OpenBIOS May 2014

openbios@openbios.org

10 participants
35 discussions

[commit] r1287 - trunk/openbios-devel/arch/sparc64

by repository service

Author: mcayland Date: Mon May 12 17:43:18 2014 New Revision: 1287 URL: http://tracker.coreboot.org/trac/openbios/changeset/1287 Log: SPARC64: fixup 100Hz timer interval As we are attempting to emulate a 100MHz CPU, then in order to get a 100Hz timer interrupt the cycle interval needs to be 1MHz and not 10MHz. Signed-off-by: Mark Cave-Ayland <mark.cave-ayland(a)ilande.co.uk> Modified: trunk/openbios-devel/arch/sparc64/entry.S trunk/openbios-devel/arch/sparc64/vectors.S Modified: trunk/openbios-devel/arch/sparc64/entry.S ============================================================================== --- trunk/openbios-devel/arch/sparc64/entry.S Fri Apr 4 11:46:24 2014 (r1286) +++ trunk/openbios-devel/arch/sparc64/entry.S Mon May 12 17:43:18 2014 (r1287) @@ -17,7 +17,7 @@ #define PROM_ADDR 0x1fff0000000 #define CFG_ADDR 0x1fe02000510 -#define HZ 10 * 1000 * 1000 +#define HZ 1 * 1000 * 1000 #define TICK_INT_DIS 0x8000000000000000 .globl entry, _entry Modified: trunk/openbios-devel/arch/sparc64/vectors.S ============================================================================== --- trunk/openbios-devel/arch/sparc64/vectors.S Fri Apr 4 11:46:24 2014 (r1286) +++ trunk/openbios-devel/arch/sparc64/vectors.S Mon May 12 17:43:18 2014 (r1287) @@ -30,7 +30,7 @@ #define PROM_ADDR 0x1fff0000000 #define SER_ADDR 0x1fe020003f8 #define TICK_INT_DIS 0x8000000000000000 -#define TICK_INTERVAL 10*1000*1000 +#define TICK_INTERVAL 1*1000*1000 .section ".text.vectors", "ax" .align 16384

7 years, 4 months

usb driver to support usb keyboard

by BALATON Zoltan

Hello, I'm trying to improve the emulation of new world Macs (mac99) in QEMU and noticed that OpenBIOS does not support usb keyboards thus it currently cannot be used with qemu-system-ppc64 -machine mac99 and still uses an ADB keyboard for new world G4 mac that usually have USB keyboards. Because of this I have looked at adding a usb driver to OpenBIOS. I was looking for other open source drivers to adapt but most I could find were GPLv3 (Grub), LGPLv3 (SeaBios) which are incompatible with the current OpenBIOS license or too difficult to adapt (Linux). This mostly left the driver in Coreboot's libpayload library (http://review.coreboot.org/gitweb?p=coreboot.git;a=tree;f=payloads/libpaylo…) which is BSD licensed that I gave a try. It wasn't too difficult to port to OpenBIOS but when I tried to use it it turned out that it does not handle the device being little endian on a big endian host and thus does not work correctly with ppc (but it might work on little endian hosts, I didn't test that; Coreboot libpayload seems to only support x86 and armv7.) Fixing this is a bit tedious that I don't feel like doing now, I'd rather work on other problems towards my goal of running MorphOS on QEMU instead but I'd like to share the results so far in case someone wants to look at it, has any idea how to easily fix the endianness problem or maybe help with finishing it. Or if someone knows about another simple boot time driver that could be adapted instead and works correctly on big endian hosts please let me know. The current work in progress version can be found here: http://goliat.eik.bme.hu/~balaton/openbios-usbhid-driver-wip.tar.xz Any comments/help are welcome. Regards, BALATON Zoltan

7 years, 4 months

[PATCH v3] Updated some device tree properties to match names used by Apple for compatibility with MorphOS

by BALATON Zoltan

The names of some device tree properties are different on old world and new world Mac hardware and the name of the root node is always "device-tree" on Apple's OpenFirmware implementation. MorphOS on Apple hardware expects these names to find these devices. Signed-off-by: BALATON Zoltan <balaton(a)eik.bme.hu> --- v2: improved commit log v3: fixed a typo in the device type of the pmu node (via-pmu instead of via_pmu) --- Index: openbios-devel/arch/ppc/qemu/init.c =================================================================== --- openbios-devel/arch/ppc/qemu/init.c (revision 1286) +++ openbios-devel/arch/ppc/qemu/init.c (working copy) @@ -702,6 +732,12 @@ push_str("/"); fword("find-device"); + /* Apple calls the root node device-tree */ + if (is_apple()) { + push_str("device-tree"); + fword("device-name"); + } + switch(machine_id) { case ARCH_HEATHROW: /* OldWorld */ Index: openbios-devel/drivers/cuda.c =================================================================== --- openbios-devel/drivers/cuda.c (revision 1286) +++ openbios-devel/drivers/cuda.c (working copy) @@ -184,32 +184,33 @@ phandle_t ph=get_cur_dev(); int props[2]; - push_str("via-cuda"); + push_str((is_oldworld() ? "via-cuda" : "via-pmu")); fword("device-type"); set_int_property(ph, "#address-cells", 1); set_int_property(ph, "#size-cells", 0); - set_property(ph, "compatible", "cuda", 5); - props[0] = __cpu_to_be32(IO_CUDA_OFFSET); props[1] = __cpu_to_be32(IO_CUDA_SIZE); set_property(ph, "reg", (char *)&props, sizeof(props)); - /* on newworld machines the cuda is on interrupt 0x19 */ + if (is_oldworld()) { + set_property(ph, "compatible", "cuda", 5); + /* we emulate an oldworld hardware, so we must use + * non-standard oldworld property (needed by linux 2.6.18) + */ + set_int_property(ph, "AAPL,interrupts", 0x12); + } else { + set_property(ph, "compatible", "pmu", 4); + /* on newworld machines the cuda is on interrupt 0x19 */ - props[0] = 0x19; - props[1] = 0; - NEWWORLD(set_property(ph, "interrupts", (char *)props, sizeof(props))); - NEWWORLD(set_int_property(ph, "#interrupt-cells", 2)); + props[0] = 0x19; + props[1] = 0; + set_property(ph, "interrupts", (char *)&props, sizeof(props)); + set_int_property(ph, "#interrupt-cells", 2); + } - /* we emulate an oldworld hardware, so we must use - * non-standard oldworld property (needed by linux 2.6.18) - */ - - OLDWORLD(set_int_property(ph, "AAPL,interrupts", 0x12)); - bind_func("ppc32-reset-all", ppc32_reset_all); push_str("' ppc32-reset-all to reset-all"); fword("eval"); @@ -380,7 +381,8 @@ ph = find_dev(buf); set_property(ph, "device_type", "power-mgt", 10); - set_property(ph, "compatible", "power-mgt", 10); + OLDWORLD(set_property(ph, "compatible", "power-mgt", 10)); + NEWWORLD(set_property(ph, "compatible", "via-pmu-99", 11)); } cuda_t *cuda_init (const char *path, phys_addr_t base) @@ -395,11 +397,13 @@ if (cuda == NULL) return NULL; - snprintf(buf, sizeof(buf), "%s/via-cuda", path); + OLDWORLD(snprintf(buf, sizeof(buf), "%s/via-cuda", path)); + NEWWORLD(snprintf(buf, sizeof(buf), "%s/via-pmu", path)); REGISTER_NAMED_NODE(ob_cuda, buf); aliases = find_dev("/aliases"); - set_property(aliases, "via-cuda", buf, strlen(buf) + 1); + OLDWORLD(set_property(aliases, "via-cuda", buf, strlen(buf) + 1)); + NEWWORLD(set_property(aliases, "via-pmu", buf, strlen(buf) + 1)); cuda->base = base; cuda_writeb(cuda, B, cuda_readb(cuda, B) | TREQ | TIP);

7 years, 4 months

[PATCH 0/7] SPARC64: *BSD fixup patchset

by Mark Cave-Ayland

This patchset is a result of my work to get the *BSD kernels to boot under QEMU SPARC64, and should be applied on top of the timer patchset previously posted. With these patches applied we now get the following: - NetBSD kernels start and enumerate PCI devices but panic due to missing device properties - Old OpenBSD kernels start and enumerate PCI devices but panic due to missing device properties - New OpenBSD kernels trap soon after trap table takeover (possible window fill/spill issue?) - FreeBSD kernels start and die during PCI device enumeration similar to above but seem to stop console output after trap table takeover? The first four patches fix bugs which enable us to unmap page zero on SPARC64 to catch NULL dereferences (in line with all other architectures) and helped further reveal the bugs fixed by the subsequent patches. Mark Cave-Ayland (7): pc_serial.c: fix up missing address property pc_kbd.c: fix up reg property and missing address property SPARC64: fix NULL pointer reference when attempting kernel boot SPARC64: Remove zero page mapping from MMU to enable detection of NULL pointer dereferences SPARC64: fix up translation property mode to make it a valid TTE ciface.fs: handle buggy callers to test-method SPARC64: only map 8MB RAM on startup openbios-devel/arch/sparc64/boot.c | 40 ++++++++++++++------------- openbios-devel/arch/sparc64/ofmem_sparc64.c | 6 ++-- openbios-devel/drivers/pc_kbd.c | 15 +++++++++- openbios-devel/drivers/pc_serial.c | 6 ++++ openbios-devel/forth/system/ciface.fs | 25 +++++++++++++---- 5 files changed, 64 insertions(+), 28 deletions(-) -- 1.7.10.4

7 years, 4 months

[PATCH 0/5] SPARC64 timer implementation and timer fixes

by Mark Cave-Ayland

This patchset came out of the work from a following SPARC64 patchset and contains a basic SPARC64 timer implementation based upon the %tick register. In addition to this, there is some extra refactoring to enable get-msecs to work across all platforms, and then finally a patch to use this to realise an implementation of the milliseconds CIF service. Mark Cave-Ayland (5): SPARC64: fixup 100Hz timer interval SPARC64: increment ms counter at obp_tick_pointer on every timer interrupt SPARC64: add basic get-msecs implementation other.fs: rework get-msecs word so it can be shared cross-platform ciface.fs: implement milliseconds service openbios-devel/arch/sparc64/entry.S | 2 +- openbios-devel/arch/sparc64/openbios.c | 6 ++++++ openbios-devel/arch/sparc64/vectors.S | 20 ++++++++++++++++++-- openbios-devel/forth/device/other.fs | 21 +++++++++------------ openbios-devel/forth/system/ciface.fs | 5 +++-- 5 files changed, 37 insertions(+), 17 deletions(-) -- 1.7.10.4

7 years, 4 months