* [PATCH v2 01/12] macfb: handle errors that occur during realize
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-05 15:34 ` Philippe Mathieu-Daudé
2021-10-04 21:19 ` [PATCH v2 02/12] macfb: fix invalid object reference in macfb_common_realize() Mark Cave-Ayland
` (10 subsequent siblings)
11 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
Make sure any errors that occur within the macfb realize chain are detected
and handled correctly to prevent crashes and to ensure that error messages are
reported back to the user.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 23 +++++++++++++++++------
1 file changed, 17 insertions(+), 6 deletions(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 76808b69cc..2ec25c5d6f 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -343,14 +343,14 @@ static const GraphicHwOps macfb_ops = {
.gfx_update = macfb_update_display,
};
-static void macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
+static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
{
DisplaySurface *surface;
if (s->depth != 1 && s->depth != 2 && s->depth != 4 && s->depth != 8 &&
s->depth != 16 && s->depth != 24) {
error_setg(errp, "unknown guest depth %d", s->depth);
- return;
+ return false;
}
s->con = graphic_console_init(dev, 0, &macfb_ops, s);
@@ -359,18 +359,20 @@ static void macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
if (surface_bits_per_pixel(surface) != 32) {
error_setg(errp, "unknown host depth %d",
surface_bits_per_pixel(surface));
- return;
+ return false;
}
memory_region_init_io(&s->mem_ctrl, OBJECT(dev), &macfb_ctrl_ops, s,
"macfb-ctrl", 0x1000);
memory_region_init_ram_nomigrate(&s->mem_vram, OBJECT(s), "macfb-vram",
- MACFB_VRAM_SIZE, errp);
+ MACFB_VRAM_SIZE, &error_abort);
s->vram = memory_region_get_ram_ptr(&s->mem_vram);
s->vram_bit_mask = MACFB_VRAM_SIZE - 1;
vmstate_register_ram(&s->mem_vram, dev);
memory_region_set_coalescing(&s->mem_vram);
+
+ return true;
}
static void macfb_sysbus_realize(DeviceState *dev, Error **errp)
@@ -378,7 +380,10 @@ static void macfb_sysbus_realize(DeviceState *dev, Error **errp)
MacfbSysBusState *s = MACFB(dev);
MacfbState *ms = &s->macfb;
- macfb_common_realize(dev, ms, errp);
+ if (!macfb_common_realize(dev, ms, errp)) {
+ return;
+ }
+
sysbus_init_mmio(SYS_BUS_DEVICE(s), &ms->mem_ctrl);
sysbus_init_mmio(SYS_BUS_DEVICE(s), &ms->mem_vram);
}
@@ -391,8 +396,14 @@ static void macfb_nubus_realize(DeviceState *dev, Error **errp)
MacfbState *ms = &s->macfb;
ndc->parent_realize(dev, errp);
+ if (*errp) {
+ return;
+ }
+
+ if (!macfb_common_realize(dev, ms, errp)) {
+ return;
+ }
- macfb_common_realize(dev, ms, errp);
memory_region_add_subregion(&nd->slot_mem, DAFB_BASE, &ms->mem_ctrl);
memory_region_add_subregion(&nd->slot_mem, VIDEO_BASE, &ms->mem_vram);
}
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* Re: [PATCH v2 01/12] macfb: handle errors that occur during realize
2021-10-04 21:19 ` [PATCH v2 01/12] macfb: handle errors that occur during realize Mark Cave-Ayland
@ 2021-10-05 15:34 ` Philippe Mathieu-Daudé
0 siblings, 0 replies; 26+ messages in thread
From: Philippe Mathieu-Daudé @ 2021-10-05 15:34 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel, laurent
On 10/4/21 23:19, Mark Cave-Ayland wrote:
> Make sure any errors that occur within the macfb realize chain are detected
> and handled correctly to prevent crashes and to ensure that error messages are
> reported back to the user.
>
> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
> Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
> ---
> hw/display/macfb.c | 23 +++++++++++++++++------
> 1 file changed, 17 insertions(+), 6 deletions(-)
> @@ -391,8 +396,14 @@ static void macfb_nubus_realize(DeviceState *dev, Error **errp)
> MacfbState *ms = &s->macfb;
>
> ndc->parent_realize(dev, errp);
> + if (*errp) {
> + return;
> + }
Sorry for being picky, but this is one fix (one patch),
> + if (!macfb_common_realize(dev, ms, errp)) {
> + return;
> + }
and this is another fix (another patch, including the bool).
>
> - macfb_common_realize(dev, ms, errp);
> memory_region_add_subregion(&nd->slot_mem, DAFB_BASE, &ms->mem_ctrl);
> memory_region_add_subregion(&nd->slot_mem, VIDEO_BASE, &ms->mem_vram);
> }
>
I'd rather split this in 2, but I'd be OK if you just reword
the description, either split or reworded:
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
^ permalink raw reply [flat|nested] 26+ messages in thread
* [PATCH v2 02/12] macfb: fix invalid object reference in macfb_common_realize()
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 01/12] macfb: handle errors that occur during realize Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 03/12] macfb: fix overflow of color_palette array Mark Cave-Ayland
` (9 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
During realize memory_region_init_ram_nomigrate() is used to initialise the RAM
memory region used for the framebuffer but the owner object reference is
incorrect since MacFbState is a typedef and not a QOM type.
Change the memory region owner to be the corresponding DeviceState to fix the
issue and prevent random crashes during macfb_common_realize().
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: 8ac919a0654 ("hw/m68k: add Nubus macfb video card")
Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 2ec25c5d6f..b363bab889 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -365,7 +365,7 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
memory_region_init_io(&s->mem_ctrl, OBJECT(dev), &macfb_ctrl_ops, s,
"macfb-ctrl", 0x1000);
- memory_region_init_ram_nomigrate(&s->mem_vram, OBJECT(s), "macfb-vram",
+ memory_region_init_ram_nomigrate(&s->mem_vram, OBJECT(dev), "macfb-vram",
MACFB_VRAM_SIZE, &error_abort);
s->vram = memory_region_get_ram_ptr(&s->mem_vram);
s->vram_bit_mask = MACFB_VRAM_SIZE - 1;
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 03/12] macfb: fix overflow of color_palette array
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 01/12] macfb: handle errors that occur during realize Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 02/12] macfb: fix invalid object reference in macfb_common_realize() Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-05 6:34 ` Laurent Vivier
2021-10-04 21:19 ` [PATCH v2 04/12] macfb: use memory_region_init_ram() in macfb_common_realize() for the framebuffer Mark Cave-Ayland
` (8 subsequent siblings)
11 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
The palette_current index counter has a maximum size of 256 * 3 to cover a full
color palette of 256 RGB entries. Linux assumes that the palette_current index
wraps back around to zero after writing 256 RGB entries so ensure that
palette_current is reset at this point to prevent data corruption within
MacfbState.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
---
hw/display/macfb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index b363bab889..39dab49026 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -303,7 +303,9 @@ static void macfb_ctrl_write(void *opaque,
s->palette_current = 0;
break;
case DAFB_LUT:
- s->color_palette[s->palette_current++] = val;
+ s->color_palette[s->palette_current] = val;
+ s->palette_current = (s->palette_current + 1) %
+ ARRAY_SIZE(s->color_palette);
if (s->palette_current % 3) {
macfb_invalidate_display(s);
}
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* Re: [PATCH v2 03/12] macfb: fix overflow of color_palette array
2021-10-04 21:19 ` [PATCH v2 03/12] macfb: fix overflow of color_palette array Mark Cave-Ayland
@ 2021-10-05 6:34 ` Laurent Vivier
0 siblings, 0 replies; 26+ messages in thread
From: Laurent Vivier @ 2021-10-05 6:34 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
> The palette_current index counter has a maximum size of 256 * 3 to cover a full
> color palette of 256 RGB entries. Linux assumes that the palette_current index
> wraps back around to zero after writing 256 RGB entries so ensure that
> palette_current is reset at this point to prevent data corruption within
> MacfbState.
>
> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
> ---
> hw/display/macfb.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
> index b363bab889..39dab49026 100644
> --- a/hw/display/macfb.c
> +++ b/hw/display/macfb.c
> @@ -303,7 +303,9 @@ static void macfb_ctrl_write(void *opaque,
> s->palette_current = 0;
> break;
> case DAFB_LUT:
> - s->color_palette[s->palette_current++] = val;
> + s->color_palette[s->palette_current] = val;
> + s->palette_current = (s->palette_current + 1) %
> + ARRAY_SIZE(s->color_palette);
> if (s->palette_current % 3) {
> macfb_invalidate_display(s);
> }
>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
^ permalink raw reply [flat|nested] 26+ messages in thread
* [PATCH v2 04/12] macfb: use memory_region_init_ram() in macfb_common_realize() for the framebuffer
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (2 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 03/12] macfb: fix overflow of color_palette array Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 05/12] macfb: add trace events for reading and writing the control registers Mark Cave-Ayland
` (7 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
Currently macfb_common_realize() defines the framebuffer RAM memory region as
being non-migrateable but then immediately registers it for migration. Replace
memory_region_init_ram_nomigrate() with memory_region_init_ram() which is clearer
and does exactly the same thing.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 39dab49026..f88f5a6523 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -367,11 +367,10 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
memory_region_init_io(&s->mem_ctrl, OBJECT(dev), &macfb_ctrl_ops, s,
"macfb-ctrl", 0x1000);
- memory_region_init_ram_nomigrate(&s->mem_vram, OBJECT(dev), "macfb-vram",
- MACFB_VRAM_SIZE, &error_abort);
+ memory_region_init_ram(&s->mem_vram, OBJECT(dev), "macfb-vram",
+ MACFB_VRAM_SIZE, &error_abort);
s->vram = memory_region_get_ram_ptr(&s->mem_vram);
s->vram_bit_mask = MACFB_VRAM_SIZE - 1;
- vmstate_register_ram(&s->mem_vram, dev);
memory_region_set_coalescing(&s->mem_vram);
return true;
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 05/12] macfb: add trace events for reading and writing the control registers
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (3 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 04/12] macfb: use memory_region_init_ram() in macfb_common_realize() for the framebuffer Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 06/12] macfb: implement mode sense to allow display type to be detected Mark Cave-Ayland
` (6 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 8 +++++++-
hw/display/trace-events | 4 ++++
2 files changed, 11 insertions(+), 1 deletion(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index f88f5a6523..1128a51c98 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -20,6 +20,7 @@
#include "qapi/error.h"
#include "hw/qdev-properties.h"
#include "migration/vmstate.h"
+#include "trace.h"
#define VIDEO_BASE 0x00001000
#define DAFB_BASE 0x00800000
@@ -289,7 +290,10 @@ static uint64_t macfb_ctrl_read(void *opaque,
hwaddr addr,
unsigned int size)
{
- return 0;
+ uint64_t val = 0;
+
+ trace_macfb_ctrl_read(addr, val, size);
+ return val;
}
static void macfb_ctrl_write(void *opaque,
@@ -311,6 +315,8 @@ static void macfb_ctrl_write(void *opaque,
}
break;
}
+
+ trace_macfb_ctrl_write(addr, val, size);
}
static const MemoryRegionOps macfb_ctrl_ops = {
diff --git a/hw/display/trace-events b/hw/display/trace-events
index f03f6655bc..6c460aaa4c 100644
--- a/hw/display/trace-events
+++ b/hw/display/trace-events
@@ -167,3 +167,7 @@ sm501_disp_ctrl_read(uint32_t addr, uint32_t val) "addr=0x%x, val=0x%x"
sm501_disp_ctrl_write(uint32_t addr, uint32_t val) "addr=0x%x, val=0x%x"
sm501_2d_engine_read(uint32_t addr, uint32_t val) "addr=0x%x, val=0x%x"
sm501_2d_engine_write(uint32_t addr, uint32_t val) "addr=0x%x, val=0x%x"
+
+# macfb.c
+macfb_ctrl_read(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx64 " value 0x%"PRIx64 " size %d"
+macfb_ctrl_write(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx64 " value 0x%"PRIx64 " size %d"
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 06/12] macfb: implement mode sense to allow display type to be detected
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (4 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 05/12] macfb: add trace events for reading and writing the control registers Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 07/12] macfb: add qdev property to specify display type Mark Cave-Ayland
` (5 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
The MacOS toolbox ROM uses the monitor sense to detect the display type and then
offer a fixed set of resolutions and colour depths accordingly. Implement the
monitor sense using information found in Apple Technical Note HW26: "Macintosh
Quadra Built-In Video" along with some local experiments.
Since the default configuration is 640 x 480 with 8-bit colour then hardcode
the sense register to return MACFB_DISPLAY_VGA for now.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 117 ++++++++++++++++++++++++++++++++++++-
hw/display/trace-events | 2 +
include/hw/display/macfb.h | 20 +++++++
3 files changed, 137 insertions(+), 2 deletions(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 1128a51c98..6e485d7aef 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -28,8 +28,66 @@
#define MACFB_PAGE_SIZE 4096
#define MACFB_VRAM_SIZE (4 * MiB)
-#define DAFB_RESET 0x200
-#define DAFB_LUT 0x213
+#define DAFB_MODE_SENSE 0x1c
+#define DAFB_RESET 0x200
+#define DAFB_LUT 0x213
+
+
+/*
+ * Quadra sense codes taken from Apple Technical Note HW26:
+ * "Macintosh Quadra Built-In Video". The sense codes and
+ * extended sense codes have different meanings:
+ *
+ * Sense:
+ * bit 2: SENSE2 (pin 10)
+ * bit 1: SENSE1 (pin 7)
+ * bit 0: SENSE0 (pin 4)
+ *
+ * 0 = pin tied to ground
+ * 1 = pin unconnected
+ *
+ * Extended Sense:
+ * bit 2: pins 4-10
+ * bit 1: pins 10-7
+ * bit 0: pins 7-4
+ *
+ * 0 = pins tied together
+ * 1 = pins unconnected
+ *
+ * Reads from the sense register appear to be active low, i.e. a 1 indicates
+ * that the pin is tied to ground, a 0 indicates the pin is disconnected.
+ *
+ * Writes to the sense register appear to activate pulldowns i.e. a 1 enables
+ * a pulldown on a particular pin.
+ *
+ * The MacOS toolbox appears to use a series of reads and writes to first
+ * determine if extended sense is to be used, and then check which pins are
+ * tied together in order to determine the display type.
+ */
+
+typedef struct MacFbSense {
+ uint8_t type;
+ uint8_t sense;
+ uint8_t ext_sense;
+} MacFbSense;
+
+static MacFbSense macfb_sense_table[] = {
+ { MACFB_DISPLAY_APPLE_21_COLOR, 0x0, 0 },
+ { MACFB_DISPLAY_APPLE_PORTRAIT, 0x1, 0 },
+ { MACFB_DISPLAY_APPLE_12_RGB, 0x2, 0 },
+ { MACFB_DISPLAY_APPLE_2PAGE_MONO, 0x3, 0 },
+ { MACFB_DISPLAY_NTSC_UNDERSCAN, 0x4, 0 },
+ { MACFB_DISPLAY_NTSC_OVERSCAN, 0x4, 0 },
+ { MACFB_DISPLAY_APPLE_12_MONO, 0x6, 0 },
+ { MACFB_DISPLAY_APPLE_13_RGB, 0x6, 0 },
+ { MACFB_DISPLAY_16_COLOR, 0x7, 0x3 },
+ { MACFB_DISPLAY_PAL1_UNDERSCAN, 0x7, 0x0 },
+ { MACFB_DISPLAY_PAL1_OVERSCAN, 0x7, 0x0 },
+ { MACFB_DISPLAY_PAL2_UNDERSCAN, 0x7, 0x6 },
+ { MACFB_DISPLAY_PAL2_OVERSCAN, 0x7, 0x6 },
+ { MACFB_DISPLAY_VGA, 0x7, 0x5 },
+ { MACFB_DISPLAY_SVGA, 0x7, 0x5 },
+};
typedef void macfb_draw_line_func(MacfbState *s, uint8_t *d, uint32_t addr,
@@ -253,6 +311,50 @@ static void macfb_invalidate_display(void *opaque)
memory_region_set_dirty(&s->mem_vram, 0, MACFB_VRAM_SIZE);
}
+static uint32_t macfb_sense_read(MacfbState *s)
+{
+ MacFbSense *macfb_sense;
+ uint8_t sense;
+
+ macfb_sense = &macfb_sense_table[MACFB_DISPLAY_VGA];
+ if (macfb_sense->sense == 0x7) {
+ /* Extended sense */
+ sense = 0;
+ if (!(macfb_sense->ext_sense & 1)) {
+ /* Pins 7-4 together */
+ if (~s->sense & 3) {
+ sense = (~s->sense & 7) | 3;
+ }
+ }
+ if (!(macfb_sense->ext_sense & 2)) {
+ /* Pins 10-7 together */
+ if (~s->sense & 6) {
+ sense = (~s->sense & 7) | 6;
+ }
+ }
+ if (!(macfb_sense->ext_sense & 4)) {
+ /* Pins 4-10 together */
+ if (~s->sense & 5) {
+ sense = (~s->sense & 7) | 5;
+ }
+ }
+ } else {
+ /* Normal sense */
+ sense = (~macfb_sense->sense & 7) | (~s->sense & 7);
+ }
+
+ trace_macfb_sense_read(sense);
+ return sense;
+}
+
+static void macfb_sense_write(MacfbState *s, uint32_t val)
+{
+ s->sense = val;
+
+ trace_macfb_sense_write(val);
+ return;
+}
+
static void macfb_update_display(void *opaque)
{
MacfbState *s = opaque;
@@ -290,8 +392,15 @@ static uint64_t macfb_ctrl_read(void *opaque,
hwaddr addr,
unsigned int size)
{
+ MacfbState *s = opaque;
uint64_t val = 0;
+ switch (addr) {
+ case DAFB_MODE_SENSE:
+ val = macfb_sense_read(s);
+ break;
+ }
+
trace_macfb_ctrl_read(addr, val, size);
return val;
}
@@ -303,6 +412,9 @@ static void macfb_ctrl_write(void *opaque,
{
MacfbState *s = opaque;
switch (addr) {
+ case DAFB_MODE_SENSE:
+ macfb_sense_write(s, val);
+ break;
case DAFB_RESET:
s->palette_current = 0;
break;
@@ -342,6 +454,7 @@ static const VMStateDescription vmstate_macfb = {
.fields = (VMStateField[]) {
VMSTATE_UINT8_ARRAY(color_palette, MacfbState, 256 * 3),
VMSTATE_UINT32(palette_current, MacfbState),
+ VMSTATE_UINT32(sense, MacfbState),
VMSTATE_END_OF_LIST()
}
};
diff --git a/hw/display/trace-events b/hw/display/trace-events
index 6c460aaa4c..5753421a43 100644
--- a/hw/display/trace-events
+++ b/hw/display/trace-events
@@ -171,3 +171,5 @@ sm501_2d_engine_write(uint32_t addr, uint32_t val) "addr=0x%x, val=0x%x"
# macfb.c
macfb_ctrl_read(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx64 " value 0x%"PRIx64 " size %d"
macfb_ctrl_write(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx64 " value 0x%"PRIx64 " size %d"
+macfb_sense_read(uint32_t value) "video sense: 0x%"PRIx32
+macfb_sense_write(uint32_t value) "video sense: 0x%"PRIx32
diff --git a/include/hw/display/macfb.h b/include/hw/display/macfb.h
index 80806b0306..febf4ce0e8 100644
--- a/include/hw/display/macfb.h
+++ b/include/hw/display/macfb.h
@@ -17,6 +17,24 @@
#include "ui/console.h"
#include "qom/object.h"
+typedef enum {
+ MACFB_DISPLAY_APPLE_21_COLOR = 0,
+ MACFB_DISPLAY_APPLE_PORTRAIT = 1,
+ MACFB_DISPLAY_APPLE_12_RGB = 2,
+ MACFB_DISPLAY_APPLE_2PAGE_MONO = 3,
+ MACFB_DISPLAY_NTSC_UNDERSCAN = 4,
+ MACFB_DISPLAY_NTSC_OVERSCAN = 5,
+ MACFB_DISPLAY_APPLE_12_MONO = 6,
+ MACFB_DISPLAY_APPLE_13_RGB = 7,
+ MACFB_DISPLAY_16_COLOR = 8,
+ MACFB_DISPLAY_PAL1_UNDERSCAN = 9,
+ MACFB_DISPLAY_PAL1_OVERSCAN = 10,
+ MACFB_DISPLAY_PAL2_UNDERSCAN = 11,
+ MACFB_DISPLAY_PAL2_OVERSCAN = 12,
+ MACFB_DISPLAY_VGA = 13,
+ MACFB_DISPLAY_SVGA = 14,
+} MacfbDisplayType;
+
typedef struct MacfbState {
MemoryRegion mem_vram;
MemoryRegion mem_ctrl;
@@ -28,6 +46,8 @@ typedef struct MacfbState {
uint8_t color_palette[256 * 3];
uint32_t width, height; /* in pixels */
uint8_t depth;
+
+ uint32_t sense;
} MacfbState;
#define TYPE_MACFB "sysbus-macfb"
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 07/12] macfb: add qdev property to specify display type
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (5 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 06/12] macfb: implement mode sense to allow display type to be detected Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM Mark Cave-Ayland
` (4 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
Since the available resolutions and colour depths are determined by the attached
display type, add a qdev property to allow the display type to be specified.
The main resolutions of interest are high resolution 1152x870 with 8-bit colour
and SVGA resolution up to 800x600 with 32-bit colour so update the q800 machine
to allow high resolution mode if specified and otherwise fall back to SVGA.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 7 ++++++-
hw/m68k/q800.c | 5 +++++
include/hw/display/macfb.h | 1 +
3 files changed, 12 insertions(+), 1 deletion(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 6e485d7aef..f98bcdec2d 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -316,7 +316,8 @@ static uint32_t macfb_sense_read(MacfbState *s)
MacFbSense *macfb_sense;
uint8_t sense;
- macfb_sense = &macfb_sense_table[MACFB_DISPLAY_VGA];
+ assert(s->type < ARRAY_SIZE(macfb_sense_table));
+ macfb_sense = &macfb_sense_table[s->type];
if (macfb_sense->sense == 0x7) {
/* Extended sense */
sense = 0;
@@ -544,6 +545,8 @@ static Property macfb_sysbus_properties[] = {
DEFINE_PROP_UINT32("width", MacfbSysBusState, macfb.width, 640),
DEFINE_PROP_UINT32("height", MacfbSysBusState, macfb.height, 480),
DEFINE_PROP_UINT8("depth", MacfbSysBusState, macfb.depth, 8),
+ DEFINE_PROP_UINT8("display", MacfbSysBusState, macfb.type,
+ MACFB_DISPLAY_VGA),
DEFINE_PROP_END_OF_LIST(),
};
@@ -551,6 +554,8 @@ static Property macfb_nubus_properties[] = {
DEFINE_PROP_UINT32("width", MacfbNubusState, macfb.width, 640),
DEFINE_PROP_UINT32("height", MacfbNubusState, macfb.height, 480),
DEFINE_PROP_UINT8("depth", MacfbNubusState, macfb.depth, 8),
+ DEFINE_PROP_UINT8("display", MacfbNubusState, macfb.type,
+ MACFB_DISPLAY_VGA),
DEFINE_PROP_END_OF_LIST(),
};
diff --git a/hw/m68k/q800.c b/hw/m68k/q800.c
index 09b3366024..5223b880bc 100644
--- a/hw/m68k/q800.c
+++ b/hw/m68k/q800.c
@@ -421,6 +421,11 @@ static void q800_init(MachineState *machine)
qdev_prop_set_uint32(dev, "width", graphic_width);
qdev_prop_set_uint32(dev, "height", graphic_height);
qdev_prop_set_uint8(dev, "depth", graphic_depth);
+ if (graphic_width == 1152 && graphic_height == 870 && graphic_depth == 8) {
+ qdev_prop_set_uint8(dev, "display", MACFB_DISPLAY_APPLE_21_COLOR);
+ } else {
+ qdev_prop_set_uint8(dev, "display", MACFB_DISPLAY_VGA);
+ }
qdev_realize_and_unref(dev, BUS(nubus), &error_fatal);
cs = CPU(cpu);
diff --git a/include/hw/display/macfb.h b/include/hw/display/macfb.h
index febf4ce0e8..e95a97ebdc 100644
--- a/include/hw/display/macfb.h
+++ b/include/hw/display/macfb.h
@@ -46,6 +46,7 @@ typedef struct MacfbState {
uint8_t color_palette[256 * 3];
uint32_t width, height; /* in pixels */
uint8_t depth;
+ uint8_t type;
uint32_t sense;
} MacfbState;
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (6 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 07/12] macfb: add qdev property to specify display type Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-05 9:50 ` Laurent Vivier
2021-10-04 21:19 ` [PATCH v2 09/12] macfb: fix up 1-bit pixel encoding Mark Cave-Ayland
` (3 subsequent siblings)
11 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
The monitor modes table is found by experimenting with the Monitors Control
Panel in MacOS and analysing the reads/writes. From this it can be found that
the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
registers.
Implement the first block of DAFB registers as a register array including the
existing sense register, the newly discovered control registers above, and also
the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
determine the current video mode.
These experiments also show that the offset of the start of video RAM and the
stride can change depending upon the monitor mode, so update macfb_draw_graphic()
and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
accordingly.
Finally update macfb_common_realize() so that only the resolution and depth
supported by the display type can be specified on the command line.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
hw/display/trace-events | 1 +
hw/m68k/q800.c | 11 ++--
include/hw/display/macfb.h | 16 ++++-
4 files changed, 131 insertions(+), 21 deletions(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index f98bcdec2d..357fe18be5 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -22,12 +22,16 @@
#include "migration/vmstate.h"
#include "trace.h"
-#define VIDEO_BASE 0x00001000
+#define VIDEO_BASE 0x0
#define DAFB_BASE 0x00800000
#define MACFB_PAGE_SIZE 4096
#define MACFB_VRAM_SIZE (4 * MiB)
+#define DAFB_MODE_VADDR1 0x0
+#define DAFB_MODE_VADDR2 0x4
+#define DAFB_MODE_CTRL1 0x8
+#define DAFB_MODE_CTRL2 0xc
#define DAFB_MODE_SENSE 0x1c
#define DAFB_RESET 0x200
#define DAFB_LUT 0x213
@@ -89,6 +93,22 @@ static MacFbSense macfb_sense_table[] = {
{ MACFB_DISPLAY_SVGA, 0x7, 0x5 },
};
+static MacFbMode macfb_mode_table[] = {
+ { MACFB_DISPLAY_VGA, 1, 0x100, 0x71e, 640, 480, 0x400, 0x1000 },
+ { MACFB_DISPLAY_VGA, 2, 0x100, 0x70e, 640, 480, 0x400, 0x1000 },
+ { MACFB_DISPLAY_VGA, 4, 0x100, 0x706, 640, 480, 0x400, 0x1000 },
+ { MACFB_DISPLAY_VGA, 8, 0x100, 0x702, 640, 480, 0x400, 0x1000 },
+ { MACFB_DISPLAY_VGA, 24, 0x100, 0x7ff, 640, 480, 0x1000, 0x1000 },
+ { MACFB_DISPLAY_VGA, 1, 0xd0 , 0x70e, 800, 600, 0x340, 0xe00 },
+ { MACFB_DISPLAY_VGA, 2, 0xd0 , 0x706, 800, 600, 0x340, 0xe00 },
+ { MACFB_DISPLAY_VGA, 4, 0xd0 , 0x702, 800, 600, 0x340, 0xe00 },
+ { MACFB_DISPLAY_VGA, 8, 0xd0, 0x700, 800, 600, 0x340, 0xe00 },
+ { MACFB_DISPLAY_VGA, 24, 0x340, 0x100, 800, 600, 0xd00, 0xe00 },
+ { MACFB_DISPLAY_APPLE_21_COLOR, 1, 0x90, 0x506, 1152, 870, 0x240, 0x80 },
+ { MACFB_DISPLAY_APPLE_21_COLOR, 2, 0x90, 0x502, 1152, 870, 0x240, 0x80 },
+ { MACFB_DISPLAY_APPLE_21_COLOR, 4, 0x90, 0x500, 1152, 870, 0x240, 0x80 },
+ { MACFB_DISPLAY_APPLE_21_COLOR, 8, 0x120, 0x5ff, 1152, 870, 0x480, 0x80 },
+};
typedef void macfb_draw_line_func(MacfbState *s, uint8_t *d, uint32_t addr,
int width);
@@ -246,7 +266,7 @@ static void macfb_draw_graphic(MacfbState *s)
ram_addr_t page;
uint32_t v = 0;
int y, ymin;
- int macfb_stride = (s->depth * s->width + 7) / 8;
+ int macfb_stride = s->mode->stride;
macfb_draw_line_func *macfb_draw_line;
switch (s->depth) {
@@ -278,7 +298,7 @@ static void macfb_draw_graphic(MacfbState *s)
DIRTY_MEMORY_VGA);
ymin = -1;
- page = 0;
+ page = s->mode->offset;
for (y = 0; y < s->height; y++, page += macfb_stride) {
if (macfb_check_dirty(s, snap, page, macfb_stride)) {
uint8_t *data_display;
@@ -323,25 +343,26 @@ static uint32_t macfb_sense_read(MacfbState *s)
sense = 0;
if (!(macfb_sense->ext_sense & 1)) {
/* Pins 7-4 together */
- if (~s->sense & 3) {
- sense = (~s->sense & 7) | 3;
+ if (~s->regs[DAFB_MODE_SENSE >> 2] & 3) {
+ sense = (~s->regs[DAFB_MODE_SENSE >> 2] & 7) | 3;
}
}
if (!(macfb_sense->ext_sense & 2)) {
/* Pins 10-7 together */
- if (~s->sense & 6) {
- sense = (~s->sense & 7) | 6;
+ if (~s->regs[DAFB_MODE_SENSE >> 2] & 6) {
+ sense = (~s->regs[DAFB_MODE_SENSE >> 2] & 7) | 6;
}
}
if (!(macfb_sense->ext_sense & 4)) {
/* Pins 4-10 together */
- if (~s->sense & 5) {
- sense = (~s->sense & 7) | 5;
+ if (~s->regs[DAFB_MODE_SENSE >> 2] & 5) {
+ sense = (~s->regs[DAFB_MODE_SENSE >> 2] & 7) | 5;
}
}
} else {
/* Normal sense */
- sense = (~macfb_sense->sense & 7) | (~s->sense & 7);
+ sense = (~macfb_sense->sense & 7) |
+ (~s->regs[DAFB_MODE_SENSE >> 2] & 7);
}
trace_macfb_sense_read(sense);
@@ -350,12 +371,64 @@ static uint32_t macfb_sense_read(MacfbState *s)
static void macfb_sense_write(MacfbState *s, uint32_t val)
{
- s->sense = val;
+ s->regs[DAFB_MODE_SENSE >> 2] = val;
trace_macfb_sense_write(val);
return;
}
+static void macfb_update_mode(MacfbState *s)
+{
+ s->width = s->mode->width;
+ s->height = s->mode->height;
+ s->depth = s->mode->depth;
+
+ trace_macfb_update_mode(s->width, s->height, s->depth);
+ macfb_invalidate_display(s);
+}
+
+static void macfb_mode_write(MacfbState *s)
+{
+ MacFbMode *macfb_mode;
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
+ macfb_mode = &macfb_mode_table[i];
+
+ if (s->type != macfb_mode->type) {
+ continue;
+ }
+
+ if ((s->regs[DAFB_MODE_CTRL1 >> 2] & 0xff) ==
+ (macfb_mode->mode_ctrl1 & 0xff) &&
+ (s->regs[DAFB_MODE_CTRL2 >> 2] & 0xff) ==
+ (macfb_mode->mode_ctrl2 & 0xff)) {
+ s->mode = macfb_mode;
+ macfb_update_mode(s);
+ break;
+ }
+ }
+}
+
+static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
+ uint16_t width, uint16_t height,
+ uint8_t depth)
+{
+ MacFbMode *macfb_mode;
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
+ macfb_mode = &macfb_mode_table[i];
+
+ if (display_type == macfb_mode->type && width == macfb_mode->width &&
+ height == macfb_mode->height && depth == macfb_mode->depth) {
+ return macfb_mode;
+ }
+ }
+
+ return NULL;
+}
+
static void macfb_update_display(void *opaque)
{
MacfbState *s = opaque;
@@ -397,6 +470,12 @@ static uint64_t macfb_ctrl_read(void *opaque,
uint64_t val = 0;
switch (addr) {
+ case DAFB_MODE_VADDR1:
+ case DAFB_MODE_VADDR2:
+ case DAFB_MODE_CTRL1:
+ case DAFB_MODE_CTRL2:
+ val = s->regs[addr >> 2];
+ break;
case DAFB_MODE_SENSE:
val = macfb_sense_read(s);
break;
@@ -413,6 +492,17 @@ static void macfb_ctrl_write(void *opaque,
{
MacfbState *s = opaque;
switch (addr) {
+ case DAFB_MODE_VADDR1:
+ case DAFB_MODE_VADDR2:
+ s->regs[addr >> 2] = val;
+ break;
+ case DAFB_MODE_CTRL1 ... DAFB_MODE_CTRL1 + 3:
+ case DAFB_MODE_CTRL2 ... DAFB_MODE_CTRL2 + 3:
+ s->regs[addr >> 2] = val;
+ if (val) {
+ macfb_mode_write(s);
+ }
+ break;
case DAFB_MODE_SENSE:
macfb_sense_write(s, val);
break;
@@ -442,7 +532,7 @@ static const MemoryRegionOps macfb_ctrl_ops = {
static int macfb_post_load(void *opaque, int version_id)
{
- macfb_invalidate_display(opaque);
+ macfb_mode_write(opaque);
return 0;
}
@@ -455,7 +545,7 @@ static const VMStateDescription vmstate_macfb = {
.fields = (VMStateField[]) {
VMSTATE_UINT8_ARRAY(color_palette, MacfbState, 256 * 3),
VMSTATE_UINT32(palette_current, MacfbState),
- VMSTATE_UINT32(sense, MacfbState),
+ VMSTATE_UINT32_ARRAY(regs, MacfbState, MACFB_NUM_REGS),
VMSTATE_END_OF_LIST()
}
};
@@ -469,9 +559,10 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
{
DisplaySurface *surface;
- if (s->depth != 1 && s->depth != 2 && s->depth != 4 && s->depth != 8 &&
- s->depth != 16 && s->depth != 24) {
- error_setg(errp, "unknown guest depth %d", s->depth);
+ s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
+ if (!s->mode) {
+ error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
+ s->width, s->height, s->depth);
return false;
}
@@ -493,6 +584,7 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
s->vram_bit_mask = MACFB_VRAM_SIZE - 1;
memory_region_set_coalescing(&s->mem_vram);
+ macfb_update_mode(s);
return true;
}
diff --git a/hw/display/trace-events b/hw/display/trace-events
index 5753421a43..075cd4614c 100644
--- a/hw/display/trace-events
+++ b/hw/display/trace-events
@@ -173,3 +173,4 @@ macfb_ctrl_read(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx
macfb_ctrl_write(uint64_t addr, uint64_t value, unsigned int size) "addr 0x%"PRIx64 " value 0x%"PRIx64 " size %d"
macfb_sense_read(uint32_t value) "video sense: 0x%"PRIx32
macfb_sense_write(uint32_t value) "video sense: 0x%"PRIx32
+macfb_update_mode(uint32_t width, uint32_t height, uint8_t depth) "setting mode to width %"PRId32 " height %"PRId32 " size %d"
diff --git a/hw/m68k/q800.c b/hw/m68k/q800.c
index 5223b880bc..df3fd3711e 100644
--- a/hw/m68k/q800.c
+++ b/hw/m68k/q800.c
@@ -74,7 +74,7 @@
* is needed by the kernel to have early display and
* thus provided by the bootloader
*/
-#define VIDEO_BASE 0xf9001000
+#define VIDEO_BASE 0xf9000000
#define MAC_CLOCK 3686418
@@ -221,6 +221,7 @@ static void q800_init(MachineState *machine)
uint8_t *prom;
const int io_slice_nb = (IO_SIZE / IO_SLICE) - 1;
int i, checksum;
+ MacFbMode *macfb_mode;
ram_addr_t ram_size = machine->ram_size;
const char *kernel_filename = machine->kernel_filename;
const char *initrd_filename = machine->initrd_filename;
@@ -428,6 +429,8 @@ static void q800_init(MachineState *machine)
}
qdev_realize_and_unref(dev, BUS(nubus), &error_fatal);
+ macfb_mode = (NUBUS_MACFB(dev)->macfb).mode;
+
cs = CPU(cpu);
if (linux_boot) {
uint64_t high;
@@ -450,12 +453,12 @@ static void q800_init(MachineState *machine)
BOOTINFO1(cs->as, parameters_base,
BI_MAC_MEMSIZE, ram_size >> 20); /* in MB */
BOOTINFO2(cs->as, parameters_base, BI_MEMCHUNK, 0, ram_size);
- BOOTINFO1(cs->as, parameters_base, BI_MAC_VADDR, VIDEO_BASE);
+ BOOTINFO1(cs->as, parameters_base, BI_MAC_VADDR,
+ VIDEO_BASE + macfb_mode->offset);
BOOTINFO1(cs->as, parameters_base, BI_MAC_VDEPTH, graphic_depth);
BOOTINFO1(cs->as, parameters_base, BI_MAC_VDIM,
(graphic_height << 16) | graphic_width);
- BOOTINFO1(cs->as, parameters_base, BI_MAC_VROW,
- (graphic_width * graphic_depth + 7) / 8);
+ BOOTINFO1(cs->as, parameters_base, BI_MAC_VROW, macfb_mode->stride);
BOOTINFO1(cs->as, parameters_base, BI_MAC_SCCBASE, SCC_BASE);
rom = g_malloc(sizeof(*rom));
diff --git a/include/hw/display/macfb.h b/include/hw/display/macfb.h
index e95a97ebdc..0aff0d84d2 100644
--- a/include/hw/display/macfb.h
+++ b/include/hw/display/macfb.h
@@ -35,6 +35,19 @@ typedef enum {
MACFB_DISPLAY_SVGA = 14,
} MacfbDisplayType;
+typedef struct MacFbMode {
+ uint8_t type;
+ uint8_t depth;
+ uint32_t mode_ctrl1;
+ uint32_t mode_ctrl2;
+ uint32_t width;
+ uint32_t height;
+ uint32_t stride;
+ uint32_t offset;
+} MacFbMode;
+
+#define MACFB_NUM_REGS 8
+
typedef struct MacfbState {
MemoryRegion mem_vram;
MemoryRegion mem_ctrl;
@@ -48,7 +61,8 @@ typedef struct MacfbState {
uint8_t depth;
uint8_t type;
- uint32_t sense;
+ uint32_t regs[MACFB_NUM_REGS];
+ MacFbMode *mode;
} MacfbState;
#define TYPE_MACFB "sysbus-macfb"
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-04 21:19 ` [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM Mark Cave-Ayland
@ 2021-10-05 9:50 ` Laurent Vivier
2021-10-05 11:38 ` Mark Cave-Ayland
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-05 9:50 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
> The monitor modes table is found by experimenting with the Monitors Control
> Panel in MacOS and analysing the reads/writes. From this it can be found that
> the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
> registers.
>
> Implement the first block of DAFB registers as a register array including the
> existing sense register, the newly discovered control registers above, and also
> the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
> determine the current video mode.
>
> These experiments also show that the offset of the start of video RAM and the
> stride can change depending upon the monitor mode, so update macfb_draw_graphic()
> and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
> accordingly.
>
> Finally update macfb_common_realize() so that only the resolution and depth
> supported by the display type can be specified on the command line.
>
> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
> ---
> hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
> hw/display/trace-events | 1 +
> hw/m68k/q800.c | 11 ++--
> include/hw/display/macfb.h | 16 ++++-
> 4 files changed, 131 insertions(+), 21 deletions(-)
>
> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
> index f98bcdec2d..357fe18be5 100644
> --- a/hw/display/macfb.c
> +++ b/hw/display/macfb.c
>
...
> +static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
> + uint16_t width, uint16_t height,
> + uint8_t depth)
> +{
> + MacFbMode *macfb_mode;
> + int i;
> +
> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
> + macfb_mode = &macfb_mode_table[i];
> +
> + if (display_type == macfb_mode->type && width == macfb_mode->width &&
> + height == macfb_mode->height && depth == macfb_mode->depth) {
> + return macfb_mode;
> + }
> + }
> +
> + return NULL;
> +}
> +
I misunderstood this part when I reviewed v1...
It means you have to provide the monitor type to QEMU to switch from the default mode?
But, as a user, how do we know which modes are allowed with which resolution?
Is possible to try to set internally the type here according to the resolution?
Could you provide an command line example how to start the q800 with the 1152x870 resolution?
Thanks,
Laurent
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-05 9:50 ` Laurent Vivier
@ 2021-10-05 11:38 ` Mark Cave-Ayland
2021-10-05 15:08 ` Laurent Vivier
0 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-05 11:38 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 05/10/2021 10:50, Laurent Vivier wrote:
> Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
>> The monitor modes table is found by experimenting with the Monitors Control
>> Panel in MacOS and analysing the reads/writes. From this it can be found that
>> the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
>> registers.
>>
>> Implement the first block of DAFB registers as a register array including the
>> existing sense register, the newly discovered control registers above, and also
>> the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
>> determine the current video mode.
>>
>> These experiments also show that the offset of the start of video RAM and the
>> stride can change depending upon the monitor mode, so update macfb_draw_graphic()
>> and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
>> accordingly.
>>
>> Finally update macfb_common_realize() so that only the resolution and depth
>> supported by the display type can be specified on the command line.
>>
>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
>> ---
>> hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
>> hw/display/trace-events | 1 +
>> hw/m68k/q800.c | 11 ++--
>> include/hw/display/macfb.h | 16 ++++-
>> 4 files changed, 131 insertions(+), 21 deletions(-)
>>
>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>> index f98bcdec2d..357fe18be5 100644
>> --- a/hw/display/macfb.c
>> +++ b/hw/display/macfb.c
>>
> ...
>> +static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>> + uint16_t width, uint16_t height,
>> + uint8_t depth)
>> +{
>> + MacFbMode *macfb_mode;
>> + int i;
>> +
>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>> + macfb_mode = &macfb_mode_table[i];
>> +
>> + if (display_type == macfb_mode->type && width == macfb_mode->width &&
>> + height == macfb_mode->height && depth == macfb_mode->depth) {
>> + return macfb_mode;
>> + }
>> + }
>> +
>> + return NULL;
>> +}
>> +
>
> I misunderstood this part when I reviewed v1...
>
> It means you have to provide the monitor type to QEMU to switch from the default mode?
Not as such: both the MacOS toolbox ROM and MacOS itself offer a fixed set of
resolutions and depths based upon the display type. What I've done for now is default
the display type to VGA since it offers both 640x480 and 800x600 in 1, 2, 4, 8, 16
and 24-bit colour which should cover the most common use of cases of people wanting
to boot using the MacOS toolbox ROM.
Even if you specify a default on the command line, MacOS still only cares about the
display type and will allow you to change the resolution and depth dynamically,
remembering the last resolution and depth across reboots.
During testing I found that having access to the 1152x870 resolution offered by the
Apple 21" monitor display type was useful to allow larger screen sizes, although only
up to 8-bit depth so I added a bit of code that will switch from a VGA display type
to a 21" display type if the graphics resolution is set to 1152x870x8.
Finally if you boot a Linux kernel directly using -kernel then the provided XxYxD is
placed directly into the relevant bootinfo fields with a VGA display type, unless a
resolution of 1152x870x8 is specified in which case the 21" display type is used as
above.
> But, as a user, how do we know which modes are allowed with which resolution?
>
> Is possible to try to set internally the type here according to the resolution?
>
> Could you provide an command line example how to start the q800 with the 1152x870 resolution?
Sure - simply add "-g 1152x870x8" to your command line. If the -g parameter is
omitted then the display type will default to VGA.
ATB,
Mark.
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-05 11:38 ` Mark Cave-Ayland
@ 2021-10-05 15:08 ` Laurent Vivier
2021-10-05 15:33 ` Mark Cave-Ayland
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-05 15:08 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 05/10/2021 à 13:38, Mark Cave-Ayland a écrit :
> On 05/10/2021 10:50, Laurent Vivier wrote:
>
>> Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
>>> The monitor modes table is found by experimenting with the Monitors Control
>>> Panel in MacOS and analysing the reads/writes. From this it can be found that
>>> the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
>>> registers.
>>>
>>> Implement the first block of DAFB registers as a register array including the
>>> existing sense register, the newly discovered control registers above, and also
>>> the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
>>> determine the current video mode.
>>>
>>> These experiments also show that the offset of the start of video RAM and the
>>> stride can change depending upon the monitor mode, so update macfb_draw_graphic()
>>> and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
>>> accordingly.
>>>
>>> Finally update macfb_common_realize() so that only the resolution and depth
>>> supported by the display type can be specified on the command line.
>>>
>>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>>> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
>>> ---
>>> hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
>>> hw/display/trace-events | 1 +
>>> hw/m68k/q800.c | 11 ++--
>>> include/hw/display/macfb.h | 16 ++++-
>>> 4 files changed, 131 insertions(+), 21 deletions(-)
>>>
>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>> index f98bcdec2d..357fe18be5 100644
>>> --- a/hw/display/macfb.c
>>> +++ b/hw/display/macfb.c
>>>
>> ...
>>> +static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>> + uint16_t width, uint16_t height,
>>> + uint8_t depth)
>>> +{
>>> + MacFbMode *macfb_mode;
>>> + int i;
>>> +
>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>> + macfb_mode = &macfb_mode_table[i];
>>> +
>>> + if (display_type == macfb_mode->type && width == macfb_mode->width &&
>>> + height == macfb_mode->height && depth == macfb_mode->depth) {
>>> + return macfb_mode;
>>> + }
>>> + }
>>> +
>>> + return NULL;
>>> +}
>>> +
>>
>> I misunderstood this part when I reviewed v1...
>>
>> It means you have to provide the monitor type to QEMU to switch from the default mode?
>
> Not as such: both the MacOS toolbox ROM and MacOS itself offer a fixed set of resolutions and depths
> based upon the display type. What I've done for now is default the display type to VGA since it
> offers both 640x480 and 800x600 in 1, 2, 4, 8, 16 and 24-bit colour which should cover the most
> common use of cases of people wanting to boot using the MacOS toolbox ROM.
>
> Even if you specify a default on the command line, MacOS still only cares about the display type and
> will allow you to change the resolution and depth dynamically, remembering the last resolution and
> depth across reboots.
>
> During testing I found that having access to the 1152x870 resolution offered by the Apple 21"
> monitor display type was useful to allow larger screen sizes, although only up to 8-bit depth so I
> added a bit of code that will switch from a VGA display type to a 21" display type if the graphics
> resolution is set to 1152x870x8.
>
> Finally if you boot a Linux kernel directly using -kernel then the provided XxYxD is placed directly
> into the relevant bootinfo fields with a VGA display type, unless a resolution of 1152x870x8 is
> specified in which case the 21" display type is used as above.
>
>> But, as a user, how do we know which modes are allowed with which resolution?
>>
>> Is possible to try to set internally the type here according to the resolution?
>>
>> Could you provide an command line example how to start the q800 with the 1152x870 resolution?
>
> Sure - simply add "-g 1152x870x8" to your command line. If the -g parameter is omitted then the
> display type will default to VGA.
>
Thank you for the explanation.
Perhaps you can add in the error message the list of the available mode and depth?
(it's not a blocker for the series, it can be added later)
As an user, it's hard to know what are the allowed values.
Thanks,
Laurent
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-05 15:08 ` Laurent Vivier
@ 2021-10-05 15:33 ` Mark Cave-Ayland
2021-10-06 12:24 ` Laurent Vivier
0 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-05 15:33 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 05/10/2021 16:08, Laurent Vivier wrote:
> Le 05/10/2021 à 13:38, Mark Cave-Ayland a écrit :
>> On 05/10/2021 10:50, Laurent Vivier wrote:
>>
>>> Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
>>>> The monitor modes table is found by experimenting with the Monitors Control
>>>> Panel in MacOS and analysing the reads/writes. From this it can be found that
>>>> the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
>>>> registers.
>>>>
>>>> Implement the first block of DAFB registers as a register array including the
>>>> existing sense register, the newly discovered control registers above, and also
>>>> the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
>>>> determine the current video mode.
>>>>
>>>> These experiments also show that the offset of the start of video RAM and the
>>>> stride can change depending upon the monitor mode, so update macfb_draw_graphic()
>>>> and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
>>>> accordingly.
>>>>
>>>> Finally update macfb_common_realize() so that only the resolution and depth
>>>> supported by the display type can be specified on the command line.
>>>>
>>>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>>>> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
>>>> ---
>>>> hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
>>>> hw/display/trace-events | 1 +
>>>> hw/m68k/q800.c | 11 ++--
>>>> include/hw/display/macfb.h | 16 ++++-
>>>> 4 files changed, 131 insertions(+), 21 deletions(-)
>>>>
>>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>>> index f98bcdec2d..357fe18be5 100644
>>>> --- a/hw/display/macfb.c
>>>> +++ b/hw/display/macfb.c
>>>>
>>> ...
>>>> +static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>>> + uint16_t width, uint16_t height,
>>>> + uint8_t depth)
>>>> +{
>>>> + MacFbMode *macfb_mode;
>>>> + int i;
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>>> + macfb_mode = &macfb_mode_table[i];
>>>> +
>>>> + if (display_type == macfb_mode->type && width == macfb_mode->width &&
>>>> + height == macfb_mode->height && depth == macfb_mode->depth) {
>>>> + return macfb_mode;
>>>> + }
>>>> + }
>>>> +
>>>> + return NULL;
>>>> +}
>>>> +
>>>
>>> I misunderstood this part when I reviewed v1...
>>>
>>> It means you have to provide the monitor type to QEMU to switch from the default mode?
>>
>> Not as such: both the MacOS toolbox ROM and MacOS itself offer a fixed set of resolutions and depths
>> based upon the display type. What I've done for now is default the display type to VGA since it
>> offers both 640x480 and 800x600 in 1, 2, 4, 8, 16 and 24-bit colour which should cover the most
>> common use of cases of people wanting to boot using the MacOS toolbox ROM.
>>
>> Even if you specify a default on the command line, MacOS still only cares about the display type and
>> will allow you to change the resolution and depth dynamically, remembering the last resolution and
>> depth across reboots.
>>
>> During testing I found that having access to the 1152x870 resolution offered by the Apple 21"
>> monitor display type was useful to allow larger screen sizes, although only up to 8-bit depth so I
>> added a bit of code that will switch from a VGA display type to a 21" display type if the graphics
>> resolution is set to 1152x870x8.
>>
>> Finally if you boot a Linux kernel directly using -kernel then the provided XxYxD is placed directly
>> into the relevant bootinfo fields with a VGA display type, unless a resolution of 1152x870x8 is
>> specified in which case the 21" display type is used as above.
>>
>>> But, as a user, how do we know which modes are allowed with which resolution?
>>>
>>> Is possible to try to set internally the type here according to the resolution?
>>>
>>> Could you provide an command line example how to start the q800 with the 1152x870 resolution?
>>
>> Sure - simply add "-g 1152x870x8" to your command line. If the -g parameter is omitted then the
>> display type will default to VGA.
>>
>
> Thank you for the explanation.
>
> Perhaps you can add in the error message the list of the available mode and depth?
> (it's not a blocker for the series, it can be added later)
>
> As an user, it's hard to know what are the allowed values.
This is where it becomes a bit trickier, since technically booting Linux with -kernel
you can use any supported values as long as everything fits in the video RAM which is
why there isn't currently a hard-coded list :)
If you would prefer a fixed set of resolutions for both MacOS and Linux booted with
-kernel then I think we should aim for VGA (640x480) in 1, 2, 4, 8, 16 and 24-bit
depths, SVGA (800x600) in 1, 2, 4, 8, 16 and 24-bit depths and 21" (1152x870) in 1,
2, 4 and 8-bit depths. Do other emulated displays show supported resolutions in the
error message, or do they rely on including this in the documentation?
One other point to note is that when MacOS detects a VGA display type, it defaults to
640x480 until you go into the Control Panel and select 800x600 to enable it. It seems
this was a precaution to prevent damaging VGA monitors that couldn't sync with an
SVGA signal, since the cable doesn't allow you to distinguish between a VGA and an
SVGA-capable monitor.
Finally after the initial boot MacOS always boots into the last resolution/depth set
successfully in the Control Panel for the detected display type. So you could specify
640x480x8 on the command line, but if you changed this to 800x600x4 before the last
reboot then MacOS would reprogram the macfb registers with the same values again,
effectively overriding the -g parameter and making it a no-op...
ATB,
Mark.
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-05 15:33 ` Mark Cave-Ayland
@ 2021-10-06 12:24 ` Laurent Vivier
2021-10-06 13:54 ` Mark Cave-Ayland
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-06 12:24 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 05/10/2021 à 17:33, Mark Cave-Ayland a écrit :
> On 05/10/2021 16:08, Laurent Vivier wrote:
>
>> Le 05/10/2021 à 13:38, Mark Cave-Ayland a écrit :
>>> On 05/10/2021 10:50, Laurent Vivier wrote:
>>>
>>>> Le 04/10/2021 à 23:19, Mark Cave-Ayland a écrit :
>>>>> The monitor modes table is found by experimenting with the Monitors Control
>>>>> Panel in MacOS and analysing the reads/writes. From this it can be found that
>>>>> the mode is controlled by writes to the DAFB_MODE_CTRL1 and DAFB_MODE_CTRL2
>>>>> registers.
>>>>>
>>>>> Implement the first block of DAFB registers as a register array including the
>>>>> existing sense register, the newly discovered control registers above, and also
>>>>> the DAFB_MODE_VADDR1 and DAFB_MODE_VADDR2 registers which are used by NetBSD to
>>>>> determine the current video mode.
>>>>>
>>>>> These experiments also show that the offset of the start of video RAM and the
>>>>> stride can change depending upon the monitor mode, so update macfb_draw_graphic()
>>>>> and both the BI_MAC_VADDR and BI_MAC_VROW bootinfo for the q800 machine
>>>>> accordingly.
>>>>>
>>>>> Finally update macfb_common_realize() so that only the resolution and depth
>>>>> supported by the display type can be specified on the command line.
>>>>>
>>>>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>>>>> Reviewed-by: Laurent Vivier <laurent@vivier.eu>
>>>>> ---
>>>>> hw/display/macfb.c | 124 ++++++++++++++++++++++++++++++++-----
>>>>> hw/display/trace-events | 1 +
>>>>> hw/m68k/q800.c | 11 ++--
>>>>> include/hw/display/macfb.h | 16 ++++-
>>>>> 4 files changed, 131 insertions(+), 21 deletions(-)
>>>>>
>>>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>>>> index f98bcdec2d..357fe18be5 100644
>>>>> --- a/hw/display/macfb.c
>>>>> +++ b/hw/display/macfb.c
>>>>>
>>>> ...
>>>>> +static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>>>> + uint16_t width, uint16_t height,
>>>>> + uint8_t depth)
>>>>> +{
>>>>> + MacFbMode *macfb_mode;
>>>>> + int i;
>>>>> +
>>>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>>>> + macfb_mode = &macfb_mode_table[i];
>>>>> +
>>>>> + if (display_type == macfb_mode->type && width == macfb_mode->width &&
>>>>> + height == macfb_mode->height && depth == macfb_mode->depth) {
>>>>> + return macfb_mode;
>>>>> + }
>>>>> + }
>>>>> +
>>>>> + return NULL;
>>>>> +}
>>>>> +
>>>>
>>>> I misunderstood this part when I reviewed v1...
>>>>
>>>> It means you have to provide the monitor type to QEMU to switch from the default mode?
>>>
>>> Not as such: both the MacOS toolbox ROM and MacOS itself offer a fixed set of resolutions and depths
>>> based upon the display type. What I've done for now is default the display type to VGA since it
>>> offers both 640x480 and 800x600 in 1, 2, 4, 8, 16 and 24-bit colour which should cover the most
>>> common use of cases of people wanting to boot using the MacOS toolbox ROM.
>>>
>>> Even if you specify a default on the command line, MacOS still only cares about the display type and
>>> will allow you to change the resolution and depth dynamically, remembering the last resolution and
>>> depth across reboots.
>>>
>>> During testing I found that having access to the 1152x870 resolution offered by the Apple 21"
>>> monitor display type was useful to allow larger screen sizes, although only up to 8-bit depth so I
>>> added a bit of code that will switch from a VGA display type to a 21" display type if the graphics
>>> resolution is set to 1152x870x8.
>>>
>>> Finally if you boot a Linux kernel directly using -kernel then the provided XxYxD is placed directly
>>> into the relevant bootinfo fields with a VGA display type, unless a resolution of 1152x870x8 is
>>> specified in which case the 21" display type is used as above.
>>>
>>>> But, as a user, how do we know which modes are allowed with which resolution?
>>>>
>>>> Is possible to try to set internally the type here according to the resolution?
>>>>
>>>> Could you provide an command line example how to start the q800 with the 1152x870 resolution?
>>>
>>> Sure - simply add "-g 1152x870x8" to your command line. If the -g parameter is omitted then the
>>> display type will default to VGA.
>>>
>>
>> Thank you for the explanation.
>>
>> Perhaps you can add in the error message the list of the available mode and depth?
>> (it's not a blocker for the series, it can be added later)
>>
>> As an user, it's hard to know what are the allowed values.
>
> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
> any supported values as long as everything fits in the video RAM which is why there isn't currently
> a hard-coded list :)
>
We need the list of "supported values". I don't want to read the code or try values combination
until it works.
In a perfect world, I would like to be able to use any value I want with "-kernel".
For instance I was using "-g 1200x800x24" and it was working fine.
Now I have:
qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
If it's not possible (because the original hardware cannot provide it, and we don't know the base
address or things like that), we don't need the list of the display id, but the list of available
modes: (width,height,depth).
Rougly, something like:
qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
Available modes:
1152x870x8
1152x870x4
1152x870x2
1152x870x1
800x600x24
800x600x8
800x600x4
800x600x2
800x600x1
640x480x24
640x480x8
640x480x4
640x480x2
640x480x1
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 5b8812e9e7d8..4b352eb89c3f 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
return NULL;
}
+static gchar *macfb_mode_list(void)
+{
+ gchar *list = NULL;
+ gchar *mode;
+ MacFbMode *macfb_mode;
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
+ macfb_mode = &macfb_mode_table[i];
+
+ mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
+ macfb_mode->height, macfb_mode->depth);
+ list = g_strconcat(mode, list, NULL);
+ g_free(mode);
+ }
+
+ return list;
+}
+
+
static void macfb_update_display(void *opaque)
{
MacfbState *s = opaque;
@@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
if (!s->mode) {
+ gchar *list;
error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
s->width, s->height, s->depth);
+ list = macfb_mode_list();
+ error_append_hint(errp, "Available modes:\n%s", list);
+ g_free(list);
+
return false;
}
Thanks,
Laurent
^ permalink raw reply related [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 12:24 ` Laurent Vivier
@ 2021-10-06 13:54 ` Mark Cave-Ayland
2021-10-06 15:46 ` Laurent Vivier
0 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-06 13:54 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 06/10/2021 13:24, Laurent Vivier wrote:
>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>> any supported values as long as everything fits in the video RAM which is why there isn't currently
>> a hard-coded list :)
>>
>
> We need the list of "supported values". I don't want to read the code or try values combination
> until it works.
>
> In a perfect world, I would like to be able to use any value I want with "-kernel".
>
> For instance I was using "-g 1200x800x24" and it was working fine.
>
> Now I have:
>
> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>
> If it's not possible (because the original hardware cannot provide it, and we don't know the base
> address or things like that), we don't need the list of the display id, but the list of available
> modes: (width,height,depth).
>
> Rougly, something like:
>
> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
> Available modes:
> 1152x870x8
> 1152x870x4
> 1152x870x2
> 1152x870x1
> 800x600x24
> 800x600x8
> 800x600x4
> 800x600x2
> 800x600x1
> 640x480x24
> 640x480x8
> 640x480x4
> 640x480x2
> 640x480x1
>
> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
> index 5b8812e9e7d8..4b352eb89c3f 100644
> --- a/hw/display/macfb.c
> +++ b/hw/display/macfb.c
> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
> return NULL;
> }
>
> +static gchar *macfb_mode_list(void)
> +{
> + gchar *list = NULL;
> + gchar *mode;
> + MacFbMode *macfb_mode;
> + int i;
> +
> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
> + macfb_mode = &macfb_mode_table[i];
> +
> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
> + macfb_mode->height, macfb_mode->depth);
> + list = g_strconcat(mode, list, NULL);
> + g_free(mode);
> + }
> +
> + return list;
> +}
> +
> +
> static void macfb_update_display(void *opaque)
> {
> MacfbState *s = opaque;
> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
>
> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
> if (!s->mode) {
> + gchar *list;
> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
> s->width, s->height, s->depth);
> + list = macfb_mode_list();
> + error_append_hint(errp, "Available modes:\n%s", list);
> + g_free(list);
> +
> return false;
> }
Hi Laurent,
Thanks for the example - I can certainly squash this into patch 8.
As for allowing extra resolutions via -kernel, since the check is being done in
macfb_common_realize() then it would be possible to add a qdev property that only
gets set when -kernel is used on the command line which bypasses the mode check if
you prefer?
I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution
with 24-bit depth) with -kernel will still work after this patchset given that the
24-bit encoding scheme has changed. Presumably this would also need a corresponding
change in the bootinfo/kernel framebuffer/X configuration somewhere?
ATB,
Mark.
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 13:54 ` Mark Cave-Ayland
@ 2021-10-06 15:46 ` Laurent Vivier
2021-10-06 16:09 ` Mark Cave-Ayland
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-06 15:46 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 06/10/2021 à 15:54, Mark Cave-Ayland a écrit :
> On 06/10/2021 13:24, Laurent Vivier wrote:
>
>>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>>> any supported values as long as everything fits in the video RAM which is why there isn't currently
>>> a hard-coded list :)
>>>
>>
>> We need the list of "supported values". I don't want to read the code or try values combination
>> until it works.
>>
>> In a perfect world, I would like to be able to use any value I want with "-kernel".
>>
>> For instance I was using "-g 1200x800x24" and it was working fine.
>>
>> Now I have:
>>
>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>
>> If it's not possible (because the original hardware cannot provide it, and we don't know the base
>> address or things like that), we don't need the list of the display id, but the list of available
>> modes: (width,height,depth).
>>
>> Rougly, something like:
>>
>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>> Available modes:
>> 1152x870x8
>> 1152x870x4
>> 1152x870x2
>> 1152x870x1
>> 800x600x24
>> 800x600x8
>> 800x600x4
>> 800x600x2
>> 800x600x1
>> 640x480x24
>> 640x480x8
>> 640x480x4
>> 640x480x2
>> 640x480x1
>>
>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>> index 5b8812e9e7d8..4b352eb89c3f 100644
>> --- a/hw/display/macfb.c
>> +++ b/hw/display/macfb.c
>> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>> return NULL;
>> }
>>
>> +static gchar *macfb_mode_list(void)
>> +{
>> + gchar *list = NULL;
>> + gchar *mode;
>> + MacFbMode *macfb_mode;
>> + int i;
>> +
>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>> + macfb_mode = &macfb_mode_table[i];
>> +
>> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
>> + macfb_mode->height, macfb_mode->depth);
>> + list = g_strconcat(mode, list, NULL);
>> + g_free(mode);
>> + }
>> +
>> + return list;
>> +}
>> +
>> +
>> static void macfb_update_display(void *opaque)
>> {
>> MacfbState *s = opaque;
>> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
>>
>> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
>> if (!s->mode) {
>> + gchar *list;
>> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
>> s->width, s->height, s->depth);
>> + list = macfb_mode_list();
>> + error_append_hint(errp, "Available modes:\n%s", list);
>> + g_free(list);
>> +
>> return false;
>> }
>
> Hi Laurent,
>
> Thanks for the example - I can certainly squash this into patch 8.
yes, please.
> As for allowing extra resolutions via -kernel, since the check is being done in
> macfb_common_realize() then it would be possible to add a qdev property that only gets set when
> -kernel is used on the command line which bypasses the mode check if you prefer?
>
I think it can wait and be done by a patch later. For the moment we can focus on MacOS.
> I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution with 24-bit
> depth) with -kernel will still work after this patchset given that the 24-bit encoding scheme has
> changed. Presumably this would also need a corresponding change in the bootinfo/kernel framebuffer/X
> configuration somewhere?
The kernel framebuffer should be easy to fix, if needed, normally we pass the needed information via
the bootinfo structure.
My X configuration is broken for a while. With debian/sid I've never been able to start X (even on a
real q800, I think), and with debian/etch we need a special kernel as the ADB stack has been broken
with old kernel. I was not able to start X for a while now...
And GNOME desktop is not available on debian/sid m68k (some packages are missing). Perhaps I should
try Xfce.
So, don't worry about that...
Thanks,
Laurent
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 15:46 ` Laurent Vivier
@ 2021-10-06 16:09 ` Mark Cave-Ayland
2021-10-06 19:24 ` Laurent Vivier
0 siblings, 1 reply; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-06 16:09 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 06/10/2021 16:46, Laurent Vivier wrote:
> Le 06/10/2021 à 15:54, Mark Cave-Ayland a écrit :
>> On 06/10/2021 13:24, Laurent Vivier wrote:
>>
>>>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>>>> any supported values as long as everything fits in the video RAM which is why there isn't currently
>>>> a hard-coded list :)
>>>>
>>>
>>> We need the list of "supported values". I don't want to read the code or try values combination
>>> until it works.
>>>
>>> In a perfect world, I would like to be able to use any value I want with "-kernel".
>>>
>>> For instance I was using "-g 1200x800x24" and it was working fine.
>>>
>>> Now I have:
>>>
>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>
>>> If it's not possible (because the original hardware cannot provide it, and we don't know the base
>>> address or things like that), we don't need the list of the display id, but the list of available
>>> modes: (width,height,depth).
>>>
>>> Rougly, something like:
>>>
>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>> Available modes:
>>> 1152x870x8
>>> 1152x870x4
>>> 1152x870x2
>>> 1152x870x1
>>> 800x600x24
>>> 800x600x8
>>> 800x600x4
>>> 800x600x2
>>> 800x600x1
>>> 640x480x24
>>> 640x480x8
>>> 640x480x4
>>> 640x480x2
>>> 640x480x1
>>>
>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>> index 5b8812e9e7d8..4b352eb89c3f 100644
>>> --- a/hw/display/macfb.c
>>> +++ b/hw/display/macfb.c
>>> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>> return NULL;
>>> }
>>>
>>> +static gchar *macfb_mode_list(void)
>>> +{
>>> + gchar *list = NULL;
>>> + gchar *mode;
>>> + MacFbMode *macfb_mode;
>>> + int i;
>>> +
>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>> + macfb_mode = &macfb_mode_table[i];
>>> +
>>> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
>>> + macfb_mode->height, macfb_mode->depth);
>>> + list = g_strconcat(mode, list, NULL);
>>> + g_free(mode);
>>> + }
>>> +
>>> + return list;
>>> +}
>>> +
>>> +
>>> static void macfb_update_display(void *opaque)
>>> {
>>> MacfbState *s = opaque;
>>> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
>>>
>>> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
>>> if (!s->mode) {
>>> + gchar *list;
>>> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
>>> s->width, s->height, s->depth);
>>> + list = macfb_mode_list();
>>> + error_append_hint(errp, "Available modes:\n%s", list);
>>> + g_free(list);
>>> +
>>> return false;
>>> }
>>
>> Hi Laurent,
>>
>> Thanks for the example - I can certainly squash this into patch 8.
>
> yes, please.
Okay I'll do that for a v3 (and also split the 1st patch that Phil suggested).
>> As for allowing extra resolutions via -kernel, since the check is being done in
>> macfb_common_realize() then it would be possible to add a qdev property that only gets set when
>> -kernel is used on the command line which bypasses the mode check if you prefer?
>>
>
> I think it can wait and be done by a patch later. For the moment we can focus on MacOS.
>
>> I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution with 24-bit
>> depth) with -kernel will still work after this patchset given that the 24-bit encoding scheme has
>> changed. Presumably this would also need a corresponding change in the bootinfo/kernel framebuffer/X
>> configuration somewhere?
>
> The kernel framebuffer should be easy to fix, if needed, normally we pass the needed information via
> the bootinfo structure.
>
> My X configuration is broken for a while. With debian/sid I've never been able to start X (even on a
> real q800, I think), and with debian/etch we need a special kernel as the ADB stack has been broken
> with old kernel. I was not able to start X for a while now...
FWIW I found that the last set of ADB fixes in mac_via.c actually fixed ADB on old
kernels again (I was able to use keyboard and mouse on the 4.15 kernel you used for
the original patches), so you may be able to get debian/etch working in QEMU. I'd
expect forcing EMILE into a 24-bit depth on a real Quadra 800 would also show the
same issue here.
> And GNOME desktop is not available on debian/sid m68k (some packages are missing). Perhaps I should
> try Xfce.
>
> So, don't worry about that...
Alright, thanks. I'll see if I can get v3 posted later this evening...
ATB,
Mark.
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 16:09 ` Mark Cave-Ayland
@ 2021-10-06 19:24 ` Laurent Vivier
2021-10-06 21:16 ` Laurent Vivier
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-06 19:24 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
Le 06/10/2021 à 18:09, Mark Cave-Ayland a écrit :
> On 06/10/2021 16:46, Laurent Vivier wrote:
>> Le 06/10/2021 à 15:54, Mark Cave-Ayland a écrit :
>>> On 06/10/2021 13:24, Laurent Vivier wrote:
>>>
>>>>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>>>>> any supported values as long as everything fits in the video RAM which is why there isn't
>>>>> currently
>>>>> a hard-coded list :)
>>>>>
>>>>
>>>> We need the list of "supported values". I don't want to read the code or try values combination
>>>> until it works.
>>>>
>>>> In a perfect world, I would like to be able to use any value I want with "-kernel".
>>>>
>>>> For instance I was using "-g 1200x800x24" and it was working fine.
>>>>
>>>> Now I have:
>>>>
>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>>
>>>> If it's not possible (because the original hardware cannot provide it, and we don't know the base
>>>> address or things like that), we don't need the list of the display id, but the list of available
>>>> modes: (width,height,depth).
>>>>
>>>> Rougly, something like:
>>>>
>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>> Available modes:
>>>> 1152x870x8
>>>> 1152x870x4
>>>> 1152x870x2
>>>> 1152x870x1
>>>> 800x600x24
>>>> 800x600x8
>>>> 800x600x4
>>>> 800x600x2
>>>> 800x600x1
>>>> 640x480x24
>>>> 640x480x8
>>>> 640x480x4
>>>> 640x480x2
>>>> 640x480x1
>>>>
>>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>>> index 5b8812e9e7d8..4b352eb89c3f 100644
>>>> --- a/hw/display/macfb.c
>>>> +++ b/hw/display/macfb.c
>>>> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>>> return NULL;
>>>> }
>>>>
>>>> +static gchar *macfb_mode_list(void)
>>>> +{
>>>> + gchar *list = NULL;
>>>> + gchar *mode;
>>>> + MacFbMode *macfb_mode;
>>>> + int i;
>>>> +
>>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>>> + macfb_mode = &macfb_mode_table[i];
>>>> +
>>>> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
>>>> + macfb_mode->height, macfb_mode->depth);
>>>> + list = g_strconcat(mode, list, NULL);
>>>> + g_free(mode);
>>>> + }
>>>> +
>>>> + return list;
>>>> +}
>>>> +
>>>> +
>>>> static void macfb_update_display(void *opaque)
>>>> {
>>>> MacfbState *s = opaque;
>>>> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error
>>>> **errp)
>>>>
>>>> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
>>>> if (!s->mode) {
>>>> + gchar *list;
>>>> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
>>>> s->width, s->height, s->depth);
>>>> + list = macfb_mode_list();
>>>> + error_append_hint(errp, "Available modes:\n%s", list);
>>>> + g_free(list);
>>>> +
>>>> return false;
>>>> }
>>>
>>> Hi Laurent,
>>>
>>> Thanks for the example - I can certainly squash this into patch 8.
>>
>> yes, please.
>
> Okay I'll do that for a v3 (and also split the 1st patch that Phil suggested).
>
>>> As for allowing extra resolutions via -kernel, since the check is being done in
>>> macfb_common_realize() then it would be possible to add a qdev property that only gets set when
>>> -kernel is used on the command line which bypasses the mode check if you prefer?
>>>
>>
>> I think it can wait and be done by a patch later. For the moment we can focus on MacOS.
>>
>>> I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution with 24-bit
>>> depth) with -kernel will still work after this patchset given that the 24-bit encoding scheme has
>>> changed. Presumably this would also need a corresponding change in the bootinfo/kernel framebuffer/X
>>> configuration somewhere?
>>
>> The kernel framebuffer should be easy to fix, if needed, normally we pass the needed information via
>> the bootinfo structure.
>>
>> My X configuration is broken for a while. With debian/sid I've never been able to start X (even on a
>> real q800, I think), and with debian/etch we need a special kernel as the ADB stack has been broken
>> with old kernel. I was not able to start X for a while now...
>
> FWIW I found that the last set of ADB fixes in mac_via.c actually fixed ADB on old kernels again (I
> was able to use keyboard and mouse on the 4.15 kernel you used for the original patches), so you may
> be able to get debian/etch working in QEMU. I'd expect forcing EMILE into a 24-bit depth on a real
> Quadra 800 would also show the same issue here.
>
In fact, I was trying the etch kernel, 2.6.32 that was never able to boot before because VBL
interrupt was not implemented. Now, with your series it boots (congratulation) but there is some
other issues in ADB but I don't think it's related to the changes in QEMU.
I've found another kernel on my disk (4.1.39) and I'm able to start a gnome-session with 1152x870x8
mode. ADB works well. There is an issue with gdm that doesn't take characters from the keyboard but
if I start X manually and then gnome-sessions I'm able to play with it.
Thanks,
Laurent
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 19:24 ` Laurent Vivier
@ 2021-10-06 21:16 ` Laurent Vivier
2021-10-07 9:19 ` Mark Cave-Ayland
0 siblings, 1 reply; 26+ messages in thread
From: Laurent Vivier @ 2021-10-06 21:16 UTC (permalink / raw)
To: Mark Cave-Ayland, qemu-devel
[-- Attachment #1: Type: text/plain, Size: 5928 bytes --]
Le 06/10/2021 à 21:24, Laurent Vivier a écrit :
> Le 06/10/2021 à 18:09, Mark Cave-Ayland a écrit :
>> On 06/10/2021 16:46, Laurent Vivier wrote:
>>> Le 06/10/2021 à 15:54, Mark Cave-Ayland a écrit :
>>>> On 06/10/2021 13:24, Laurent Vivier wrote:
>>>>
>>>>>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>>>>>> any supported values as long as everything fits in the video RAM which is why there isn't
>>>>>> currently
>>>>>> a hard-coded list :)
>>>>>>
>>>>>
>>>>> We need the list of "supported values". I don't want to read the code or try values combination
>>>>> until it works.
>>>>>
>>>>> In a perfect world, I would like to be able to use any value I want with "-kernel".
>>>>>
>>>>> For instance I was using "-g 1200x800x24" and it was working fine.
>>>>>
>>>>> Now I have:
>>>>>
>>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>>>
>>>>> If it's not possible (because the original hardware cannot provide it, and we don't know the base
>>>>> address or things like that), we don't need the list of the display id, but the list of available
>>>>> modes: (width,height,depth).
>>>>>
>>>>> Rougly, something like:
>>>>>
>>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>>> Available modes:
>>>>> 1152x870x8
>>>>> 1152x870x4
>>>>> 1152x870x2
>>>>> 1152x870x1
>>>>> 800x600x24
>>>>> 800x600x8
>>>>> 800x600x4
>>>>> 800x600x2
>>>>> 800x600x1
>>>>> 640x480x24
>>>>> 640x480x8
>>>>> 640x480x4
>>>>> 640x480x2
>>>>> 640x480x1
>>>>>
>>>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>>>> index 5b8812e9e7d8..4b352eb89c3f 100644
>>>>> --- a/hw/display/macfb.c
>>>>> +++ b/hw/display/macfb.c
>>>>> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>>>> return NULL;
>>>>> }
>>>>>
>>>>> +static gchar *macfb_mode_list(void)
>>>>> +{
>>>>> + gchar *list = NULL;
>>>>> + gchar *mode;
>>>>> + MacFbMode *macfb_mode;
>>>>> + int i;
>>>>> +
>>>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>>>> + macfb_mode = &macfb_mode_table[i];
>>>>> +
>>>>> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
>>>>> + macfb_mode->height, macfb_mode->depth);
>>>>> + list = g_strconcat(mode, list, NULL);
>>>>> + g_free(mode);
>>>>> + }
>>>>> +
>>>>> + return list;
>>>>> +}
>>>>> +
>>>>> +
>>>>> static void macfb_update_display(void *opaque)
>>>>> {
>>>>> MacfbState *s = opaque;
>>>>> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error
>>>>> **errp)
>>>>>
>>>>> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
>>>>> if (!s->mode) {
>>>>> + gchar *list;
>>>>> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
>>>>> s->width, s->height, s->depth);
>>>>> + list = macfb_mode_list();
>>>>> + error_append_hint(errp, "Available modes:\n%s", list);
>>>>> + g_free(list);
>>>>> +
>>>>> return false;
>>>>> }
>>>>
>>>> Hi Laurent,
>>>>
>>>> Thanks for the example - I can certainly squash this into patch 8.
>>>
>>> yes, please.
>>
>> Okay I'll do that for a v3 (and also split the 1st patch that Phil suggested).
>>
>>>> As for allowing extra resolutions via -kernel, since the check is being done in
>>>> macfb_common_realize() then it would be possible to add a qdev property that only gets set when
>>>> -kernel is used on the command line which bypasses the mode check if you prefer?
>>>>
>>>
>>> I think it can wait and be done by a patch later. For the moment we can focus on MacOS.
>>>
>>>> I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution with 24-bit
>>>> depth) with -kernel will still work after this patchset given that the 24-bit encoding scheme has
>>>> changed. Presumably this would also need a corresponding change in the bootinfo/kernel framebuffer/X
>>>> configuration somewhere?
>>>
>>> The kernel framebuffer should be easy to fix, if needed, normally we pass the needed information via
>>> the bootinfo structure.
>>>
>>> My X configuration is broken for a while. With debian/sid I've never been able to start X (even on a
>>> real q800, I think), and with debian/etch we need a special kernel as the ADB stack has been broken
>>> with old kernel. I was not able to start X for a while now...
>>
>> FWIW I found that the last set of ADB fixes in mac_via.c actually fixed ADB on old kernels again (I
>> was able to use keyboard and mouse on the 4.15 kernel you used for the original patches), so you may
>> be able to get debian/etch working in QEMU. I'd expect forcing EMILE into a 24-bit depth on a real
>> Quadra 800 would also show the same issue here.
>>
>
> In fact, I was trying the etch kernel, 2.6.32 that was never able to boot before because VBL
> interrupt was not implemented. Now, with your series it boots (congratulation) but there is some
> other issues in ADB but I don't think it's related to the changes in QEMU.
>
> I've found another kernel on my disk (4.1.39) and I'm able to start a gnome-session with 1152x870x8
> mode. ADB works well. There is an issue with gdm that doesn't take characters from the keyboard but
> if I start X manually and then gnome-sessions I'm able to play with it.
And I'm able to run a 800x600x24 session with the attached xorg.conf
To run 1152x870, the DefaultDepth and DefaultFbBpp must be changed to 8.
Thanks,
Laurent
[-- Attachment #2: xorg.conf --]
[-- Type: text/plain, Size: 3012 bytes --]
# /etc/X11/xorg.conf (xorg X Window System server configuration file)
#
# This file was generated by dexconf, the Debian X Configuration tool, using
# values from the debconf database.
#
# Edit this file with caution, and see the /etc/X11/xorg.conf manual page.
# (Type "man /etc/X11/xorg.conf" at the shell prompt.)
#
# This file is automatically updated on xserver-xorg package upgrades *only*
# if it has not been modified since the last upgrade of the xserver-xorg
# package.
#
# If you have edited this file but would like it to be automatically updated
# again, run the following command:
# sudo dpkg-reconfigure -phigh xserver-xorg
Section "Files"
FontPath "/usr/share/fonts/X11/misc"
FontPath "/usr/X11R6/lib/X11/fonts/misc"
FontPath "/usr/share/fonts/X11/cyrillic"
FontPath "/usr/X11R6/lib/X11/fonts/cyrillic"
FontPath "/usr/share/fonts/X11/100dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi/:unscaled"
FontPath "/usr/share/fonts/X11/75dpi/:unscaled"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi/:unscaled"
FontPath "/usr/share/fonts/X11/Type1"
FontPath "/usr/X11R6/lib/X11/fonts/Type1"
FontPath "/usr/share/fonts/X11/100dpi"
FontPath "/usr/X11R6/lib/X11/fonts/100dpi"
FontPath "/usr/share/fonts/X11/75dpi"
FontPath "/usr/X11R6/lib/X11/fonts/75dpi"
# path to defoma fonts
FontPath "/var/lib/defoma/x-ttcidfont-conf.d/dirs/TrueType"
EndSection
Section "Module"
Load "i2c"
Load "bitmap"
Load "ddc"
Load "dri"
Load "extmod"
Load "freetype"
Load "glx"
Load "int10"
Load "vbe"
EndSection
Section "InputDevice"
Identifier "Generic Keyboard"
Driver "keyboard"
Option "CoreKeyboard"
Option "XkbRules" "xorg"
Option "XkbModel" "macintosh"
Option "XkbLayout" "us"
EndSection
Section "InputDevice"
Identifier "Configured Mouse"
Driver "mouse"
Option "CorePointer"
Option "Device" "/dev/input/mice"
Option "Protocol" "ImPS/2"
Option "Emulate3Buttons" "true"
EndSection
Section "Device"
Identifier "Generic Video Card"
Driver "fbdev"
Option "UseFBDev" "true"
EndSection
Section "Monitor"
Identifier "Generic Monitor"
Option "DPMS"
HorizSync 28-60
VertRefresh 43-60
EndSection
Section "Screen"
Identifier "Default Screen"
Device "Generic Video Card"
Monitor "Generic Monitor"
DefaultDepth 24
DefaultFbBpp 32
SubSection "Display"
Depth 1
Modes "1152x870" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 4
Modes "1152x870" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 8
Modes "1152x870" "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 15
Modes "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 16
Modes "800x600" "640x480"
EndSubSection
SubSection "Display"
Depth 24
FbBpp 32
Modes "800x600" "640x480"
EndSubSection
EndSection
Section "ServerLayout"
Identifier "Default Layout"
Screen "Default Screen"
InputDevice "Generic Keyboard"
InputDevice "Configured Mouse"
EndSection
Section "DRI"
Mode 0666
EndSection
^ permalink raw reply [flat|nested] 26+ messages in thread* Re: [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM
2021-10-06 21:16 ` Laurent Vivier
@ 2021-10-07 9:19 ` Mark Cave-Ayland
0 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-07 9:19 UTC (permalink / raw)
To: Laurent Vivier, qemu-devel
On 06/10/2021 22:16, Laurent Vivier wrote:
> Le 06/10/2021 à 21:24, Laurent Vivier a écrit :
>> Le 06/10/2021 à 18:09, Mark Cave-Ayland a écrit :
>>> On 06/10/2021 16:46, Laurent Vivier wrote:
>>>> Le 06/10/2021 à 15:54, Mark Cave-Ayland a écrit :
>>>>> On 06/10/2021 13:24, Laurent Vivier wrote:
>>>>>
>>>>>>> This is where it becomes a bit trickier, since technically booting Linux with -kernel you can use
>>>>>>> any supported values as long as everything fits in the video RAM which is why there isn't
>>>>>>> currently
>>>>>>> a hard-coded list :)
>>>>>>>
>>>>>>
>>>>>> We need the list of "supported values". I don't want to read the code or try values combination
>>>>>> until it works.
>>>>>>
>>>>>> In a perfect world, I would like to be able to use any value I want with "-kernel".
>>>>>>
>>>>>> For instance I was using "-g 1200x800x24" and it was working fine.
>>>>>>
>>>>>> Now I have:
>>>>>>
>>>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>>>>
>>>>>> If it's not possible (because the original hardware cannot provide it, and we don't know the base
>>>>>> address or things like that), we don't need the list of the display id, but the list of available
>>>>>> modes: (width,height,depth).
>>>>>>
>>>>>> Rougly, something like:
>>>>>>
>>>>>> qemu-system-m68k: unknown display mode: width 1200, height 800, depth 24
>>>>>> Available modes:
>>>>>> 1152x870x8
>>>>>> 1152x870x4
>>>>>> 1152x870x2
>>>>>> 1152x870x1
>>>>>> 800x600x24
>>>>>> 800x600x8
>>>>>> 800x600x4
>>>>>> 800x600x2
>>>>>> 800x600x1
>>>>>> 640x480x24
>>>>>> 640x480x8
>>>>>> 640x480x4
>>>>>> 640x480x2
>>>>>> 640x480x1
>>>>>>
>>>>>> diff --git a/hw/display/macfb.c b/hw/display/macfb.c
>>>>>> index 5b8812e9e7d8..4b352eb89c3f 100644
>>>>>> --- a/hw/display/macfb.c
>>>>>> +++ b/hw/display/macfb.c
>>>>>> @@ -438,6 +438,26 @@ static MacFbMode *macfb_find_mode(MacfbDisplayType display_type,
>>>>>> return NULL;
>>>>>> }
>>>>>>
>>>>>> +static gchar *macfb_mode_list(void)
>>>>>> +{
>>>>>> + gchar *list = NULL;
>>>>>> + gchar *mode;
>>>>>> + MacFbMode *macfb_mode;
>>>>>> + int i;
>>>>>> +
>>>>>> + for (i = 0; i < ARRAY_SIZE(macfb_mode_table); i++) {
>>>>>> + macfb_mode = &macfb_mode_table[i];
>>>>>> +
>>>>>> + mode = g_strdup_printf(" %dx%dx%d\n", macfb_mode->width,
>>>>>> + macfb_mode->height, macfb_mode->depth);
>>>>>> + list = g_strconcat(mode, list, NULL);
>>>>>> + g_free(mode);
>>>>>> + }
>>>>>> +
>>>>>> + return list;
>>>>>> +}
>>>>>> +
>>>>>> +
>>>>>> static void macfb_update_display(void *opaque)
>>>>>> {
>>>>>> MacfbState *s = opaque;
>>>>>> @@ -620,8 +640,13 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error
>>>>>> **errp)
>>>>>>
>>>>>> s->mode = macfb_find_mode(s->type, s->width, s->height, s->depth);
>>>>>> if (!s->mode) {
>>>>>> + gchar *list;
>>>>>> error_setg(errp, "unknown display mode: width %d, height %d, depth %d",
>>>>>> s->width, s->height, s->depth);
>>>>>> + list = macfb_mode_list();
>>>>>> + error_append_hint(errp, "Available modes:\n%s", list);
>>>>>> + g_free(list);
>>>>>> +
>>>>>> return false;
>>>>>> }
>>>>>
>>>>> Hi Laurent,
>>>>>
>>>>> Thanks for the example - I can certainly squash this into patch 8.
>>>>
>>>> yes, please.
>>>
>>> Okay I'll do that for a v3 (and also split the 1st patch that Phil suggested).
>>>
>>>>> As for allowing extra resolutions via -kernel, since the check is being done in
>>>>> macfb_common_realize() then it would be possible to add a qdev property that only gets set when
>>>>> -kernel is used on the command line which bypasses the mode check if you prefer?
>>>>>
>>>>
>>>> I think it can wait and be done by a patch later. For the moment we can focus on MacOS.
>>>>
>>>>> I'm not sure that your existing example of "-g 1200x800x24" (or indeed any resolution with 24-bit
>>>>> depth) with -kernel will still work after this patchset given that the 24-bit encoding scheme has
>>>>> changed. Presumably this would also need a corresponding change in the bootinfo/kernel framebuffer/X
>>>>> configuration somewhere?
>>>>
>>>> The kernel framebuffer should be easy to fix, if needed, normally we pass the needed information via
>>>> the bootinfo structure.
>>>>
>>>> My X configuration is broken for a while. With debian/sid I've never been able to start X (even on a
>>>> real q800, I think), and with debian/etch we need a special kernel as the ADB stack has been broken
>>>> with old kernel. I was not able to start X for a while now...
>>>
>>> FWIW I found that the last set of ADB fixes in mac_via.c actually fixed ADB on old kernels again (I
>>> was able to use keyboard and mouse on the 4.15 kernel you used for the original patches), so you may
>>> be able to get debian/etch working in QEMU. I'd expect forcing EMILE into a 24-bit depth on a real
>>> Quadra 800 would also show the same issue here.
>>>
>>
>> In fact, I was trying the etch kernel, 2.6.32 that was never able to boot before because VBL
>> interrupt was not implemented. Now, with your series it boots (congratulation) but there is some
>> other issues in ADB but I don't think it's related to the changes in QEMU.
That's definitely progress! There are a couple of extra ADB fixes in my gitlab
q800.upstream branch for NetBSD which may or may not help here?
>> I've found another kernel on my disk (4.1.39) and I'm able to start a gnome-session with 1152x870x8
>> mode. ADB works well. There is an issue with gdm that doesn't take characters from the keyboard but
>> if I start X manually and then gnome-sessions I'm able to play with it.
>
> And I'm able to run a 800x600x24 session with the attached xorg.conf
>
> To run 1152x870, the DefaultDepth and DefaultFbBpp must be changed to 8.
Excellent - thanks for the example config. Assuming that this is matching the
behaviour on the real Quadra 800 then I'm happy these patches are moving the
emulation in the right direction :)
ATB,
Mark.
^ permalink raw reply [flat|nested] 26+ messages in thread
* [PATCH v2 09/12] macfb: fix up 1-bit pixel encoding
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (7 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 08/12] macfb: add common monitor modes supported by the MacOS toolbox ROM Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 10/12] macfb: fix 24-bit RGB " Mark Cave-Ayland
` (2 subsequent siblings)
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
The MacOS driver expects the RGB values for the pixel to be in entries 0 and 1
of the colour palette.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index 357fe18be5..ae0cf02dff 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -128,7 +128,9 @@ static void macfb_draw_line1(MacfbState *s, uint8_t *d, uint32_t addr,
for (x = 0; x < width; x++) {
int bit = x & 7;
int idx = (macfb_read_byte(s, addr) >> (7 - bit)) & 1;
- r = g = b = ((1 - idx) << 7);
+ r = s->color_palette[idx * 3];
+ g = s->color_palette[idx * 3 + 1];
+ b = s->color_palette[idx * 3 + 2];
addr += (bit == 7);
*(uint32_t *)d = rgb_to_pixel32(r, g, b);
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 10/12] macfb: fix 24-bit RGB pixel encoding
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (8 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 09/12] macfb: fix up 1-bit pixel encoding Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 11/12] macfb: add vertical blank interrupt Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 12/12] q800: wire macfb IRQ to separate video interrupt on VIA2 Mark Cave-Ayland
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
According to Apple Technical Note HW26: "Macintosh Quadra Built-In Video" the
in-built framebuffer encodes each 24-bit pixel into 4 bytes. Adjust the 24-bit
RGB pixel encoding accordingly which agrees with the encoding expected by MacOS
when changing into 24-bit colour mode.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index ae0cf02dff..ac5d46fa43 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -224,10 +224,10 @@ static void macfb_draw_line24(MacfbState *s, uint8_t *d, uint32_t addr,
int x;
for (x = 0; x < width; x++) {
- r = macfb_read_byte(s, addr);
- g = macfb_read_byte(s, addr + 1);
- b = macfb_read_byte(s, addr + 2);
- addr += 3;
+ r = macfb_read_byte(s, addr + 1);
+ g = macfb_read_byte(s, addr + 2);
+ b = macfb_read_byte(s, addr + 3);
+ addr += 4;
*(uint32_t *)d = rgb_to_pixel32(r, g, b);
d += 4;
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 11/12] macfb: add vertical blank interrupt
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (9 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 10/12] macfb: fix 24-bit RGB " Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
2021-10-04 21:19 ` [PATCH v2 12/12] q800: wire macfb IRQ to separate video interrupt on VIA2 Mark Cave-Ayland
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
The MacOS driver expects a 60.15Hz vertical blank interrupt to be generated by
the framebuffer which in turn schedules the mouse driver via the Vertical Retrace
Manager.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/display/macfb.c | 83 ++++++++++++++++++++++++++++++++++++++
include/hw/display/macfb.h | 8 ++++
2 files changed, 91 insertions(+)
diff --git a/hw/display/macfb.c b/hw/display/macfb.c
index ac5d46fa43..5b8812e9e7 100644
--- a/hw/display/macfb.c
+++ b/hw/display/macfb.c
@@ -33,9 +33,16 @@
#define DAFB_MODE_CTRL1 0x8
#define DAFB_MODE_CTRL2 0xc
#define DAFB_MODE_SENSE 0x1c
+#define DAFB_INTR_MASK 0x104
+#define DAFB_INTR_STAT 0x108
+#define DAFB_INTR_CLEAR 0x10c
#define DAFB_RESET 0x200
#define DAFB_LUT 0x213
+#define DAFB_INTR_VBL 0x4
+
+/* Vertical Blank period (60.15Hz) */
+#define DAFB_INTR_VBL_PERIOD_NS 16625800
/*
* Quadra sense codes taken from Apple Technical Note HW26:
@@ -450,6 +457,36 @@ static void macfb_update_display(void *opaque)
macfb_draw_graphic(s);
}
+static void macfb_update_irq(MacfbState *s)
+{
+ uint32_t irq_state = s->irq_state & s->irq_mask;
+
+ if (irq_state) {
+ qemu_irq_raise(s->irq);
+ } else {
+ qemu_irq_lower(s->irq);
+ }
+}
+
+static int64_t macfb_next_vbl(void)
+{
+ return (qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + DAFB_INTR_VBL_PERIOD_NS) /
+ DAFB_INTR_VBL_PERIOD_NS * DAFB_INTR_VBL_PERIOD_NS;
+}
+
+static void macfb_vbl_timer(void *opaque)
+{
+ MacfbState *s = opaque;
+ int64_t next_vbl;
+
+ s->irq_state |= DAFB_INTR_VBL;
+ macfb_update_irq(s);
+
+ /* 60 Hz irq */
+ next_vbl = macfb_next_vbl();
+ timer_mod(s->vbl_timer, next_vbl);
+}
+
static void macfb_reset(MacfbState *s)
{
int i;
@@ -478,6 +515,9 @@ static uint64_t macfb_ctrl_read(void *opaque,
case DAFB_MODE_CTRL2:
val = s->regs[addr >> 2];
break;
+ case DAFB_INTR_STAT:
+ val = s->irq_state;
+ break;
case DAFB_MODE_SENSE:
val = macfb_sense_read(s);
break;
@@ -493,6 +533,8 @@ static void macfb_ctrl_write(void *opaque,
unsigned int size)
{
MacfbState *s = opaque;
+ int64_t next_vbl;
+
switch (addr) {
case DAFB_MODE_VADDR1:
case DAFB_MODE_VADDR2:
@@ -508,8 +550,23 @@ static void macfb_ctrl_write(void *opaque,
case DAFB_MODE_SENSE:
macfb_sense_write(s, val);
break;
+ case DAFB_INTR_MASK:
+ s->irq_mask = val;
+ if (val & DAFB_INTR_VBL) {
+ next_vbl = macfb_next_vbl();
+ timer_mod(s->vbl_timer, next_vbl);
+ } else {
+ timer_del(s->vbl_timer);
+ }
+ break;
+ case DAFB_INTR_CLEAR:
+ s->irq_state &= ~DAFB_INTR_VBL;
+ macfb_update_irq(s);
+ break;
case DAFB_RESET:
s->palette_current = 0;
+ s->irq_state &= ~DAFB_INTR_VBL;
+ macfb_update_irq(s);
break;
case DAFB_LUT:
s->color_palette[s->palette_current] = val;
@@ -586,6 +643,7 @@ static bool macfb_common_realize(DeviceState *dev, MacfbState *s, Error **errp)
s->vram_bit_mask = MACFB_VRAM_SIZE - 1;
memory_region_set_coalescing(&s->mem_vram);
+ s->vbl_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, macfb_vbl_timer, s);
macfb_update_mode(s);
return true;
}
@@ -601,6 +659,16 @@ static void macfb_sysbus_realize(DeviceState *dev, Error **errp)
sysbus_init_mmio(SYS_BUS_DEVICE(s), &ms->mem_ctrl);
sysbus_init_mmio(SYS_BUS_DEVICE(s), &ms->mem_vram);
+
+ qdev_init_gpio_out(dev, &ms->irq, 1);
+}
+
+static void macfb_nubus_set_irq(void *opaque, int n, int level)
+{
+ MacfbNubusState *s = NUBUS_MACFB(opaque);
+ NubusDevice *nd = NUBUS_DEVICE(s);
+
+ nubus_set_irq(nd, level);
}
static void macfb_nubus_realize(DeviceState *dev, Error **errp)
@@ -621,6 +689,19 @@ static void macfb_nubus_realize(DeviceState *dev, Error **errp)
memory_region_add_subregion(&nd->slot_mem, DAFB_BASE, &ms->mem_ctrl);
memory_region_add_subregion(&nd->slot_mem, VIDEO_BASE, &ms->mem_vram);
+
+ ms->irq = qemu_allocate_irq(macfb_nubus_set_irq, s, 0);
+}
+
+static void macfb_nubus_unrealize(DeviceState *dev)
+{
+ MacfbNubusState *s = NUBUS_MACFB(dev);
+ MacfbNubusDeviceClass *ndc = NUBUS_MACFB_GET_CLASS(dev);
+ MacfbState *ms = &s->macfb;
+
+ ndc->parent_unrealize(dev);
+
+ qemu_free_irq(ms->irq);
}
static void macfb_sysbus_reset(DeviceState *d)
@@ -671,6 +752,8 @@ static void macfb_nubus_class_init(ObjectClass *klass, void *data)
device_class_set_parent_realize(dc, macfb_nubus_realize,
&ndc->parent_realize);
+ device_class_set_parent_unrealize(dc, macfb_nubus_unrealize,
+ &ndc->parent_unrealize);
dc->desc = "Nubus Macintosh framebuffer";
dc->reset = macfb_nubus_reset;
dc->vmsd = &vmstate_macfb;
diff --git a/include/hw/display/macfb.h b/include/hw/display/macfb.h
index 0aff0d84d2..e52775aa21 100644
--- a/include/hw/display/macfb.h
+++ b/include/hw/display/macfb.h
@@ -14,7 +14,9 @@
#define MACFB_H
#include "exec/memory.h"
+#include "hw/irq.h"
#include "ui/console.h"
+#include "qemu/timer.h"
#include "qom/object.h"
typedef enum {
@@ -63,6 +65,11 @@ typedef struct MacfbState {
uint32_t regs[MACFB_NUM_REGS];
MacFbMode *mode;
+
+ uint32_t irq_state;
+ uint32_t irq_mask;
+ QEMUTimer *vbl_timer;
+ qemu_irq irq;
} MacfbState;
#define TYPE_MACFB "sysbus-macfb"
@@ -81,6 +88,7 @@ struct MacfbNubusDeviceClass {
DeviceClass parent_class;
DeviceRealize parent_realize;
+ DeviceUnrealize parent_unrealize;
};
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread* [PATCH v2 12/12] q800: wire macfb IRQ to separate video interrupt on VIA2
2021-10-04 21:19 [PATCH v2 00/12] macfb: fixes for booting MacOS Mark Cave-Ayland
` (10 preceding siblings ...)
2021-10-04 21:19 ` [PATCH v2 11/12] macfb: add vertical blank interrupt Mark Cave-Ayland
@ 2021-10-04 21:19 ` Mark Cave-Ayland
11 siblings, 0 replies; 26+ messages in thread
From: Mark Cave-Ayland @ 2021-10-04 21:19 UTC (permalink / raw)
To: qemu-devel, laurent
Whilst the in-built Quadra 800 framebuffer exists within the Nubus address
space for slot 9, it has its own dedicated interrupt on VIA2. Force the
macfb device to occupy slot 9 in the q800 machine and wire its IRQ to the
separate video interrupt since this is what is expected by the MacOS
interrupt handler.
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
---
hw/m68k/q800.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/hw/m68k/q800.c b/hw/m68k/q800.c
index df3fd3711e..fd4855047e 100644
--- a/hw/m68k/q800.c
+++ b/hw/m68k/q800.c
@@ -407,8 +407,10 @@ static void q800_init(MachineState *machine)
MAC_NUBUS_FIRST_SLOT * NUBUS_SUPER_SLOT_SIZE);
sysbus_mmio_map(SYS_BUS_DEVICE(dev), 1, NUBUS_SLOT_BASE +
MAC_NUBUS_FIRST_SLOT * NUBUS_SLOT_SIZE);
-
- for (i = 0; i < VIA2_NUBUS_IRQ_NB; i++) {
+ qdev_connect_gpio_out(dev, 9,
+ qdev_get_gpio_in_named(via2_dev, "nubus-irq",
+ VIA2_NUBUS_IRQ_INTVIDEO));
+ for (i = 1; i < VIA2_NUBUS_IRQ_NB; i++) {
qdev_connect_gpio_out(dev, 9 + i,
qdev_get_gpio_in_named(via2_dev, "nubus-irq",
VIA2_NUBUS_IRQ_9 + i));
@@ -419,6 +421,7 @@ static void q800_init(MachineState *machine)
/* framebuffer in nubus slot #9 */
dev = qdev_new(TYPE_NUBUS_MACFB);
+ qdev_prop_set_uint32(dev, "slot", 9);
qdev_prop_set_uint32(dev, "width", graphic_width);
qdev_prop_set_uint32(dev, "height", graphic_height);
qdev_prop_set_uint8(dev, "depth", graphic_depth);
--
2.20.1
^ permalink raw reply related [flat|nested] 26+ messages in thread