assuming the max packet size
is 64 bytes).
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 31 Dec 2014 06:40:07 +0000 (01:40 -0500)]
uhci: Enable "depth" tree traversal for bulk transfers
Set the "depth" flag on bulk transactions. Since SeaBIOS doesn't use
bandwidth reclamation, without the depth flag the uhci controller will
only transfer one bulk packet per 1 ms frame. This results in a
maximum of 64 bytes per millisecond, which severely limits the
transfer rate.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Mon, 29 Dec 2014 18:06:23 +0000 (13:06 -0500)]
usb: Add support for OHCI bulk transfers
Support bulk transfers (usb drives) on OHCI USB controllers. Now that
there is support for 32bit only drive controllers, it is not that hard
to support disks on OHCI controllers, and it may be useful for some
old hardware.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Mon, 29 Dec 2014 15:29:34 +0000 (10:29 -0500)]
block: Check for read/write requests over 64K
The standard BIOS disk read/write request interface should never get a
request for more than 64K of data. Explicitly check for overly large
requests and reject them. This way, the low-level drivers do not need
to check for or attempt to handle very large requests.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Mon, 29 Dec 2014 15:15:57 +0000 (10:15 -0500)]
cdrom: Break up very large read requests into smaller requests
A cdrom boot image could be over 64K in size, but the low level
drivers may not support very large reads. If a large cdrom image is
found, issue multiple reads so that a read request over 64K is never
issued.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 16 Dec 2014 14:55:16 +0000 (09:55 -0500)]
sdcard: Initial support for SD cards on PCI SDHCI controllers on QEMU
This adds basic read/write support for SD cards emulated by QEMU.
This code is not expected to work on real hardware, because the
current controller and card initialization is not robust.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 16 Dec 2014 02:44:13 +0000 (21:44 -0500)]
Simplify README files - point to online documentation instead
The README file and README.CSM file have gotten a bit out of date.
Instead of maintaining technical information in the README file, point
new users to the SeaBIOS wiki.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 3 Dec 2014 17:53:01 +0000 (12:53 -0500)]
Eliminate FUNCFSEG - only force portions of inline asm to f-segment
The FUNCFSEG macro was introduced to force a C function into the
f-segment. This was needed for some C functions that used inline
assembler that contained some 16bit code. Instead of forcing the
entire C function into the f-segment, just force the small subset of
inline assembler into the f-segment.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 12 Nov 2014 22:49:33 +0000 (17:49 -0500)]
Fix build issue on gcc34
Older versions of gcc may not inline on_extra_stack() and thus cause a
link error when compiling in 32bit segmented mode. Test for MODE16
explicitly in stack_hop_back() to prevent the problem.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 5 Nov 2014 14:05:36 +0000 (09:05 -0500)]
Use an aligned stack offset when entering on the extra stack
The size of 'struct bregs' is not evenly divisible by four and where
the assembler placed a 'struct bregs' on the extra stack as part of
entering into the C functions it caused the C functions to run with a
non-aligned stack. It's technically not correct to use an unaligned
stack and it is certainly less efficient.
This patch avoids using BREGS_size (the sizeof struct bregs) and
instead introduces PUSHBREGS_size (the size of the general purpose
registers in struct bregs) in the assembler. Where the code actually
did use the %cs:%ip and flags, an extra 8 (instead of 6) bytes are
added to maintain a sane alignment.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Sat, 18 Oct 2014 02:15:42 +0000 (22:15 -0400)]
vgabios: Add software cursor capability
Add mechanism for drawing a cursor to the framebuffer to implement a
cursor in software. The timer interrupt is "hooked" so that the
cursor can blink. This can be useful for "coreboot native vga".
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Thu, 30 Oct 2014 16:09:19 +0000 (12:09 -0400)]
vgabios: Support emulating text mode attributes while in graphics mode
Add support for simple text mode attribute emulation while in graphics
mode. This improves text highlighting and background color on some
boot-loaders. Enable it only for CBVGA vgabios and only when a text
mode is requested.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Hannes Reinecke [Thu, 6 Nov 2014 13:31:56 +0000 (14:31 +0100)]
megasas: read addional PCI I/O bar
Some newer cards have the I/O bar at BAR 0, so we need to check
that, too, if we cannot get an address for BAR 2.
Without this patch the new 'megasas-gen2' emulation in qemu
is not detected.
Signed-off-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Kevin O'Connor [Thu, 23 Oct 2014 00:27:28 +0000 (20:27 -0400)]
vgabios: Set cursor shape fixes
The "cursor scaling" mode is determined by the first bit of the
video_ctl bda value, not the modeset_ctl value. Also, on a
set_scan_lines() call, the shape should be set to lines-3/lines-2 and
the shape must be set after char_height has been updated.
The existing cursor shape code matches the legacy "lgpl vgabios", but
the behavior doesn't match bios documentation nor observations with an
ati vga bios.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Sat, 18 Oct 2014 01:47:10 +0000 (21:47 -0400)]
vgabios: Don't pass vmode_g to vgafb_move_chars() / vgafb_clear_chars()
Now that a pointer to the current video mode info struct is cached in
memory, it's not necessary to pass the struct into the vgafb code.
The vgafb code can lookup the struct itself.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Sat, 18 Oct 2014 01:37:23 +0000 (21:37 -0400)]
vgabios: Cache a pointer to the current mode struct in the BDA
Cache a pointer to the current mode 'vgamode_s' struct in the BDA to
avoid doing a linear scan of all available vga modes when the struct
is needed.
This uses an additional two bytes in the BDA (at offset 0xbc). It's
possible this could conflict with some other software, but that seams
unlikely because that part of the BDA seems reserved for BIOS and
VGABIOS uses. (And neither SeaBIOS nor Bochs BIOS currently make use
of that area.)
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Sat, 18 Oct 2014 01:17:48 +0000 (21:17 -0400)]
vgabios: Don't declare custom internal BDA storage in std/bda.h
The vgabios uses storage in the BDA at offset 0xb9 for internal custom
storage (the contents do not appear to be part of any bios standard).
Move the description of this custom vgabios area from std/bda.h to
vgasrc/vgabios.h. Add two new macros (GET_BDA_EXT and SET_BDA_EXT).
This should make it more clear that the area is for custom internal
storage.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 21 Oct 2014 18:34:06 +0000 (14:34 -0400)]
Do full BREGS backup/restore for pmm, pnp, and irqentry_extrastack
Although these entry points only require backup and restore of the
registers that the C code clobbers, there is no harm in backing up
some additional registers. This allows the BREGS macros to be used
which makes the code a little more readable.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 21 Oct 2014 06:23:02 +0000 (02:23 -0400)]
Create assembler macros for saving and restoring 'struct bregs'
Create macros SAVEBREGS_POP_DSEAX and RESTOREBREGS_DSEAX for saving
and restoring the cpu state. These are similar to the existing
PUSHBREGS and POPBREGS macros.
This also fixes a bug in __farcall16 which inadvertently restored %ds
in %es and vice-versa.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 21 Oct 2014 03:06:04 +0000 (23:06 -0400)]
Use 32bit memcpy in int1587 when applicable
If the amount of data to be copied is an even number of four, then
copy the data in four byte chunks. This 32bit copy is more efficient,
in particular when copying to/from memory mapped io. This should
improve the performance of framebuffer draws in cbvga SeaVGABIOS.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Thu, 16 Oct 2014 18:03:39 +0000 (14:03 -0400)]
usb: Use usb_realloc_pipe for pipe alloc, update, and free.
Now that the usb controller drivers all support the realloc method,
use that for all pipe allocations, reallocations, and freeing. This
gives the driver more control over the pipe life cycle.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Sat, 11 Oct 2014 17:16:12 +0000 (13:16 -0400)]
Move a20 code from system.c and ps2port.h to x86.h
Although the a20 functionality was originally implemented in the ps2
controller, that is just a historical artifact. It's a core feature
of modern x86 cpus and the code is better located in the x86.h header.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Fri, 11 Apr 2014 15:20:41 +0000 (11:20 -0400)]
Implement call32 mechanism using SMIs.
Add support for jumping into 32bit mode using a System Management Mode
(SMM) handler. When available, this allows SeaBIOS to transition to
32bit mode even when called in vm86 mode. It will also prevent the
clobbering of the segment registers.
Currently, the SMM mode is only supported in QEMU when running in TCG
mode. Also, QEMU v2.1 (or later) is needed for it to work when in
vm86 mode.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 30 Sep 2014 16:55:54 +0000 (12:55 -0400)]
Fully restore 16bit state during call16_sloppy()
When transitioning back to 16bit mode from within call32(), restore
the full state (cmos index, gdt, fs/gs) in addition to restoring the
original stack.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 24 Sep 2014 21:48:28 +0000 (17:48 -0400)]
Move call16() functions from romlayout.S to inline assembler in stacks.c
Use inline assembler in call16 type functions instead of using
__call16() in romlayout.S.
Since call16() and call16big() are now only called with %ss==0 they do
not need to update the stack pointer. Only call16_sloppy() requires
the stack manipulation code.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Wed, 24 Sep 2014 19:58:12 +0000 (15:58 -0400)]
build: Support declaring 32bit C functions that must reside in the f-segment
Add support for a FUNCFSEG macro that will force a "32bit flat" C
function to be located in the f-segment. This is useful for 32bit
code with inline assembler that thunks to 16bit mode.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
Kevin O'Connor [Tue, 30 Sep 2014 04:11:38 +0000 (00:11 -0400)]
Simplify farcall16 code
With this change, farcall16() is only used for external API calls and
is only invoked from a 32bit mode entered directly via transition32.
farcall16big() is also only used for external API calls and is only
invoked from a 32bit mode entered directly via transition32.
call16_int() now calls _farcall16() directly, and it will use normal
16bit mode or big real mode as required.
Signed-off-by: Kevin O'Connor <kevin@koconnor.net>
projects / seabios.git / log