October 2015 - OpenBIOS - mail.coreboot.org

[PATCH v2 0/4] Build openbios on powerpc64 host
by Laurent Vivier Nov. 6, 2015

Nov. 6, 2015

This series of patches allows to build openbios on a powerpc64 host. It has been tested with the following targets: ppc, sparc32, sparc64. It has also been tested on x86_64 host with the same targets. v2: remove patches 3/5 and 4/5 of the first series and replace it with patch 4/4. We don't include the system includes using full pathname but use '-iquote' intead of '-I' A preliminary patch is needed for fcode-utils, as "toke" is broken on big endian architecture: diff -ru fcode-utils/toke/scanner.c fcode-utils-new/toke/scanner.c --- fcode-utils/toke/scanner.c 2006-10-30 09:48:28.000000000 +0100 +++ fcode-utils-new/toke/scanner.c 2015-10-19 00:21:43.417373690 +0200 @@ -1625,7 +1625,7 @@ void user_message( tic_param_t pfield ) { - char delim = (char)pfield.deflt_elem ; + char delim = (char)pfield.fw_token ; handle_user_message( delim, TRUE); } @@ -5295,7 +5295,7 @@ void process_remark( tic_param_t pfield ) { - char until_char = (char)pfield.deflt_elem ; + char until_char = (char)pfield.fw_token ; unsigned int start_lineno = lineno; #ifdef DEBUG_SCANNER Laurent Vivier (4): switch-arch: factorize code to compute architecture properties. switch-arch: compute base arch and allow native compiler for 32bit and 64bit. switch-arch: as for powerpc64, select sparc64 compiler to compile sparc32 bootstrap: don't include files from target/include as system includes. Makefile.target | 2 +- config/scripts/switch-arch | 85 +++++++++++++++++++++++++--------------------- include/arch/ppc/types.h | 8 ----- 3 files changed, 48 insertions(+), 47 deletions(-) -- 2.4.3

4 9

[PATCH 0/5] Build openbios on powerpc64 host
by Laurent Vivier Oct. 24, 2015

Oct. 24, 2015

This series of patches allows to build openbios on a powerpc64 host. It has been tested with following targets: ppc, sparc32, sparc64. A preliminary patch is needed for fcode-utils, as "toke" is broken on big endan architecture: diff -ru fcode-utils/toke/scanner.c fcode-utils-new/toke/scanner.c --- fcode-utils/toke/scanner.c 2006-10-30 09:48:28.000000000 +0100 +++ fcode-utils-new/toke/scanner.c 2015-10-19 00:21:43.417373690 +0200 @@ -1625,7 +1625,7 @@ void user_message( tic_param_t pfield ) { - char delim = (char)pfield.deflt_elem ; + char delim = (char)pfield.fw_token ; handle_user_message( delim, TRUE); } @@ -5295,7 +5295,7 @@ void process_remark( tic_param_t pfield ) { - char until_char = (char)pfield.deflt_elem ; + char until_char = (char)pfield.fw_token ; unsigned int start_lineno = lineno; #ifdef DEBUG_SCANNER Laurent Vivier (5): switch-arch: factorize code to compute architecture properties. switch-arch: compute base arch and allow native compiler for 32bit and 64bit. On ppc64, asm/sigcontext.h includes asm/elf.h On ppc64 host, elf_vrreg_t is defined as __vector128 switch-arch: as for powerpc64, select sparc64 compiler to compile sparc32 config/scripts/switch-arch | 85 ++++++++++++++++++++++++-------------------- include/arch/ppc/elf.h | 4 +++ include/arch/ppc/types.h | 5 +++ include/arch/sparc32/elf.h | 4 +++ include/arch/sparc32/types.h | 13 +++++++ include/arch/sparc64/elf.h | 4 +++ include/arch/sparc64/types.h | 13 +++++++ 7 files changed, 90 insertions(+), 38 deletions(-) -- 2.4.3

3 15

[commit] r1353 - in trunk/openbios-devel: arch/ppc/qemu drivers include/drivers
by repository service Oct. 23, 2015

Oct. 23, 2015

Author: mcayland Date: Fri Oct 23 14:58:22 2015 New Revision: 1353 URL: http://tracker.coreboot.org/trac/openbios/changeset/1353 Log: ppc: add ob_ide_quiesce() and call it from the PPC CIF quiesce word Add a new ob_ide_quiesce() function and hook it into QEMU PPC's ciface_quiesce() routine to properly prepare the drives for transferring control to the operating system. Without this OS 9 becomes confused about the initial drive state on boot and refuses to correctly detect and initialise it. Based upon an original patch by Cormac O'Brien <cormac(a)c-obrien.org>. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/arch/ppc/qemu/methods.c trunk/openbios-devel/drivers/ide.c trunk/openbios-devel/include/drivers/drivers.h Modified: trunk/openbios-devel/arch/ppc/qemu/methods.c ============================================================================== --- trunk/openbios-devel/arch/ppc/qemu/methods.c Fri Oct 23 14:58:20 2015 (r1352) +++ trunk/openbios-devel/arch/ppc/qemu/methods.c Fri Oct 23 14:58:22 2015 (r1353) @@ -114,6 +114,8 @@ ciface_quiesce( unsigned long args[], unsigned long ret[] ) { usb_exit(); + + ob_ide_quiesce(); #if 0 unsigned long msr; /* This seems to be the correct thing to do - but I'm not sure */ Modified: trunk/openbios-devel/drivers/ide.c ============================================================================== --- trunk/openbios-devel/drivers/ide.c Fri Oct 23 14:58:20 2015 (r1352) +++ trunk/openbios-devel/drivers/ide.c Fri Oct 23 14:58:22 2015 (r1353) @@ -1488,6 +1488,28 @@ return 0; } +void ob_ide_quiesce(void) +{ + struct ide_channel *channel; + int i; + + channel = channels; + while (channel) { + for (i = 0; i < 2; i++) { + struct ide_drive *drive = &channel->drives[i]; + + if (!drive->present) + continue; + + ob_ide_select_drive(drive); + ob_ide_software_reset(drive); + ob_ide_device_type_check(drive); + } + + channel = channel->next; + } +} + #if defined(CONFIG_DRIVER_MACIO) static unsigned char macio_ide_inb(struct ide_channel *chan, unsigned int port) Modified: trunk/openbios-devel/include/drivers/drivers.h ============================================================================== --- trunk/openbios-devel/include/drivers/drivers.h Fri Oct 23 14:58:20 2015 (r1352) +++ trunk/openbios-devel/include/drivers/drivers.h Fri Oct 23 14:58:22 2015 (r1353) @@ -52,6 +52,7 @@ /* drivers/ide.c */ int ob_ide_init(const char *path, uint32_t io_port0, uint32_t ctl_port0, uint32_t io_port1, uint32_t ctl_port1); +void ob_ide_quiesce(void); int macio_ide_init(const char *path, uint32_t addr, int nb_channels); #endif #ifdef CONFIG_DRIVER_ESP

1 0

[commit] r1352 - trunk/openbios-devel/drivers
by repository service Oct. 23, 2015

Oct. 23, 2015

Author: mcayland Date: Fri Oct 23 14:58:20 2015 New Revision: 1352 URL: http://tracker.coreboot.org/trac/openbios/changeset/1352 Log: ppc: fix up IDE device tree Mac OS 9 is finicky about devices having the correct properties, so we change some IDE data to more closely model a real New World Mac. Signed-off-by: Cormac O'Brien <cormac(a)c-obrien.org> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/drivers/ide.c Modified: trunk/openbios-devel/drivers/ide.c ============================================================================== --- trunk/openbios-devel/drivers/ide.c Fri Oct 23 14:58:18 2015 (r1351) +++ trunk/openbios-devel/drivers/ide.c Fri Oct 23 14:58:20 2015 (r1352) @@ -1581,6 +1581,12 @@ set_property(dnode, "compatible", (is_oldworld() ? "heathrow-ata" : "keylargo-ata"), 13); + set_property(dnode, "model", ((current_channel == 3) ? + "ata-3" : "ata-4"), strlen("ata-*") + 1); + + set_property(dnode, "AAPL,connector", "ata", + strlen("ata") + 1); + props[0] = 0x00000526; props[1] = 0x00000085; props[2] = 0x00000025; @@ -1589,8 +1595,8 @@ props[5] = 0x00000000; props[6] = 0x00000000; props[7] = 0x00000000; - OLDWORLD(set_property(dnode, "AAPL,pio-timing", - (char *)&props, 8*sizeof(props[0]))); + set_property(dnode, "AAPL,pio-timing", + (char *)&props, 8*sizeof(props[0])); /* The first interrupt entry is the ide interrupt, the second the dbdma interrupt */ @@ -1634,8 +1640,8 @@ (char *)&props, 2*sizeof(props[0]))); props[0] = 0; - OLDWORLD(set_property(dnode, "AAPL,bus-id", (char*)props, - 1 * sizeof(props[0]))); + set_property(dnode, "AAPL,bus-id", (char*)props, + 1 * sizeof(props[0])); IDE_DPRINTF(DEV_NAME": [io ports 0x%lx]\n", current_channel, chan->mmio);

1 0

[commit] r1351 - trunk/openbios-devel/drivers
by repository service Oct. 23, 2015

Oct. 23, 2015

Author: mcayland Date: Fri Oct 23 14:58:18 2015 New Revision: 1351 URL: http://tracker.coreboot.org/trac/openbios/changeset/1351 Log: ppc: fix CUDA ADB error packet detection CUDA ADB error packets contain a 3-byte payload rather than a standard 2-byte header plus payload (as seen in QEMU <= 2.4). Allow cuda_adb_req() to differentiate between the two formats and act accordingly. Signed-off-by: Cormac O'Brien <cormac(a)c-obrien.org> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/drivers/cuda.c Modified: trunk/openbios-devel/drivers/cuda.c ============================================================================== --- trunk/openbios-devel/drivers/cuda.c Fri Oct 23 14:58:16 2015 (r1350) +++ trunk/openbios-devel/drivers/cuda.c Fri Oct 23 14:58:18 2015 (r1351) @@ -144,8 +144,22 @@ // CUDA_DPRINTF("len: %d %02x\n", len, snd_buf[0]); len = cuda_request(host, ADB_PACKET, snd_buf, len, buffer); if (len > 1 && buffer[0] == ADB_PACKET) { - pos = buffer + 2; - len -= 2; + /* We handle 2 types of ADB packet here: + Normal: <type> <status> <data> ... + Error : <type> <status> <cmd> (<data> ...) + Ideally we should use buffer[1] (status) to determine whether this + is a normal or error packet but this requires a corresponding fix + in QEMU <= 2.4. Hence we temporarily handle it this way to ease + the transition. */ + if (len > 2 && buffer[2] == snd_buf[0]) { + /* Error */ + pos = buffer + 3; + len -= 3; + } else { + /* Normal */ + pos = buffer + 2; + len -= 2; + } } else { pos = buffer + 1; len = -1;

1 0

[commit] r1350 - trunk/openbios-devel/drivers
by repository service Oct. 23, 2015

Oct. 23, 2015

Author: mcayland Date: Fri Oct 23 14:58:16 2015 New Revision: 1350 URL: http://tracker.coreboot.org/trac/openbios/changeset/1350 Log: ppc: add ESCC legacy support Mac OS 9 requires that the host provide escc-legacy serial ports, so we add them here to get further in the boot process. Signed-off-by: Cormac O'Brien <cormac(a)c-obrien.org> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/drivers/escc.c trunk/openbios-devel/drivers/escc.h trunk/openbios-devel/drivers/pci.c Modified: trunk/openbios-devel/drivers/escc.c ============================================================================== --- trunk/openbios-devel/drivers/escc.c Fri Oct 23 14:58:13 2015 (r1349) +++ trunk/openbios-devel/drivers/escc.c Fri Oct 23 14:58:16 2015 (r1350) @@ -380,12 +380,22 @@ static void escc_add_channel(const char *path, const char *node, phys_addr_t addr, - uint32_t offset) + int esnum) { char buf[64], tty[32]; phandle_t dnode, aliases; - int len; - cell props[2]; + + cell props[10]; + int offset; + int legacy; + + switch (esnum) { + case 2: offset = 1; legacy = 0; break; + case 3: offset = 0; legacy = 0; break; + case 4: offset = 1; legacy = 1; break; + case 5: offset = 0; legacy = 1; break; + default: return; + } /* add device */ @@ -411,16 +421,28 @@ set_property(dnode, "device_type", "serial", strlen("serial") + 1); - snprintf(buf, sizeof(buf), "ch-%s", node); - len = strlen(buf) + 1; - snprintf(buf + len, sizeof(buf) - len, "CHRP,es2"); - set_property(dnode, "compatible", buf, len + 9); - - props[0] = IO_ESCC_OFFSET + offset * 0x20; - props[1] = 0x00000020; - set_property(dnode, "reg", (char *)&props, 2 * sizeof(cell)); + snprintf(buf, sizeof(buf), "chrp,es%d", esnum); + set_property(dnode, "compatible", buf, 9); - props[0] = addr + IO_ESCC_OFFSET + offset * 0x20; + if (legacy) { + props[0] = IO_ESCC_LEGACY_OFFSET + offset * 0x4; + props[1] = 0x00000001; + props[2] = IO_ESCC_LEGACY_OFFSET + offset * 0x4 + 2; + props[3] = 0x00000001; + props[4] = IO_ESCC_LEGACY_OFFSET + offset * 0x4 + 6; + props[5] = 0x00000001; + set_property(dnode, "reg", (char *)&props, 6 * sizeof(cell)); + } else { + props[0] = IO_ESCC_OFFSET + offset * 0x20; + props[1] = 0x00000020; + set_property(dnode, "reg", (char *)&props, 2 * sizeof(cell)); + } + + if (legacy) { + props[0] = addr + IO_ESCC_LEGACY_OFFSET + offset * 0x4; + } else { + props[0] = addr + IO_ESCC_OFFSET + offset * 0x20; + } OLDWORLD(set_property(dnode, "AAPL,address", (char *)&props, 1 * sizeof(cell))); @@ -430,13 +452,21 @@ props[0] = (0x24) + offset; props[1] = 0; + props[2] = 0; NEWWORLD(set_property(dnode, "interrupts", - (char *)&props, 2 * sizeof(cell))); + (char *)&props, 3 * sizeof(cell))); device_end(); - uart_init_line((unsigned char*)addr + IO_ESCC_OFFSET + offset * 0x20, - CONFIG_SERIAL_SPEED); + if (legacy) { + uart_init_line( + (unsigned char*)addr + IO_ESCC_LEGACY_OFFSET + offset * 0x4, + CONFIG_SERIAL_SPEED); + } else { + uart_init_line( + (unsigned char*)addr + IO_ESCC_OFFSET + offset * 0x20, + CONFIG_SERIAL_SPEED); + } } void @@ -467,10 +497,34 @@ fword("finish-device"); - escc_add_channel(buf, "a", addr, 1); - escc_add_channel(buf, "b", addr, 0); + escc_add_channel(buf, "a", addr, 2); + escc_add_channel(buf, "b", addr, 3); escc_serial_dev = (unsigned char *)addr + IO_ESCC_OFFSET + (CONFIG_SERIAL_PORT ? 0 : 0x20); + + push_str(path); + fword("find-device"); + fword("new-device"); + + push_str("escc-legacy"); + fword("device-name"); + + snprintf(buf, sizeof(buf), "%s/escc-legacy", path); + + dnode = find_dev(buf); + + set_int_property(dnode, "#address-cells", 1); + props[0] = __cpu_to_be32(IO_ESCC_LEGACY_OFFSET); + props[1] = __cpu_to_be32(IO_ESCC_LEGACY_SIZE); + set_property(dnode, "reg", (char *)&props, sizeof(props)); + set_property(dnode, "device_type", "escc-legacy", + strlen("escc-legacy") + 1); + set_property(dnode, "compatible", "chrp,es1", 9); + + fword("finish-device"); + + escc_add_channel(buf, "a", addr, 4); + escc_add_channel(buf, "b", addr, 5); } #endif Modified: trunk/openbios-devel/drivers/escc.h ============================================================================== --- trunk/openbios-devel/drivers/escc.h Fri Oct 23 14:58:13 2015 (r1349) +++ trunk/openbios-devel/drivers/escc.h Fri Oct 23 14:58:16 2015 (r1350) @@ -1,6 +1,8 @@ #define IO_ESCC_SIZE 0x00001000 #define IO_ESCC_OFFSET 0x00013000 +#define IO_ESCC_LEGACY_SIZE 0x00001000 +#define IO_ESCC_LEGACY_OFFSET 0x00012000 #define ZS_REGS 8 Modified: trunk/openbios-devel/drivers/pci.c ============================================================================== --- trunk/openbios-devel/drivers/pci.c Fri Oct 23 14:58:13 2015 (r1349) +++ trunk/openbios-devel/drivers/pci.c Fri Oct 23 14:58:16 2015 (r1350) @@ -1420,6 +1420,12 @@ target_node = find_dev("/pci/mac-io/escc/ch-b"); set_int_property(target_node, "interrupt-parent", dnode); + target_node = find_dev("/pci/mac-io/escc-legacy/ch-a"); + set_int_property(target_node, "interrupt-parent", dnode); + + target_node = find_dev("/pci/mac-io/escc-legacy/ch-b"); + set_int_property(target_node, "interrupt-parent", dnode); + /* QEMU only emulates 2 of the 3 ata buses currently */ /* On a new world Mac these are not numbered but named by the * ATA version they support. Thus we have: ata-3, ata-3, ata-4

1 0

[commit] r1349 - trunk/openbios-devel/drivers
by repository service Oct. 23, 2015

Oct. 23, 2015

Author: mcayland Date: Fri Oct 23 14:58:13 2015 New Revision: 1349 URL: http://tracker.coreboot.org/trac/openbios/changeset/1349 Log: ppc: fix up power-mgt properties Modify some basic device tree properties to more accurately emulate a real Mac machine. Signed-off-by: Cormac O'Brien <cormac(a)c-obrien.org> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/drivers/cuda.c Modified: trunk/openbios-devel/drivers/cuda.c ============================================================================== --- trunk/openbios-devel/drivers/cuda.c Sun Oct 4 21:24:13 2015 (r1348) +++ trunk/openbios-devel/drivers/cuda.c Fri Oct 23 14:58:13 2015 (r1349) @@ -380,7 +380,8 @@ ph = find_dev(buf); set_property(ph, "device_type", "power-mgt", 10); - set_property(ph, "compatible", "power-mgt", 10); + set_property(ph, "mgt-kind", "min-consumption-pwm-led", strlen("min-consumption-pwm-led") + 1); + set_property(ph, "compatible", "cuda", strlen("cuda") + 1); } cuda_t *cuda_init (const char *path, phys_addr_t base)

1 0

[PATCH 0/5] GSoC OS 9 boot patchset update
by Mark Cave-Ayland Oct. 23, 2015

Oct. 23, 2015

Hi all, Here is a slightly reworked version of Cormac's GSoC MacOS 9 patchset which I believe is ready to apply. A considerable amount of time was spent by Cormac finding the correct set of properties required by OS 9 whilst also not regressing existing images. The patchset in it's current form here passes all of my local PPC image tests while still progressing OS 9 boot as before. The patches are mostly untouched other than adding my SoB except for patch 3 where I altered the version referenced in the comment and commit message, and splitting patch 4 into 2 separate patches as requested by Alex, along with some other minor tidy-ups. Note: patch 1 from Cormac's original patchset was a hack from me to allow the OS 9 bootloader to run, however it is not included here because it causes problems with yaboot. People interested in the OS 9 work can apply the patch by hand in the meantime. Several review comments suggested the use of a helper function for string properties, however recent work on SPARC has shown that building up string properties in this way is prevalent throughout the OpenBIOS codebase. In the interest of progressing work in this area, I'd like to apply the patches as-is and then globally rework string properties using a helper at a later date. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Cormac O'Brien (4): ppc: fix up power-mgt properties ppc: add ESCC legacy support ppc: fix CUDA ADB packet header format ppc: fix up IDE device tree and control transfer Mark Cave-Ayland (1): ppc: add ob_ide_quiesce() and call it from the PPC CIF quiesce word openbios-devel/arch/ppc/qemu/methods.c | 2 + openbios-devel/drivers/cuda.c | 17 +++++- openbios-devel/drivers/escc.c | 88 ++++++++++++++++++++++++------ openbios-devel/drivers/escc.h | 2 + openbios-devel/drivers/ide.c | 36 ++++++++++-- openbios-devel/drivers/pci.c | 6 ++ openbios-devel/include/drivers/drivers.h | 1 + 7 files changed, 128 insertions(+), 24 deletions(-) -- 1.7.10.4

2 8

[commit] r1348 - trunk/openbios-devel/drivers
by repository service Oct. 4, 2015

Oct. 4, 2015

Author: mcayland Date: Sun Oct 4 19:24:13 2015 New Revision: 1348 URL: http://tracker.coreboot.org/trac/openbios/changeset/1348 Log: SPARC64: fix PCI-EBus bridge ranges property The child address of the PCI-EBus bridge "ranges" property should be offset from the BAR address and not the beginning of the address space (similar as to how the "assigned-addresses" property is generated). This fixes EBus enumeration under FreeBSD SPARC64. Reported-by: Marius Strobl <marius(a)alchemy.franken.de> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/drivers/pci.c Modified: trunk/openbios-devel/drivers/pci.c ============================================================================== --- trunk/openbios-devel/drivers/pci.c Sun Jul 12 12:24:11 2015 (r1347) +++ trunk/openbios-devel/drivers/pci.c Sun Oct 4 19:24:13 2015 (r1348) @@ -824,7 +824,7 @@ ncells += pci_encode_phys_addr(props + ncells, flags, space_code, config->dev, PCI_BASE_ADDR_0 + (i * sizeof(uint32_t)), - 0); + config->assigned[i] & ~mask); props[ncells++] = config->sizes[i]; }

1 0