If some bug happens during early bootup, the mem_cons might not have
been initialized yet. In that case use default PARISC_SERIAL_CONSOLE
to show something.
Signed-off-by: Helge Deller <deller@gmx.de>
The existing esp-scsi state machine checks for the STAT_TC bit to exit state 1
but in the case where there is no data phase, a non-DMA command is executed
which doesn't set STAT_TC. This only works because QEMU currently always sets
STAT_TC just after issuing every SCSI command.
Update the esp-scsi state machine so that in the case where there is no data
phase, we immediately execute CMD_ICCS instead of waiting for STAT_TC to be
set which will never happen with a non-DMA CMD_SELATN command.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-ID: <20230807065300.366070-4-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The ESP SELATN command used to send SCSI commands from the ESP to the SCSI bus
is not a DMA command and therefore does not affect the STAT_TC bit. The only
reason this works at all is due to a bug in QEMU which (currently) always
updates the STAT_TC bit in ESP_RSTAT regardless of the state of the ESP_CMD_DMA
bit.
According to the NCR datasheet [1] the INTR_BS/INTR_FC bits are set when the
SELATN command has completed, so update the existing logic to check for these
bits in ESP_RINTR instead. Note that the read of ESP_RINTR needs to be
restricted to state == 0 as reading ESP_RINTR resets the ESP_RSTAT register
which breaks the STAT_TC check when state == 1.
This commit also includes an extra read of ESP_INTR to clear all the interrupt
bits before submitting the SELATN command to ensure that we don't accidentally
immediately progress to the data phase handling logic where ESP_RINTR bits have
already been set by a previous ESP command.
[1] "NCR 53C94, 53C95, 53C96 Advanced SCSI Controller"
NCR_53C94_53C95_53C96_Data_Sheet_Feb90.pdf
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20230807065300.366070-3-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The ESP FIFO is used as a buffer for DMA requests and so isn't guaranteed to
be empty in the case of SCSI errors or a mixed DMA/non-DMA request. Flush the
FIFO before sending a SCSI command to guarantee that it is correctly
positioned at the start of the FIFO.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-ID: <20230807065300.366070-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
SeaBIOS has no information about the lower implementation details for
Block-TLB (or generic TLB) support, so it simply hands over the
parameters in the CPU registers provided by the operating system to the
PDC_BLOCK_TLB PDC function to QEMU.
Calling QEMU happens via a diagnostics instruction (diag 0x100) which
QEMU will respond to. Older QEMU versions without the BTLB support will
simply log that they ignored the diag instruction, and SeaBIOS will
return PDC_BAD_OPTION for this PDC call in that case.
Signed-off-by: Helge Deller <deller@gmx.de>
Fixes
- boot of HP-UX with SMP, and
- reboot of Linux and HP-UX with SMP
Enhancements:
- show qemu version in boot menu
- adds exit menu entry in boot menu to quit emulation
- allow to trace PCD_CHASSIS codes more specifically
Signed-off-by: Helge Deller <deller@gmx.de>
Do not store the boot args in the pad0 field of PAGE0.
This breaks booting HP-UX SMP kernels.
Use pad608 instead.
Signed-off-by: Helge Deller <deller@gmx.de>
Fixes boot of the Debian-12 installation CD-ROM
The size of the ramdisc of Debian-12 is 17123328 bytes and occupies 8361
CD-ROM sectors (of 2048 bytes each). As such the palo bootloader tries to
read 8361 sectors, which fails with DISK_RET_EBADTRACK error because the
emulated LSI SCSI driver only supports (due to hardware limitation)
reading 8192 sectors at once.
Fix it in the SeaBIOS firmware by limiting the sectors which can be read
at once to 8192-16 sectors. The palo bootloader will then simply do
additional calls to read the remaining sectors.
This firmware bug didn't show up with Debian-11 because that ramdisc
was smaller and below the 8192 sector limit.
Signed-off-by: Helge Deller <deller@gmx.de>
Fixes the serial port emulation
Qemu versions which request a SEABIOS_HPPA_VERSION < 6 have the bug that
they use the DINO UART instead of the LASI UART as serial port #0.
Staring with SEABIOS_HPPA_VERSION 6 the serial ports are now emulated as
on physical hardware, with LASI UART being serial port #0.
This patch adds a compat patch, which will use the DINO UART port
for console I/O if it detects that qemu requests a SEABIOS_HPPA_VERSION
< 6.
Signed-off-by: Helge Deller <deller@gmx.de>
Add this option to avoid some warnings like:
src/tcgbios.c: In function 'tpm_interrupt_handler32':
src/tcgbios.c:1721:45: warning: taking address of packed member of 'struct bregs' may result
in an unaligned pointer value [-Waddress-of-packed-member]
Signed-off-by: Helge Deller <deller@gmx.de>
Prevent usage of the hardcoded PARISC_SERIAL_CONSOLE value and instead
check the given hpa and provide the correct serial port name.
Signed-off-by: Helge Deller <deller@gmx.de>
The structs mem_kbd and mem_cons holds the HPA of the serial ports
used for console I/O, so use those instead of hardcoding to
PARISC_SERIAL_CONSOLE.
There is no functional change with this, but now it complies to how
the values from PAGE0 should be used.
Signed-off-by: Helge Deller <deller@gmx.de>
On real hardware, Linux detects the UART which is built-into LASI as serial port #1,
and the UART which is built-into DINO as serial port #2.
But to now qemu and SeaBIOS used the UART on DINO as serial port #1,
which breaks qemu's possibility to use the "-serial" option to
pass-through host serial ports to the guest.
Fix this by swapping the addresses of PORT_SERIAL1 and PORT_SERIAL2.
Signed-off-by: Helge Deller <deller@gmx.de>
On german keyboards, the Z key is positioned at the same location
as Y on english keyboards. German users may thus type Z instead of
Y. So this patch allows Z as synonym for Y.
Signed-off-by: Helge Deller <deller@gmx.de>
New features and fixes:
* Fixed SeaBIOS PS/2 keyboard support.
This now allows to use the qemu "-boot menu=on" option to start the
firmware boot selection menu and provide a working menu with keyboard on
serial (qemu "-nographic" option) and PS/2 keyboard (qemu "-graphic"
option) on graphical screen.
Finally add the real boot menu.
If qemu is started with the "-boot menu=on" option, the PDC boot menu is
shown and the user can then boot any of the available SCSI disc/CDs.
Just type "BOOT" (to boot the default device), or "BOOT FWSCSI.4" (to boot
SCSI device #4).
Signed-off-by: Helge Deller <deller@gmx.de>
lasips2_kbd_in() is used to read input from an emulated PS/2 keyboard.
Since there is no interrupt support durig PDC/IODC we need to poll the
PS/2 status flags to see if a new key was pressed. In order to be able
to poll regularily, exit the loop to read already pressed chars as soon
as possible. This can be done by calling handle_16() with register ah =
0x11 to check if keys are queued up and exit if none are available.
This patch fixes the problem that keys couldn't be read when using a
graphical output (artist graphics). This patch fixes palo and other
bootloaders to be able to use input from the emulated PS/2 keyboard.
Signed-off-by: Helge Deller <deller@gmx.de>
New features and fixes:
* Fix firmware rendenzvous code to clear all pending external intrrupts
before entering the waiting loop.
* STI firmware now contains additional fonts built-in, which
can be selected with qemu command-line options:
-fw_cfg opt/font,string=1 - a HP 8x16 font
-fw_cfg opt/font,string=2 - a HP 6x13 font
-fw_cfg opt/font,string=3 - a HP 10x20 font
-fw_cfg opt/font,string=4 - a Linux 16x32 font
Signed-off-by: Helge Deller <deller@gmx.de>
Add additional fonts to STI ROM.
They can be selected with qemu command-line options:
-fw_cfg opt/font,string=1 - a HP 8x16 font
-fw_cfg opt/font,string=2 - a HP 6x13 font
-fw_cfg opt/font,string=3 - a HP 10x20 font
-fw_cfg opt/font,string=4 - a Linux 16x32 font
Signed-off-by: Helge Deller <deller@gmx.de>
Fix the PDC rendenzvous code to clear all pending external intrrupts
before entering the waiting loop.
Without clearing them before, the CPU was waken up immediately again.
This fixes the CPU hotplug on Linux, which can be run with
chcpu -d 2 # to disable CPU2
chcpu -e 2 # to wake up CPU2
Signed-off-by: Helge Deller <deller@gmx.de>
New features and fixes:
* Allow up to 16 CPUs
* Add TOC button support:
To trigger a TOC, execute "nmi" in the qemu monitor (Ctrl-A C)
* New opt/hostid fw_cfg option to change hostid:
-fw_cfg opt/hostid,string=334455
* Add opt/console fw_cfg option to select default console:
-fw_cfg opt/console,string=serial
-fw_cfg opt/console,string=graphics
* Add Linux TER16x32 font to STI firmware:
-fw_cfg opt/font,string=2
* Leave IRQs disabled after rendevouz
Signed-off-by: Helge Deller <deller@gmx.de>
The new 16x32 font can be selected as default firmware font with
this Qemu option:
-fw_cfg opt/font,string=2
Any other values will choose the current HP 8x16 font.
Signed-off-by: Helge Deller <deller@gmx.de>
By default the firmware will choose the (virtual) graphics card
as firmware console if the graphics card is available.
If not it defaults to serial console.
In some configurations it's useful to select the serial console
als default output device even if the graphics card is available.
For that, this patch adds the opt/console fw_cfg qemu option.
Qemu option usage to switch the firmare to use the serial or graphics
default output:
-fw_cfg opt/console,string=serial
-fw_cfg opt/console,string=graphics
Signed-off-by: Helge Deller <deller@gmx.de>
Move the HPA for MEMORY_HPA out of the address space of the 16th CPU,
and fix the PDC_COPROC_CFG PDC function to report functional FPUs for
all configured CPUs. The documentation specifies that cr10 can only
report for 8 CPUs, but that seems outdated as both HP-UX 10 and HP-UX 11
accepts this change.
Signed-off-by: Helge Deller <deller@gmx.de>
Add support for an emulated TOC/NMI button.
Almost all PA-RISC machines have either a button that is labeled with
'TOC' or a BMC/GSP function to trigger a TOC. TOC is a non-maskable
interrupt that is sent to the processor. This can be used for
diagnostic purposes like obtaining a stack trace/register dump or to
enter KDB/KGDB in Linux.
This patch adds support for such an emulated TOC button.
This patch writes the CPU registers into PIM (processor internal
memmory) for later analysis. It uses the CPU "shadow registers", the
IAOQ-back and IASQ-back values are provided by qemu in r24/r25. This
patch uses the new qemu aritificial opcode "getshadowregs" (0xfffdead2)
which restores the original values of the shadow registers.
To trigger a TOC, switch to the "qemu monitor" with Ctrl-A C, and type
in the command "nmi". After the TOC started the OS-debugger, exit the
qemu monitor with Ctrl-A C.
Signed-off-by: Helge Deller <deller@gmx.de>
Allow the user to select another hostid value.
In HP/UX the "uname -i" command shows the number,
and in Linux it's visible in the /proc/cpuinfo output.
The default hostid number is 2006243326, any other value can
be selected with this qemu command line option:
-fw_cfg opt/hostid,string=334455
Signed-off-by: Helge Deller <deller@gmx.de>
This version adds fixes which allow choosing the primary boot device.
Either via qemu bootindex parameter, e.g.:
-device scsi-hd,drive=drive1,bus=scsi0.0,channel=0,scsi-id=6,lun=0,bootindex=1
-drive file=image.img,if=none,id=drive1
or via qemu boot order option to select the SCSI ID, e.g.:
-boot order=g-m: machine implementation dependent drives (SCSI ID)
where "g" means SCSI ID 0, "h" is SCSI ID 1 and so on.
Signed-off-by: Helge Deller <deller@gmx.de>
Enable firmware debug info at startup with this option:
-fw_cfg pdc_debug,string=<val>
<val> is an integer with this bitmask:
DEBUG_PDC 0x0001, enables debug info for PDC calls
DEBUG_IODC 0x0002, enables debug info for IODC calls
Signed-off-by: Helge Deller <deller@gmx.de>
Make the malloc area initialization the first step after bootup.
It's important, because the first mallocs otherwise ends up in lower
memory area (PAGE0 from parisc) which will gets overwritten later.
That change fixes the fw_cfg() interface as well. Before that change
fw_cfg got initialized and zeroed-out shortly later in the boot process.
And now the initialization of the boot structures succeeds too, so we
can run boot_init() directly after setting up malloc and fw_cfg, and the
bootindex parameter from qemu suddenly works out of the box.
Signed-off-by: Helge Deller <deller@gmx.de>
