Re: PowerPC port for the e500 core
Hi, On Tue, Feb 24, 2009 at 03:16:02PM +0100, Sebastian Andrzej Siewior wrote: Hello PowerPC team, I have a Lenny snapshot of a new Debian architecture called gnuspe located at [0]. I haven't announced it anywhere (until now). It is called gnuspe because the gcc triplet is powerpc-linux-gnuspe-. This port is for all PowerPCs which have an E500 core [1]. The main important differences between PowerPC and this one are: Are you building userspace with SPE instructions enabled, or are you using only the base power ISA for the binaries? - the normal FPU unit is not present. Instead the APU offers float point operations. This is _not_ SW-emulation it is HW support but different at the assembly level. The ABI is also different (there are no dedicated FPU regs). - those opcodes (or some of them) have the same binary representation like AltiVec. The spec forbids AltiVec. No BookE cores have altivec, besides I don't think the base distro uses any altivec since it would not be usable on IBM big iron before power6, nor on 603/604 (G3) systems. - It is not 100% compatible with the PowerPC ISA. An opcode, lwsync, is not supported by the core and raises an invalid opcode exception. This is a core bug. This is not a core bug per se, but somewhat annoying interpretation of the spec by FSL, as far as I know? In the past I used the normal PowerPC port which was sufficient for most things, however it got very slowly once it came to floating point. Starting with gcc 4.2 (I think but starting with Lenny is correct) the gcc starts using the lwsync opcode more frequently and every C++ program is effected via libstdc++ as you can see in #495120. Replacing the opcode in the PowerPc port would cost performance on all other PowerPCs. Another alternative is to have the kernel rewrite the instruction to be a sync instead (which is the expected behaviour on cores not implementing lwsync), when the fault is taken. Now, I wanted to ask if there are more people that could be interrested in this port. Any comments are welcome. I don't have a personal stake in this, but I think a regular booke port (compiled for softfpu and no SPE) would be more valuable, since that would also run on IBM/AMCC 4xx CPUs. -Olof -- To UNSUBSCRIBE, email to debian-powerpc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Can I open the macro CONFIG_8xx in function do_page_fault() in file kernel\linux-2.6.14-cgl\arch\ppc\mm\fault.c
On Jun 18, 2008, at 11:36 PM, Hema wrote: Hi All, For function do_page_fault(), my understanding is that if a push stack actions occurs for a protected area (such as thread stack guard area protected by mprotect), it will come here. But it seems SIGSEGV can only be sent when CONFIG_8xx is defined. My CPU is 8541 and I want to open this macro, but I also found there are sveral parts in the kernel related to this macro. And I'm not sure about it's impact. Could any one give some proposals? You reference kernel version 2.6.14-cgl. That's not one that debian ships, so you should take your comments to whatever vendor provided you with it. It might be changed in ways we have no idea about. That being said, note that 8xx and 85xx is completely different processors, and the MMU on them works differently. 85xx defines CONFIG_E500. The CONFIG_.* defines are generated by the kernel configuration subsystem. But please, take these questions to the appropriate forum for the kernel you are working with, not the debian-powerpc list. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bug#469030: debian ppc64 not booting after install
On Fri, Mar 07, 2008 at 03:49:21PM +0200, Ulrich Enslin wrote: Hi Frans Since you are going to update the code, can you also check that you are catering for a IBM RS6000 (7026-H70) and IBM pseries 630 (7028-6C4). IBM RS6000 (7026-H70): It would be a better catch-all to check for RS64.* -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [Cbe-oss-dev] Debian for PS3?
On Fri, Feb 22, 2008 at 10:18:52AM -0800, Geoff Levand wrote: Geert Uytterhoeven wrote: Hi, Although it's been possible to install Debian on PS3 since a long time (cfr. http://www.keshi.org/moin/PS3/Debian/Live, or you could just use plain debootstrap), there's no official support for PS3 in Debian yet. So I downloaded debian-testing-powerpc-netinst.iso (dated 2008-02-20), booted it using a self-compiled kernel and /install/powerpc/initrd.gz from the ISO image, and installed Debian on the PS3's hard disk without any glitches! I noticed Olof Johansson is recently working on getting support for his PA Semi boards into Debian, so it might be worthwhile to look at his efforts: For the kernel: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=463200 For the installer: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=464429 http://lists.debian.org/debian-boot/2008/02/msg00153.html As already mentioned on the list, there are already equivalent PS3 bugs open with patches attached. FWIW, mine haven't been picked up yet either but I'm still waiting patiently (with an occasional ping). I can't force code onto anyone, and pushing too hard normally just causes frustration for everyone. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compile quirk linux-2.6.24 (with workaround)
On Tue, Feb 05, 2008 at 09:38:20AM -0600, Josh Boyer wrote: Taking a step back though, there will always be odd cases like this as we move forward. Toolchain XXX will eventually not support instruction which will eventually be used, etc. I'll try to make this specific case work because it's scope is quite limited. But this problem as a whole will still remain. Worst case, if it's only a couple of instructions, they could be covered the way we usually handle new instructions in the ISA: make a macro for them and hand-encode them. See real_205_{read,write}b() in arch/powerpc/kernel/misc_64.S for example. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Powermac G5 won't boot headless?
On Thu, 12 Oct 2006 08:10:03 +1000 Benjamin Herrenschmidt [EMAIL PROTECTED] wrote: On Wed, 2006-10-11 at 10:59 -0400, Alexandre Gauthier wrote: But it never boots completely. Add a monitor, and it's fine... interesting... I have a dual with a serial port, I'll do some tests. It could be a yaboot issue. The quad is pickier than the (old) dual for me. I have a monitor and keyboard plugged in on the quad for this reason, the dual (PM7,3) never needed it. Alexandre, is your dual an old or new model? -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: The powerpc port should be removed from etch release candidates ...
On Thu, May 04, 2006 at 12:12:56PM +0200, Gabriel Paubert wrote: On Thu, May 04, 2006 at 08:41:34AM +0200, Sven Luther wrote: On Thu, May 04, 2006 at 04:38:07PM +1000, Benjamin Herrenschmidt wrote: Hey, cool. so ARCH=ppc will work both for apus and prep, and the rest should go with ARCH=powerpc. This is the case both for 2.6.16 and the upcoming 2.6.17, right ? I don't remember when he fixed it precisely but I think 2.6.16 got it yes. Do you know if there are SMP PReP machines around ? I think i will do only a UP -prep flavour. There were at least dual 604 PreP boards. The IBM 7043-240 is a dual-cpu system, I think it's PReP? -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 7447A strange problem with MSR:POW (WAS: can't boot 2.6.17-rc1)
On Tue, Apr 18, 2006 at 04:32:46PM +1000, Paul Mackerras wrote: understand now why you were talking about putting the code in the exit path on irc ... I don't like it that way Also, if you want to keep it, maybe use a separate CONFIG_PPC_970STYLE_NAP or something that gets selected by platforms that can do it ? The config option makes sense. How about a CPU feature bit instead? That way it's possible to build a multiplatform kernel without getting the overhead on other platforms. 4 nops should be manageable? -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 7447A strange problem with MSR:POW (WAS: can't boot 2.6.17-rc1)
On Tue, Apr 18, 2006 at 09:56:00AM -0500, Olof Johansson wrote: On Tue, Apr 18, 2006 at 04:32:46PM +1000, Paul Mackerras wrote: understand now why you were talking about putting the code in the exit path on irc ... I don't like it that way Also, if you want to keep it, maybe use a separate CONFIG_PPC_970STYLE_NAP or something that gets selected by platforms that can do it ? The config option makes sense. How about a CPU feature bit instead? That way it's possible to build a multiplatform kernel without getting the overhead on other platforms. 4 nops should be manageable? DOH! I'm going blind, that's already there. So, Ben, you're unhappy with nopping it out? -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 7447A strange problem with MSR:POW (WAS: can't boot 2.6.17-rc1)
Hi, On Fri, Apr 14, 2006 at 12:07:23PM -0700, Paul Mackerras wrote: Becky Bruce writes: Actually, I think the problem is that the code linux is using to turn on nap mode is not guaranteed to put the processor in nap mode by the time the blr in ppc6xx_idle occurs. Thanks, Becky. This patch fixes it for me. Comments, anyone? The bf mnemonics had me scratching my head a while, it's not listed as a simplified mnemonic in the 64-bit PEM. Two questions below. _GLOBAL(power_save_6xx_restore) + tophys(r11, r1) /* Make the idle task do a blr */ + lwz r9,_LINK(r11) + stw r9,_NIP(r11) mfspr r11,SPRN_HID0 - rlwinm. r11,r11,0,10,8 /* Clear NAP copy NAP bit !state to cr1 EQ */ - cror4*cr1+eq,4*cr0+eq,4*cr0+eq + rlwinm r11,r11,0,10,8 /* Clear NAP */ BEGIN_FTR_SECTION rlwinm r11,r11,0,9,7 /* Clear DOZE */ END_FTR_SECTION_IFSET(CPU_FTR_CAN_DOZE) mtspr SPRN_HID0, r11 #ifdef DEBUG - beq cr1,1f + bf 9,1f Where is cr0 set now -- you took the dot off of rlwinm? lis r11,(nap_return_count-KERNELBASE)@ha lwz r9,[EMAIL PROTECTED](r11) addir9,r9,1 stw r9,[EMAIL PROTECTED](r11) 1: #endif - + +#ifdef CONFIG_SMP rlwinm r9,r1,0,0,18 tophys(r9,r9) lwz r11,TI_CPU(r9) slwir11,r11,2 +#else + li r11,0 +#endif /* Todo make sure all these are in the same page - * and load r22 (@ha part + CPU offset) only once + * and load r11 (@ha part + CPU offset) only once */ BEGIN_FTR_SECTION - beq cr1,1f + bf 9,1f Same comment as above w.r.t. cr0? -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 7447A strange problem with MSR:POW (WAS: can't boot 2.6.17-rc1)
On Fri, Apr 14, 2006 at 01:19:36PM -0700, Paul Mackerras wrote: Olof Johansson writes: Where is cr0 set now -- you took the dot off of rlwinm? transfer_to_handler does mfspr r11,SPRN_HID0; mtcr r11 before jumping to power_save_6xx_restore. The rlwinm. was wrong anyway since it was setting cr0.eq based on all the *other* bits in HID0, not HID0_NAP (doh!). Oh, right, you even updated the comment to reflect this. My bad. Patch looks good to me then. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 7447A strange problem with MSR:POW (WAS: can't boot 2.6.17-rc1)
On Fri, Apr 14, 2006 at 08:37:21AM +1000, Benjamin Herrenschmidt wrote: I need to verify what's up with the 970. I noticed Apple has some 970 keeps executing too. I guess it's never bitten hard enough to trigger anything serious. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: G4-cube console switching problem
On Thu, Apr 13, 2006 at 09:17:04PM -0400, Rich Johnson wrote: How do I switch text-consoles with an Apple Pro Keyboard. Neither cmd-ctl-F(n) nor opt-ctl-F(n) do anything. I can switch _from_ the gdm using cmd-ctl-F(n)...but I can't get back! What am I missing? Did you try Alt-Fx? Also known as Opt-Fx. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: On the status of support for the current generation iBooks
On Wed, Mar 22, 2006 at 05:20:06PM -0500, James Montgomery wrote: now. My only concern is; will PPC development be droping off? I know I doubt that neither software nor hardware development is dropping off. Main blank spot right now is the lack of consumer/developer machines in a decent price range. it won't disappear overnight, but how do the next 2-5 years look? And if they can get XP to boot on the x86 MacBook, how long before linux x86 will be running on Apple hardware? They already have booted linux on the new x86 hardware. It's just not covered here since it's not PPC-related. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Thermal control problem with G5 tower (PowerMac 9.1) with Kubuntu Breezy PPC install.
On Tue, Feb 14, 2006 at 05:53:03AM -0600, Peter Seebach wrote: In message [EMAIL PROTECTED], Brian Durant w rites: On 14. feb 2006, at 12.42, Peter Seebach wrote: In message [EMAIL PROTECTED], Brian Durant w rites: No, I have a G5 tower 1.8 GHz. single. Is that a 9.1? I thought that was the 7.2. -s It is a PowerMac 9.1, but apparently there is someone with a 7.2 that is experiencing the same problem. Huh, weird. My dual 2Ghz is a 7.3, and it's from the last model line before the 970MP systems. PowerMac7,2 were the initial uniprocessor PPC970 models (1.6 and 1.8GHz). PowerMac7,3 have been most models since then, everything from initial dual 2.0GHz PPC970 all the way to the dual 2.7GHz PPC970FX. PowerMac9,1 is a low-cost uniprocessor model that's based on the iMac design, and uses the U3L northbridge that the iMacs have, with a 1:3 bus speed ratio. PowerMac11,2 are all the new PPC970MP models, i.e. U4 northbridge with one or two chips (2 or 4 cores) per system. I _think_, but I'm not sure, that all 7,3 models use the U3H northbridge, and 7,2 has the original U3. Numbering seems to indicate that Apple bumps version numbers when they upgrade chipsets. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ppc64 missing from upstream 2.6.15-rc6?
On Mon, Dec 19, 2005 at 03:07:05PM +0200, Eduardo TrĂ¡pani wrote: I downloaded the sources from kernel.org, but there is no arch/ppc64 ... I bet I'm missing something obvious, but I cannot find it. Yep, ppc64 has folded into the new shared 32/64-bit powerpc arch (arch/powerpc), and is now controlled by config options therein. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Can't Install on new PowerBooks - NO IDE
On Sat, Nov 26, 2005 at 06:35:19PM +, Chris Burdess wrote: I can confirm all the above. Regarding suspend, it seems to work (with pbbuttonsd) but oopses on wakeup. There are a bunch of USB bugs that causes problems at wakeup. Ben has posted some of the patches for it to this list, check the archives. PCMCIA also appears to have issues. I spent several days trying to get a wireless ethernet card up and running in the PCMCIA slot (works without problems on 2 i386 Debian systems), but although it shows up in lspci there is no trace of it in sysfs. NB this may simply be down to the recent kernel changes (cardmgr-pcmciautils). I'm using a D-Link pccard wireless card on mine, works fine with madwifi drivers: 0001:11:00.0 Ethernet controller: Atheros Communications, Inc. AR5212 802.11abg NIC (rev 01) -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: yaboot command-T
On Thu, Nov 17, 2005 at 08:16:42PM +0100, Fritz Wettstein wrote: Hi, Is there a solution to boot with command-t - mac acts as a firewire-disk server - with yaboot? not sure how you would do it from within yaboot, but you can always do: yaboot: bye ... to get you back to OF, then: 0 target-mode then ctrl-z when you want to break out of it. -Olof -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]