OpenBIOS August 2011

openbios@openbios.org

19 participants
88 discussions

sbus slot probing
by Bob Breuer 22 Aug '11

22 Aug '11

Here's a preliminary patch for sbus slot probing. This can be used for the preliminary dbri device patch, see my dbri-pre1 branch at github: https://github.com/breuerr/qemu/tree/dbri-pre1 That branch also has a rewritten fcode rom to avoid the use of the obsolete intr forth word. Use the SS-20 machine to get the dbri device. Todo: non-faulting reads so that an empty slot can be probed extend to probe other (all?) slots as well only add the slot E range for SS-20, but not SS-10 Is there any existing code for sparc32 which does a non-faulting read? It looks like even cpeek doesn't try to catch any faults. Bob Index: drivers/sbus.c =================================================================== --- drivers/sbus.c (revision 1046) +++ drivers/sbus.c (working copy) @@ -451,11 +451,36 @@ static void sbus_probe_slot_ss10(unsigned int slot, uint64_t base) { + ucell fcode; + unsigned char start; + // OpenBIOS and Qemu don't know how to do Sbus probing switch(slot) { case 2: // SUNW,tcx ob_tcx_init(slot, "/iommu/sbus/SUNW,tcx"); break; + case 0xe: // dbri + fcode = ofmem_map_io(base, 0x1000); + // TODO: non-faulting read + start = *(unsigned char*)fcode; + if (start == 0xfd || start == 0xf1) { + push_str("/iommu/sbus"); + fword("select-dev"); + fword("new-device"); + push_str(NULL); + PUSH(0); // offset + PUSH(slot); // space + push_str("encode-unit"); + fword("$call-parent"); + fword("set-args"); + PUSH(fcode); + PUSH(1); + fword("byte-load"); + fword("finish-device"); + fword("unselect-dev"); + } + ofmem_unmap(fcode, 0x1000); + break; case 0xf: // le, esp, bpp, power-management ob_macio_init(slot, base, 0); // Power management (APC) XXX should not exist @@ -534,6 +559,7 @@ { 1, 0, 0xe10000000ULL, 0x10000000,}, { 2, 0, 0xe20000000ULL, 0x10000000,}, { 3, 0, 0xe30000000ULL, 0x10000000,}, + [0xe] = { 0xe, 0, 0xee0000000ULL, 0x10000000,}, [0xf] = { 0xf, 0, 0xef0000000ULL, 0x10000000,}, };

3 3

[PATCH 10/10] Adding filll Forth primitive (v2)
by William Hahne 22 Aug '11

22 Aug '11

This is a forth word that is required by BootX. It just fills some specified memory address and length with 32 bit integers. It is a non-standard addition and therefore isn't documented. Thanks to Segher Boessenkool for the corrected for loop. Index: arch/ppc/qemu/init.c =================================================================== --- arch/ppc/qemu/init.c (revision 1047) +++ arch/ppc/qemu/init.c (working copy) @@ -539,6 +539,7 @@ } static void go(void); +static void filll(void); static void go(void) @@ -551,6 +552,23 @@ call_elf(0, 0, addr); } +/* This is an Apple specific word that fills a specified location and length + of memory with 32bit integers. + + ( dest count value -- ) */ +static void +filll(void) +{ + uint32_t value = POP(); + int count = POP(); + uint32_t *dest = (uint32_t*)cell2pointer(POP()); + + int i; + for (i = count; i > 0; i -= sizeof(uint32_t)) { + *dest++ = value; + } +} + static void kvm_of_init(void) { char hypercall[4 * 4]; @@ -879,4 +897,5 @@ bind_func("platform-boot", boot); bind_func("(go)", go); + bind_func("filll", filll); }

1 0

Re: [OpenBIOS] [PATCH 2/10] Ignore extraneous partitions
by William Hahne 22 Aug '11

22 Aug '11

Updated commit message: Currently every entry in the Apple Partition Map is counted as a partition. This behavior prevents BootX from finding the boot partition since BootX will only check the first few partitions before giving up. This patch skips over entries that do not specify partitions such as void entries and drivers. While documentation doesn't say specifically that these partitions should be skipped, it can be inferred from BootX behavior as well as how the documentation doesn't refer to drivers as partitions.

1 0

Re: [OpenBIOS] [PATCH 3/10] Don't stop at APPLE_VOID partition
by William Hahne 22 Aug '11

22 Aug '11

Updated commit message: This patch modifies mac-parts to continue reading partitions even when encountering an APPLE_VOID entry. An APPLE_VOID entry is just place holder allowing another partition to be inserted without rewriting the whole partition table (see Inside Macintosh books, Volume 2 or 4). So it can be safely ignored, meaning there is no reason to stop at an APPLE_VOID entry.

1 0

Re: [OpenBIOS] [PATCH 6/10] Wrapped HFS+ incorrectly recognized
by William Hahne 22 Aug '11

22 Aug '11

Updated commit message: The code for checking for a wrapped HFS+ volume looks at the wrong offset. The offset of the magic number for the wrapped volume is 0x7c, but vol is declared as a uint16 meaning the correct array index to check is 0x3e (see http://developer.apple.com/library/mac/#technotes/tn/tn1150.html)

1 0

[commit] r1047 - trunk/openbios-devel/drivers
by repository service 20 Aug '11

20 Aug '11

Author: blueswirl Date: Sat Aug 20 14:10:07 2011 New Revision: 1047 URL: http://tracker.coreboot.org/trac/openbios/changeset/1047 Log: ESP: clear Unit Attention condition after reset In recent QEMU, SCSI devices are in Unit Attention condition after reset and this made Read Capacity command fail. Fix by sending a Test Unit Ready command to the devices first. Also avoid division by zero if block size is 0 when Read Capacity fails. Thanks to Paolo Bonzini for analysis. Signed-off-by: Blue Swirl <blauwirbel(a)gmail.com> Modified: trunk/openbios-devel/drivers/esp.c Modified: trunk/openbios-devel/drivers/esp.c ============================================================================== --- trunk/openbios-devel/drivers/esp.c Mon Aug 8 14:33:10 2011 (r1046) +++ trunk/openbios-devel/drivers/esp.c Sat Aug 20 14:10:07 2011 (r1047) @@ -108,11 +108,17 @@ DPRINTF("do_command: id %d, cmd[0] 0x%x, status 0x%x\n", sd->id, esp->buffer[0], status); - // Target didn't want all command data or went to status phase - // instead of data phase? - if ((status & ESP_STAT_TCNT) != ESP_STAT_TCNT - || (status & ESP_STAT_PMASK) == ESP_STATP) + /* Target didn't want all command data? */ + if ((status & ESP_STAT_TCNT) != ESP_STAT_TCNT) { return status; + } + if (replylen == 0) { + return 0; + } + /* Target went to status phase instead of data phase? */ + if ((status & ESP_STAT_PMASK) == ESP_STATP) { + return status; + } // Get reply // Set DMA address @@ -176,12 +182,31 @@ if (do_command(esp, sd, 11, 8)) { sd->sectors = 0; sd->bs = 0; - + DPRINTF("read_capacity id %d failed\n", sd->id); return 0; } sd->bs = (esp->buffer[4] << 24) | (esp->buffer[5] << 16) | (esp->buffer[6] << 8) | esp->buffer[7]; sd->sectors = ((esp->buffer[0] << 24) | (esp->buffer[1] << 16) | (esp->buffer[2] << 8) | esp->buffer[3]) * (sd->bs / 512); + DPRINTF("read_capacity id %d bs %d sectors %d\n", sd->id, sd->bs, + sd->sectors); + return 1; +} + +static unsigned int +test_unit_ready(esp_private_t *esp, sd_private_t *sd) +{ + /* Setup command = Test Unit Ready */ + memset(esp->buffer, 0, 6); + esp->buffer[0] = 0x80; + esp->buffer[1] = TEST_UNIT_READY; + + if (do_command(esp, sd, 6, 0)) { + DPRINTF("test_unit_ready id %d failed\n", sd->id); + return 0; + } + + DPRINTF("test_unit_ready id %d success\n", sd->id); return 1; } @@ -234,6 +259,10 @@ DPRINTF("ob_sd_read_blocks id %d %lx block=%d n=%d\n", (*sd)->id, (unsigned long)dest, blk, n ); + if ((*sd)->bs == 0) { + PUSH(0); + return; + } spb = (*sd)->bs / 512; while (n) { sect_offset = blk / spb; @@ -438,6 +467,7 @@ int id, diskcount = 0, cdcount = 0, *counter_ptr; char nodebuff[256], aliasbuff[256]; esp_private_t *esp; + unsigned int i; DPRINTF("Initializing SCSI..."); @@ -476,8 +506,21 @@ for (id = 0; id < 8; id++) { esp->sd[id].id = id; - if (!inquiry(esp, &esp->sd[id])) + if (!inquiry(esp, &esp->sd[id])) { + DPRINTF("Unit %d not present\n", id); continue; + } + /* Clear Unit Attention condition from reset */ + for (i = 0; i < 5; i++) { + if (test_unit_ready(esp, &esp->sd[id])) { + break; + } + } + if (i == 5) { + DPRINTF("Unit %d present but won't become ready\n", id); + continue; + } + DPRINTF("Unit %d present\n", id); read_capacity(esp, &esp->sd[id]); #ifdef CONFIG_DEBUG_ESP

1 0

Re: [OpenBIOS] [PATCH 5/10] Add draw-rectangle to the display package
by Mark Cave-Ayland 19 Aug '11

19 Aug '11

On 16/08/11 02:25, William Hahne wrote: > On Mon, Aug 15, 2011 at 6:57 AM, Mark Cave-Ayland > <mark.cave-ayland(a)siriusit.co.uk> wrote: >> Ah I see - so the real bug is that OpenBIOS can't add the word correctly? Do >> you have any DEBUG_CIF traces so we can determine the Forth being sent to >> OpenBIOS and come up with a fix for that? > > I would recommend looking at the BootX source directly for this. If > you look at http://opensource.apple.com/source/BootX/BootX-45/bootx.tproj/sl.subproj/di… > line 403. You can see that it just throws the phandle directly into > active-package. The device package implementation in OpenBIOS expects > active-package to only be set by calling the active-package! word. If > it is set directly then it doesn't actually change the active-package. So looking at the OpenBIOS code I can see that active-package! not only changes the active-package value, but also switches in the private wordlist for that package. I wonder how OBP does it? I can only assume that there is a hook in the dictionary search routine which interrogates the private wordlist first based upon the current value of active-package. I have a feeling in the back of my mind that I've seen Fcode elsewhere that does a similar thing in some of my SPARC investigations, so it seems like a strong assumption held by various developers. Perhaps it may be possible to alter the $find word to search the active-package wordlist first if set to a non-zero value? > I would like to note that this particular patch is optional for > booting Mac OS X. It is just nice since it allows the Apple logo to be > displayed. Understood. ATB, Mark. -- Mark Cave-Ayland - Senior Technical Architect PostgreSQL - PostGIS Sirius Corporation plc - control through freedom http://www.siriusit.co.uk t: +44 870 608 0063 Sirius Labs: http://www.siriusit.co.uk/labs

2 1

Re: [OpenBIOS] [PATCH 10/10] Adding filll Forth primitive
by Segher Boessenkool 19 Aug '11

19 Aug '11

> +static void filll(void) > +{ > + ucell value = POP(); > + ucell count = POP(); > + ucell *dest = (ucell *)cell2pointer(POP()); > + > + int i; > + for (i = 0; i <= count / 4; i++) { > + dest[i] = value; > + } > +} Apple's code does: int *dest ... ; for (i = count; i > 0; i -= 4) *dest++ = value; which is not the same thing for huge count, or for count a multiple of four (like all callers I have seen use; you write a word too many). Segher

2 2

Re: [OpenBIOS] [PATCH 10/10] Adding filll Forth primitive
by Andreas Färber 17 Aug '11

17 Aug '11

Am 16.08.2011 um 04:04 schrieb William Hahne: > On 8/15/11, Andreas Färber <andreas.faerber(a)web.de> wrote: >> Am 09.08.2011 um 23:55 schrieb William Hahne: >> >>> This is a forth primitive that is required by BootX. It just fills >>> some specified memory address and length with longs. >>> >>> >>> Index: kernel/forth.c >>> =================================================================== >>> --- kernel/forth.c (revision 1041) >>> +++ kernel/forth.c (working copy) >>> @@ -1610,6 +1616,20 @@ >>> memset(src, value, count); >>> } >>> >>> +/* >>> + * filll ( addr len byte -- ) >>> + */ >>> +static void filll(void) >>> +{ >>> + ucell value = POP(); >>> + ucell count = POP(); >>> + ucell *dest = (ucell *)cell2pointer(POP()); >>> + >>> + int i; >>> + for (i = 0; i <= count / 4; i++) { >>> + dest[i] = value; >>> + } >>> +} >> >> This word puzzles me: The extra l in filll seems to be for "long" >> according to the patch description. But there's a hardcoded >> arithmetic >> with 4 and the parameter is documented as byte yet is being written >> ucell-wide. That strikes me as inconsistent. > > Honestly, my problem is I was unable to find documentation on filll. > This suggests that it may be an non-standard addition by Apple. You > are right about the inconsistency. I will make a revised version of > the patch to address this. > >> Either 4 needs to be replaced with sizeof(ucell) to fit sparc64, or >> the division by 4 dropped and uint8_t* written len times. > > It seems likely that it is supposed to fill with 32bit integers, but > this would have to be checked on a 64bit platform which I do not have > access to. I have access to both Apple and IBM ppc64 machines and could check if you provide paste'able Forth code. :) > This may not be applicable to sparc at all but again > someone with access to the hardware would have to confirm. In its present form, the patch seems in no way limited to ppc - you might want to look at the [IFDEF word or so if that's your intent. Also note, your division by 4 leaves up to three trailing bytes un- filled for unaligned lengths. If intentional, you should add a comment saying so. Further, it was suggested we use QEMU coding style for new contributions - the above function uses an inconsistent indentation. Andreas

2 1

Re: [OpenBIOS] [PATCH 10/10] Adding filll Forth primitive
by Segher Boessenkool 17 Aug '11

17 Aug '11

>> This word puzzles me: The extra l in filll seems to be for "long" >> according to the patch description. But there's a hardcoded >> arithmetic >> with 4 and the parameter is documented as byte yet is being written >> ucell-wide. That strikes me as inconsistent. "l" in OF refers to 32-bit memory, like in "la1+" and "rl@" for example. 64-bit is "x". > Honestly, my problem is I was unable to find documentation on filll. > This suggests that it may be an non-standard addition by Apple. It is. Segher

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →