Re: iBook G3 owners
> An earlier email from you on this subject mentioned three instances of "lis > r4,0xfff0" to change. Should I be changing all three as per that earlier > email, or just the one as you noted here? it seems that changing the L1 one is enough. This is what I commited to 2.6.12-rc and should be in -rc4 when released. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Saturday 16 April 2005 07:12 pm, Benjamin Herrenschmidt wrote: > Hrm... can you try to edit > arch/ppc/platforms/pmac_cache.S, and spot that bit of code: > > /* disp-flush L1 */ > li r4,0x4000 > mtctr r4 > lis r4,0xfff0 > 1: lwz r0,0(r4) > addir4,r4,32 > bdnz1b > sync > isync > > And change the line > > lis r4,0xfff0 > > to > > li r4,0 > > and tell me if that helps... Hi Ben, An earlier email from you on this subject mentioned three instances of "lis r4,0xfff0" to change. Should I be changing all three as per that earlier email, or just the one as you noted here? Thanks! Steve -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sun, 17 Apr 2005 00:27:08 -0700 Steve Freitas <[EMAIL PROTECTED]> wrote: > By the way, if I run into USB trouble, are these patches > (http://colino.net/ibookg4/readme.txt) worth applying for the USB > stuff for a G3? Btw, these patches are now at ftp://ftp.kernel.org/pub/linux/kernel/people/colin/patches/ ; they're also in -mm. -- Colin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sun, 2005-04-17 at 12:35 -0700, Steve Freitas wrote: > On Sunday 17 April 2005 08:40 am, Michel Dänzer wrote: > > On Sun, 2005-17-04 at 01:55 -0700, Steve Freitas wrote: > > > Somehow Xorg decided to start looking for its config file in /root, [...] > > > > This has been the case since XFree86 4.0 at least. man xorg.conf > > Since I've just run distros that have put it in /etc/X11, I didn't expect my > dpkg-reconfigure to do that, but it's fine. X.org still looks for it's config file in /etc/X11, but a config file in the home dir of the user (root in your case as you run it as root) will override that. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sunday 17 April 2005 08:40 am, Michel DÃnzer wrote: > On Sun, 2005-17-04 at 01:55 -0700, Steve Freitas wrote: > > Somehow Xorg decided to start looking for its config file in /root, [...] > > This has been the case since XFree86 4.0 at least. man xorg.conf Since I've just run distros that have put it in /etc/X11, I didn't expect my dpkg-reconfigure to do that, but it's fine. Steve
Re: iBook G3 owners
On Sun, 2005-17-04 at 01:55 -0700, Steve Freitas wrote: > > Somehow Xorg decided to start looking for its config file in /root, [...] This has been the case since XFree86 4.0 at least. man xorg.conf -- Earthling Michel DÃnzer | Debian (powerpc), X and DRI developer Libre software enthusiast| http://svcs.affero.net/rm.php?r=daenzer
Re: iBook G3 owners
speaking of xorg, is it in sarge or sid for ppc yet im still runnung xfree86 - but im assuming one magic day ill have an enormous update as xorg takes over and everything is compiled against it - and consequently repackaged Dean Steve Freitas wrote: On Sunday 17 April 2005 12:14 am, Benjamin Herrenschmidt wrote: The X server behaviour has nothing to do with the kernel here. Remind me never to challenge your advice again. ;-) Somehow Xorg decided to start looking for its config file in /root, and that's why it wasn't getting the picture. I'll start banging on the sleep stuff Monday, but until then, thanks! Steve -- WWW: http://dean.bong.com.au LAN: http://www.bong.com.au EMAIL: [EMAIL PROTECTED] or [EMAIL PROTECTED] ICQ: 16867613 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sunday 17 April 2005 12:14 am, Benjamin Herrenschmidt wrote: > The X server behaviour has nothing to do with the kernel here. Remind me never to challenge your advice again. ;-) Somehow Xorg decided to start looking for its config file in /root, and that's why it wasn't getting the picture. I'll start banging on the sleep stuff Monday, but until then, thanks! Steve -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sunday 17 April 2005 12:14 am, Benjamin Herrenschmidt wrote: > There are still issues with USB, I would be careful if I were you, that > is unplug stuffs before suspend. By the way, if I run into USB trouble, are these patches (http://colino.net/ibookg4/readme.txt) worth applying for the USB stuff for a G3? Thanks, Steve -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sun, 2005-04-17 at 00:27 -0700, Steve Freitas wrote: > On Sunday 17 April 2005 12:14 am, Benjamin Herrenschmidt wrote: > > There are still issues with USB, I would be careful if I were you, that > > is unplug stuffs before suspend. > > By the way, if I run into USB trouble, are these patches > (http://colino.net/ibookg4/readme.txt) worth applying for the USB stuff for a > G3? Test and tell us :) Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sunday 17 April 2005 12:14 am, Benjamin Herrenschmidt wrote: > > You're a genius! I gave that a go (actually I made that edit in 3 > > locations in that file, per your earlier email on the topic -- was that > > okay, or should I do it with just the one line?) and it seemed to help. > > There isn't that much genius in there, and I still have no explanation > why that works in fact. There is something really strange happening with > those machines. I wish you were in Southern California. I'd volunteer my machine as a guinea pig for you. > There are still issues with USB, I would be careful if I were you, that > is unplug stuffs before suspend. Okay. Vanilla 2.6.7 worked absolutely beautifully for me -- I could do anything I want with USB devices, suspend worked great, etc. It only had trouble with DRI, and of course no HAL, which is what prompted this painful process. :-) > > Right now I'm tracking down why the X server won't go past 640x480. It's > > a problem also with unpatched 2.6.12-rc2, so now I'm compiling vanilla > > 2.6.11 to see if the problem manifests itself there. The Xorg log reads: > > The X server behaviour has nothing to do with the kernel here. > The problem is that the dislay in this machine doesn't do DDC and the X > server is too dumb to try to obtain the panel size without it. Just do > what you did, that is put some wide enough refresh ranges and tell it to > do 1024x768x60. Yeah, I'd think so, except that 1024x768 worked under Ubuntu's 2.6.10, and my vanilla 2.6.7, and the kernel's the only change I've been making on the machine. Steve -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> > So I rebooted, turned HAL off, and slept it. I was delighted to see that it > > woke back up. But when I tried to do anything, it seemed unable to access > > the disk. > > Hrm... can you try to edit > arch/ppc/platforms/pmac_cache.S, and spot that bit of code: > > /* disp-flush L1 */ > li r4,0x4000 > mtctr r4 > lis r4,0xfff0 > 1: lwz r0,0(r4) > addir4,r4,32 > bdnz1b > sync > isync > > And change the line > > lis r4,0xfff0 > > to > > li r4,0 > > and tell me if that helps... Hi Ben, You're a genius! I gave that a go (actually I made that edit in 3 locations in that file, per your earlier email on the topic -- was that okay, or should I do it with just the one line?) and it seemed to help. Sleep seems to work now, including with the X server (DRI on or off), although I haven't yet subjected it to more extensive testing (with HAL, playing with USB and Firewire devices, doing post-wake tests to see if I can reproduce those funny disk errors, etc). There are several second long delays when the X server is involved, with lots of onscreen psychadelics. I'll remove the delays in the radeon driver per your earlier post and see if that helps. Right now I'm tracking down why the X server won't go past 640x480. It's a problem also with unpatched 2.6.12-rc2, so now I'm compiling vanilla 2.6.11 to see if the problem manifests itself there. The Xorg log reads: (II) RADEON(0): I2C bus "DDC" initialized. (WW) RADEON(0): Video BIOS not detected in PCI space! (WW) RADEON(0): Attempting to read Video BIOS from legacy ISA space! (WW) RADEON(0): Unrecognized BIOS signature, BIOS data will not be used (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): DDC Type: 2, Detected Type: 0 (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. (II) RADEON(0): I2C device "DDC:ddc2" removed. (II) RADEON(0): DDC Type: 3, Detected Type: 0 (II) RADEON(0): (II) RADEON(0): Primary: Monitor -- NONE Connector -- DVI-D DAC Type -- TVDAC/ExtDAC TMDS Type -- Internal DDC Type -- DVI_DDC (II) RADEON(0): Secondary: Monitor -- NONE Connector -- VGA DAC Type -- Primary TMDS Type -- External DDC Type -- VGA_DDC (WW) RADEON(0): Video BIOS not detected, using default clock settings! (WW) RADEON(0): Failed to probe xtal value ! Using default 27Mhz (II) RADEON(0): Probed PLL values: xtal: 27.00 Mhz, sclk: 83.076920 Mhz, mclk: 83.076920 Mhz (EE) RADEON(0): MergedFB does not work with Option UseFBDev, MergedFB mode is disabled (==) RADEON(0): Using gamma correction (1.0, 1.0, 1.0) (II) RADEON(0): Validating modes on Primary head - (II) RADEON(0): Color LCD: Using default hsync range of 28.00-33.00 kHz (II) RADEON(0): Color LCD: Using default vrefresh range of 43.00-72.00 Hz (II) RADEON(0): Clock range: 12.50 to 350.00 MHz Now, under Ubuntu's 2.6.10, it also defaulted to 640x480, but by adding... HorizSync 31.5-48.5 VertRefresh 40-70 ...to xorg.conf, I was able to get it to happily do 1024x768. If 2.6.11 also gives me the problem, can you recommend anything to me? Might there be something I'm doing wrong in my kernel's .config? Thanks for all your help Ben, Steve -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> You're a genius! I gave that a go (actually I made that edit in 3 locations > in > that file, per your earlier email on the topic -- was that okay, or should I > do it with just the one line?) and it seemed to help. There isn't that much genius in there, and I still have no explanation why that works in fact. There is something really strange happening with those machines. > Sleep seems to work > now, including with the X server (DRI on or off), although I haven't yet > subjected it to more extensive testing (with HAL, playing with USB and > Firewire devices, doing post-wake tests to see if I can reproduce those funny > disk errors, etc). There are several second long delays when the X server is > involved, with lots of onscreen psychadelics. I'll remove the delays in the > radeon driver per your earlier post and see if that helps. There are still issues with USB, I would be careful if I were you, that is unplug stuffs before suspend. > Right now I'm tracking down why the X server won't go past 640x480. It's a > problem also with unpatched 2.6.12-rc2, so now I'm compiling vanilla 2.6.11 > to see if the problem manifests itself there. The Xorg log reads: The X server behaviour has nothing to do with the kernel here. > (II) RADEON(0): I2C bus "DDC" initialized. > (WW) RADEON(0): Video BIOS not detected in PCI space! > (WW) RADEON(0): Attempting to read Video BIOS from legacy ISA space! > (WW) RADEON(0): Unrecognized BIOS signature, BIOS data will not be used > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): DDC Type: 2, Detected Type: 0 > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): I2C device "DDC:ddc2" registered at address 0xA0. > (II) RADEON(0): I2C device "DDC:ddc2" removed. > (II) RADEON(0): DDC Type: 3, Detected Type: 0 > (II) RADEON(0): > (II) RADEON(0): Primary: > Monitor -- NONE > Connector -- DVI-D > DAC Type -- TVDAC/ExtDAC > TMDS Type -- Internal > DDC Type -- DVI_DDC > (II) RADEON(0): Secondary: > Monitor -- NONE > Connector -- VGA > DAC Type -- Primary > TMDS Type -- External > DDC Type -- VGA_DDC > (WW) RADEON(0): Video BIOS not detected, using default clock settings! > (WW) RADEON(0): Failed to probe xtal value ! Using default 27Mhz > (II) RADEON(0): Probed PLL values: xtal: 27.00 Mhz, sclk: 83.076920 Mhz, > mclk: 83.076920 Mhz > (EE) RADEON(0): MergedFB does not work with Option UseFBDev, MergedFB mode is > disabled > (==) RADEON(0): Using gamma correction (1.0, 1.0, 1.0) > (II) RADEON(0): Validating modes on Primary head - > (II) RADEON(0): Color LCD: Using default hsync range of 28.00-33.00 kHz > (II) RADEON(0): Color LCD: Using default vrefresh range of 43.00-72.00 Hz > (II) RADEON(0): Clock range: 12.50 to 350.00 MHz > > Now, under Ubuntu's 2.6.10, it also defaulted to 640x480, but by adding... > HorizSync 31.5-48.5 > VertRefresh 40-70 > ...to xorg.conf, I was able to get it to happily do 1024x768. If 2.6.11 also > gives me the problem, can you recommend anything to me? Might there be > something I'm doing wrong in my kernel's .config? > > Thanks for all your help Ben, The problem is that the dislay in this machine doesn't do DDC and the X server is too dumb to try to obtain the panel size without it. Just do what you did, that is put some wide enough refresh ranges and tell it to do 1024x768x60. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sat, 2005-04-16 at 14:36 -0700, Steve Freitas wrote: > Okay, I'm using 2.6.12-rc2 with your patches as downloaded today. (Don't know > what's responsible for the Zilog and dm errors below, I'm using Ubuntu > Hoary's config from 2.6.10 brought forward with make oldconfig.) No Firewire > or USB devices attached. I'm booting without X ever running and without the > radeon module loaded. I've pasted my lsmod, dmesg and /proc/cpuinfo below. > > The experience has actually been a little bizarre. The first time I tried > waking it up (with HAL running), it put some stuff about radeon and eth0 on > the screen and then froze. Couldn't SSH in, caps lock key light was working. > I figured I'd try again without HAL and capture the text of this particular > error later. > > So I rebooted, turned HAL off, and slept it. I was delighted to see that it > woke back up. But when I tried to do anything, it seemed unable to access the > disk. Hrm... can you try to edit arch/ppc/platforms/pmac_cache.S, and spot that bit of code: /* disp-flush L1 */ li r4,0x4000 mtctr r4 lis r4,0xfff0 1: lwz r0,0(r4) addir4,r4,32 bdnz1b sync isync And change the line lis r4,0xfff0 to li r4,0 and tell me if that helps... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Okay, I'm using 2.6.12-rc2 with your patches as downloaded today. (Don't know what's responsible for the Zilog and dm errors below, I'm using Ubuntu Hoary's config from 2.6.10 brought forward with make oldconfig.) No Firewire or USB devices attached. I'm booting without X ever running and without the radeon module loaded. I've pasted my lsmod, dmesg and /proc/cpuinfo below. The experience has actually been a little bizarre. The first time I tried waking it up (with HAL running), it put some stuff about radeon and eth0 on the screen and then froze. Couldn't SSH in, caps lock key light was working. I figured I'd try again without HAL and capture the text of this particular error later. So I rebooted, turned HAL off, and slept it. I was delighted to see that it woke back up. But when I tried to do anything, it seemed unable to access the disk. [EMAIL PROTECTED]:~ # ll /etc /bin/ls- Success Trying dmesg, it complained it didn't know about /bin/dmesg. Couldn't SSH in, either. So I rebooted. Now, any time I sleep, it just wakes back up to a black screen, HAL or no HAL. Steve - [EMAIL PROTECTED]:~ # lsmod Module Size Used by cpufreq_userspace 5676 1 cpufreq_powersave 2016 0 cpufreq_ondemand8060 0 ipv6 328040 15 af_packet 21480 2 ide_cd 51492 0 cdrom 50716 1 ide_cd sungem 38852 0 sungem_phy 10368 1 sungem ohci1394 42516 0 ieee1394 443504 1 ohci1394 ohci_hcd 26468 0 usbcore 146164 2 ohci_hcd uninorth_agp 11272 1 agpgart44188 1 uninorth_agp evdev 12160 2 tsdev 9344 0 md 56156 0 dm_mod 73140 0 ext3 153320 1 jbd72600 1 ext3 mbcache11244 1 ext3 ide_disk 21952 3 via82cxxx 14876 0 [permanent] trm290 5762 0 [permanent] sl82c1058032 0 [permanent] siimage14944 0 [permanent] sc1200 10464 0 [permanent] pdc202xx_old 14240 0 [permanent] pdc202xx_new 13600 0 [permanent] ns87415 5092 0 [permanent] hpt366 24480 0 [permanent] hpt34x 6464 0 [permanent] generic 5248 0 [permanent] cmd64x 14780 0 [permanent] aec62xx 9280 0 [permanent] unix 32532 166 [EMAIL PROTECTED]:~ # cat /proc/cpuinfo processor : 0 cpu : 750FX temperature : 11-13 C (uncalibrated) clock : 400MHz revision: 1.18 (pvr 7000 0112) bogomips: 793.45 machine : PowerBook4,3 motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh detected as : 257 (iBook 2 rev. 2) pmac flags : 001b L2 cache: 512K unified memory : 640MB pmac-generation : NewWorld [EMAIL PROTECTED]:~ # dmesg Total memory = 640MB; using 2048kB for hash table (at c040) Linux version 2.6.12-rc2-1 ([EMAIL PROTECTED]) (gcc version 3.3.5 (Debian 1:3.3.5-8ubuntu2)) #2 Sat Apr 16 13:58:50 PDT 2005 Found UniNorth memory controller & host bridge, revision: 192 Mapped at 0xfdf0 Found a Pangea mac-io controller, rev: 0, mapped at 0xfde8 Processor NAP mode on idle enabled. PowerMac motherboard: iBook 2 rev. 2 Found UniNorth PCI host bridge at 0xf000. Firmware bus number: 0->0 Found UniNorth PCI host bridge at 0xf200. Firmware bus number: 0->0 Found UniNorth PCI host bridge at 0xf400. Firmware bus number: 0->0 via-pmu: Server Mode is disabled PMU driver 2 initialized for Core99, firmware: 0c nvram: Checking bank 0... nvram: gen0=693, gen1=692 nvram: Active bank is: 0 nvram: OF partition at 0x410 nvram: XP partition at 0x1020 nvram: NR partition at 0x1120 On node 0 totalpages: 163840 DMA zone: 163840 pages, LIFO batch:16 Normal zone: 0 pages, LIFO batch:1 HighMem zone: 0 pages, LIFO batch:1 Built 1 zonelists Kernel command line: root=/dev/hda4 ro quiet splash PowerMac using OpenPIC irq controller at 0x8004 OpenPIC Version 1.2 (4 CPUs and 64 IRQ sources) at fc62f000 OpenPIC timer frequency is 4.16 MHz PID hash table entries: 4096 (order: 12, 65536 bytes) GMT Delta read from XPRAM: -480 minutes, DST: off time_init: decrementer frequency = 24.835213 MHz Console: colour dummy device 80x25 Dentry cache hash table entries: 131072 (order: 7, 524288 bytes) Inode-cache hash table entries: 65536 (order: 6, 262144 bytes) Memory: 639168k available (1892k kernel code, 1184k data, 176k init, 0k highmem) AGP special page: 0xe7fff000 Calibrating delay loop... 1388.54 BogoMIPS (lpj=694272) Mount-cache hash table entries: 512 checking if image is initramfs...it isn't (bad gzip magic numbers); looks like an initrd Freeing initrd memory: 4376k freed NET: Registered protocol family 1
Re: iBook G3 owners
> if 2.6.11 plain doesn't work, go backward from that. > > The ordering is: > > 2.6.10 > 2.6.10-bk* > 2.6.11-rc1 > 2.6.11-rc1-bk* > 2.6.11-rc2 > 2.6.11-rc2-bk* > etc... > 2.6.11 > 2.6.11.x (1..7) are just small bug/security fixes Hi Ben. Sorry about the delay. After a long work, testing, compiling, configuring and re-compiling, I'm in a point where I don't know what is going on. I'm going to explain you in a summarized way all the work I've done, and after that, the point where I'm now and others comments about this. 2.6.10 plain: all works. ... 2.6.11-rc2: all works. ... 2.6.11-rc2-bk10: all works. 2.6.11-rc3: not tested, compilation error. 2.6.11-rc3-bk1: I can't resume/wakeup, but X works witn DRI enabled. 2.6.11-rc3-bk2: I can't resume/wakeup, but X works witn DRI enabled. 2.6.11-rc3-bk3: not tested. 2.6.11-rc3-bk4: not tested. 2.6.11-rc3-bk5: I can't resume/wakeup, but X works witn DRI enabled. 2.6.11-rc3-bk6: not tested. 2.6.11-rc3-bk7: not tested. 2.6.11-rc3-bk8: not tested. 2.6.11-rc3-bk9: it doesn't work, but X works with DRI enabled. 2.6.11-rc4: it doesn't work resume/wakeup, but it works DRI enabled. ... 2.6.11 plain: it doesn't work resume/wakeup, but it works DRI enabled. First all, now I'm testing 2.6.11-rc3-bk* kernel branch. I've tried with 2.6.11-rc2 plain and with 2.6.11-rc2-bk9, and all works. When I say "all works" I'm refering to resume/wakeup cycle works and start X with DRI enabled also it works. I also tried 2.6.11-rc4 and above and the resume/wakeup crash the ibook and I've to power off the ibook manually. But it curious because X with DRI enabled whether works. I've to say that all test I've done, they have been done without your patches, just like I've downloaded from the upstream (www.kernel.org). Well, to summarize: - 2.6.11-rc2-bk*: all works. - 2.6.11-rc3-bk*: the resume doesn't work. No crash, but it doesn't resume. - 2.6.11-rc4-bk*: crash when I try to wakeup, but X with DRI enabled works. Then, I'm in this point, trying to test 2.6.11-rc3-bk*. When I've compiled this kernel, I've seen the compilation time is less than when I compile the others kernels from branches 2.6.11-rc2-bk* and 2.6.11-rc4-bk*. With "rc3" the total compilation time is about 22 minutes, and with the others ("rc2", "rc4", ...) is about 33 minutes. The size in all the cases is almost equal, about 5,3MB I will like to be able to test resume/wakeup cycle on 2.6.11-rc3-bk* kernel branch, but for some reason, the resume doesn't work when I close the lid of my ibook. I've tried to sleep it manually with: "echo mem > /sys/power/state" but in my ibook that doesn't work (with my old kernel neither it works). Also, I've tried to sleep with a little program of a friend of mine, that basically it does a system call: "ioctl(fd, PMU_IOC_SLEEP, 0)" but it doesn't work. Furthermore, the system shows that message: "Sleep mode not supported on this machine" Due to those troubles and problems, I can't say you exactly when the sleep just doesn't work. I will like to test all the 2.6.11-rc3-bk* kernel branch to figure out where is the problem with resume/wakeup cycle to help to you. I don't know if all this information can helps to you. Tell me about it when you can. Thanks for your time and work. :-) Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On 18:04:10 09/Apr , Kristian Peters wrote: > X doesn't work. I only see some wierd flickering when starting X. > > Somehow sleep-support isn't supported by the kernel. Any ideas? strange, because in my identical ibook all is working really good, both with xfree and xorg. I haven't any problem. You can find my kconfig here: http://shine.homelinux.org/ppc/ibook/ HTH > *Kristian Bye Menoz -- Free Software Enthusiast Debian Powered Linux User #332564 http://shine.homelinux.org -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Benjamin Herrenschmidt <[EMAIL PROTECTED]> schrieb: > Do you have CONFIG_PMAC_PBOOK in your .config ? I don't know what's up > with X, what video chip do you have ? Did X broke with the latest kernel > or was it always broken ? Can you try figuing out what kernel version > broke it ? (and please CC me or I may miss your replies) Sorry for the delay. Maybe it was a wrong config but now X works. Suspend also works as always on both X and pure console. Please tell me how I can help you further. PS: My video chip is Radeon M7 LW [Radeon Mobility 7500]. *Kristian # lspci -vvv :00:0b.0 Host bridge: Apple Computer Inc. UniNorth/Pangea AGP Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- TAbort- SERR- http://www.korseby.net /\\ ::_\_V :.: pgpj9NHsALs4H.pgp Description: PGP signature
Re: iBook G3 owners
On Sat, 2005-04-09 at 18:04 +0200, Kristian Peters wrote: > Hi. > > I own an iBook 2.2 900 MHz and did some testing with your recent patches. > > Thermal management is somewhat better as before (2.6.8). My iBook isn't > getting much hot anymore. btw: I'm using cpudynd to control cpu speed. > > X doesn't work. I only see some wierd flickering when starting X. > > Somehow sleep-support isn't supported by the kernel. Any ideas? > > # pmud > pmud [treshold = 420, margin = 15] started > PMU version 12: iBook/G3 Pismo/G4 Titanium > No sleep support on this hardware, exiting! > daemon stopped (missing sleep support) Do you have CONFIG_PMAC_PBOOK in your .config ? I don't know what's up with X, what video chip do you have ? Did X broke with the latest kernel or was it always broken ? Can you try figuing out what kernel version broke it ? (and please CC me or I may miss your replies) > > $ cat /proc/cpuinfo > processor : 0 > cpu : 750FX > temperature : 52 C (uncalibrated) > clock : 900MHz > revision: 2.3 (pvr 7000 0203) > bogomips: 1785.85 > machine : PowerBook4,3 > motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh > detected as : 257 (iBook 2 rev. 2) > pmac flags : 001a > L2 cache: 512K unified > memory : 640MB > pmac-generation : NewWorld > > [EMAIL PROTECTED] /sys/devices/system/cpu/cpu0/cpufreq]# for i in *; do echo > $i:; cat $i; done > affected_cpus: > 0 > cpuinfo_cur_freq: > 40 > cpuinfo_max_freq: > 90 > cpuinfo_min_freq: > 40 > scaling_available_governors: > ondemand powersave userspace performance > scaling_cur_freq: > 40 > scaling_driver: > powermac > scaling_governor: > powersave > scaling_max_freq: > 90 > scaling_min_freq: > 40 > > Thanks for your help. > > *Kristian > -- Benjamin Herrenschmidt <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Sat, 2005-04-09 at 17:36 +0100, David Woodhouse wrote: > On Mon, 2005-04-04 at 16:06 +1000, Benjamin Herrenschmidt wrote: > > There have been various reports of issues with sleep among others on > > iBook G3 equiped with the 750FX processor. Also, the cpufreq code on > > these so far didn't change the CPU voltage, which limited the actual > > power saving at low frequency. > > Sleep hasn't worked properly on my PowerBook G4 since I applied your > first pair of patches. The current behaviour seems to be that short > periods of sleep appear to work OK, but a fairly random death happens > after a longer sleep. Last time it was ohci_hcd complaining of an > uninitialised spinlock. > > I'm half inclined to suspect RAM isn't being refreshed properly -- but > most of the kernel text still seems to be there and it manages to oops > and panic OK, so perhaps it's just that the dcache isn't correctly > disabled, or isn't correctly flushed when we enable it again? That kind of sleep is usually related to cache flush issues ... the setup of RAM auto refresh isn't under kernel code control, it's entirely done by the chipset when entering sleep state. What is strange however is that none of these 2 patches have _any_ effect on your machine model. Can you confirm that's indeed the patches you were talking about ? ppc32-750-errata-fix.diff ppc32-pmac-sleep-fix.diff > I'll rebuild the Fedora kernel without any of the patches to eliminate > the possibility that it's caused by some other change. Make sure you have the 2 cpufreq related patches though, there are issues with CPU voltage possibly not correct on wakeup that can cause similar problems. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Mon, 2005-04-04 at 16:06 +1000, Benjamin Herrenschmidt wrote: > There have been various reports of issues with sleep among others on > iBook G3 equiped with the 750FX processor. Also, the cpufreq code on > these so far didn't change the CPU voltage, which limited the actual > power saving at low frequency. Sleep hasn't worked properly on my PowerBook G4 since I applied your first pair of patches. The current behaviour seems to be that short periods of sleep appear to work OK, but a fairly random death happens after a longer sleep. Last time it was ohci_hcd complaining of an uninitialised spinlock. I'm half inclined to suspect RAM isn't being refreshed properly -- but most of the kernel text still seems to be there and it manages to oops and panic OK, so perhaps it's just that the dcache isn't correctly disabled, or isn't correctly flushed when we enable it again? I'll rebuild the Fedora kernel without any of the patches to eliminate the possibility that it's caused by some other change. -- dwmw2 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Hi. I own an iBook 2.2 900 MHz and did some testing with your recent patches. Thermal management is somewhat better as before (2.6.8). My iBook isn't getting much hot anymore. btw: I'm using cpudynd to control cpu speed. X doesn't work. I only see some wierd flickering when starting X. Somehow sleep-support isn't supported by the kernel. Any ideas? # pmud pmud [treshold = 420, margin = 15] started PMU version 12: iBook/G3 Pismo/G4 Titanium No sleep support on this hardware, exiting! daemon stopped (missing sleep support) $ cat /proc/cpuinfo processor : 0 cpu : 750FX temperature : 52 C (uncalibrated) clock : 900MHz revision: 2.3 (pvr 7000 0203) bogomips: 1785.85 machine : PowerBook4,3 motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh detected as : 257 (iBook 2 rev. 2) pmac flags : 001a L2 cache: 512K unified memory : 640MB pmac-generation : NewWorld [EMAIL PROTECTED] /sys/devices/system/cpu/cpu0/cpufreq]# for i in *; do echo $i:; cat $i; done affected_cpus: 0 cpuinfo_cur_freq: 40 cpuinfo_max_freq: 90 cpuinfo_min_freq: 40 scaling_available_governors: ondemand powersave userspace performance scaling_cur_freq: 40 scaling_driver: powermac scaling_governor: powersave scaling_max_freq: 90 scaling_min_freq: 40 Thanks for your help. *Kristian -- :... [snd.science] ...: :: _o) :: http://www.korseby.net /\\ ::_\_V :.: pgps15databOU.pgp Description: PGP signature
Re: iBook G3 owners
Hi again. > [...] > Actually ... just in case, do one test with > radeon.default_dynclk=-1 Well, I've tried it with the "radeon.default_dynclk=-1" option. Just booted, I start X with DRI disabled and it works. But if I enable DRI, it crashes. After that, I've applied this last patch and make the same tests. And in both cases it ocurrs the same: DRI disabled -> works, DRI enabled -> crashes. With the patch applied there is no pause as you said. > [...] > If that doesn't help, then I don't know at this point what can be > causing the crash. All I can suggest is you go back kernel versions and > sub versions until you can isolate more precisely when the crash started > to happen... Ok, don't worry. As soon as possible I will try the others kernels. I don't know if I've to try with 2.6.11.[1-6] or try with 2.6.11-rc1-bk[1-6]. I'm a bit confused with this. By the moment I'll start with 2.6.11 plain, after 2.6.11-rc1, after that 2.6.11-rc1-bk1, and so on. At the moment I see it doesn't work, I will let you know. I tried with 2.6.10 plain and it worked all. Thank you. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Fri, 2005-04-08 at 05:11 +0200, Esteban Martinez wrote: > Hi again. > > > [...] > > Actually ... just in case, do one test with > > radeon.default_dynclk=-1 > Well, I've tried it with the "radeon.default_dynclk=-1" option. Just booted, I > start X with DRI disabled and it works. But if I enable DRI, it crashes. > > After that, I've applied this last patch and make the same tests. And in both > cases it ocurrs the same: DRI disabled -> works, DRI enabled -> crashes. > With the patch applied there is no pause as you said. > > > [...] > > If that doesn't help, then I don't know at this point what can be > > causing the crash. All I can suggest is you go back kernel versions and > > sub versions until you can isolate more precisely when the crash started > > to happen... > > Ok, don't worry. As soon as possible I will try the others kernels. I don't > know if I've to try with 2.6.11.[1-6] or try with 2.6.11-rc1-bk[1-6]. I'm a > bit confused with this. By the moment I'll start with 2.6.11 plain, after > 2.6.11-rc1, after that 2.6.11-rc1-bk1, and so on. At the moment I see it > doesn't work, I will let you know. I tried with 2.6.10 plain and it worked > all. if 2.6.11 plain doesn't work, go backward from that. The ordering is: 2.6.10 2.6.10-bk* 2.6.11-rc1 2.6.11-rc1-bk* 2.6.11-rc2 2.6.11-rc2-bk* etc... 2.6.11 2.6.11.x (1..7) are just small bug/security fixes -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Fri, 2005-04-08 at 03:26 +0200, Esteban Martinez wrote: > > The pause is expected, but shouldn't be _that_ long, it seems some > > problem is triggered by the errata workaround for those chips. If you > > edit drivers/video/radeonfb.h, and comment out that bit of code: > > > > /* we can't deal with posted writes here ... */ > > _radeon_msleep(rinfo, 5); > > > > In the definition of radeon_pll_errata_after_data(), the pause will go > > away. Does it have any effect on your crashes ? > > Ok. With that line commented, the pause disappears at boot time. I've tried to > start X with DRI _enabled_ and it crashes like the other times. Also, I've > tried with DRI _disabled_ and it doesn't crash. The same like the other times. > In both cases I've passed the option "radeonfb.default_dynclk=0" to the > kernel. Ok, so the crash is unrelated to that workaround. Here is a patch that should fix the pause problem completely. You should remove the default_dynclk=0 thing, as its not related to the crash you are having. Actually ... just in case, do one test with radeon.default_dynclk=-1 (That will instruct radeonfb not to touch the dynamic clock settings at all, that is leave it to the firmware setting). And test DRI after boot (not after sleep/wake as this will cause the dynamic clocks to be modified anyway). If that doesn't help, then I don't know at this point what can be causing the crash. All I can suggest is you go back kernel versions and sub versions until you can isolate more precisely when the crash started to happen... The patch below should address the "pause" issue. Index: linux-work/include/linux/fb.h === --- linux-work.orig/include/linux/fb.h 2005-04-01 09:04:19.0 +1000 +++ linux-work/include/linux/fb.h 2005-04-08 10:24:56.0 +1000 @@ -563,6 +563,9 @@ int (*fb_setcolreg)(unsigned regno, unsigned red, unsigned green, unsigned blue, unsigned transp, struct fb_info *info); + /* set color registers in batch */ + int (*fb_setcmap)(struct fb_cmap *cmap, struct fb_info *info); + /* blank display */ int (*fb_blank)(int blank, struct fb_info *info); Index: linux-work/drivers/video/aty/radeon_base.c === --- linux-work.orig/drivers/video/aty/radeon_base.c 2005-04-01 09:04:18.0 +1000 +++ linux-work/drivers/video/aty/radeon_base.c 2005-04-08 11:15:13.0 +1000 @@ -1057,13 +1057,14 @@ return radeon_screen_blank(rinfo, blank, 0); } -static int radeonfb_setcolreg (unsigned regno, unsigned red, unsigned green, - unsigned blue, unsigned transp, struct fb_info *info) +static int radeon_setcolreg (unsigned regno, unsigned red, unsigned green, + unsigned blue, unsigned transp, +struct radeonfb_info *rinfo) { -struct radeonfb_info *rinfo = info->par; u32 pindex; unsigned int i; - + + if (regno > 255) return 1; @@ -1078,20 +1079,7 @@ pindex = regno; if (!rinfo->asleep) { - u32 dac_cntl2, vclk_cntl = 0; - radeon_fifo_wait(9); - if (rinfo->is_mobility) { - vclk_cntl = INPLL(VCLK_ECP_CNTL); - OUTPLL(VCLK_ECP_CNTL, vclk_cntl & ~PIXCLK_DAC_ALWAYS_ONb); - } - - /* Make sure we are on first palette */ - if (rinfo->has_CRTC2) { - dac_cntl2 = INREG(DAC_CNTL2); - dac_cntl2 &= ~DAC2_PALETTE_ACCESS_CNTL; - OUTREG(DAC_CNTL2, dac_cntl2); - } if (rinfo->bpp == 16) { pindex = regno * 8; @@ -1101,24 +1089,27 @@ if (rinfo->depth == 15 && regno > 31) return 1; - /* For 565, the green component is mixed one order below */ + /* For 565, the green component is mixed one order +* below +*/ if (rinfo->depth == 16) { OUTREG(PALETTE_INDEX, pindex>>1); - OUTREG(PALETTE_DATA, (rinfo->palette[regno>>1].red << 16) | - (green << 8) | (rinfo->palette[regno>>1].blue)); + OUTREG(PALETTE_DATA, + (rinfo->palette[regno>>1].red << 16) | + (green << 8) | + (rinfo->palette[regno>>1].blue)); green = rinfo->palette[regno<<1].green; } } if (rinfo->depth != 16 || regno
Re: iBook G3 owners
> The pause is expected, but shouldn't be _that_ long, it seems some > problem is triggered by the errata workaround for those chips. If you > edit drivers/video/radeonfb.h, and comment out that bit of code: > > /* we can't deal with posted writes here ... */ > _radeon_msleep(rinfo, 5); > > In the definition of radeon_pll_errata_after_data(), the pause will go > away. Does it have any effect on your crashes ? Ok. With that line commented, the pause disappears at boot time. I've tried to start X with DRI _enabled_ and it crashes like the other times. Also, I've tried with DRI _disabled_ and it doesn't crash. The same like the other times. In both cases I've passed the option "radeonfb.default_dynclk=0" to the kernel. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> Well, there is something weird going on with the radeon chip, I have no > idea what exactly at this point. I don't remember if I already asked > you, but did you try booting with that line on the kernel command line: > "radeonfb.default_dynclk=0" ? (That should turn the above message into > Dynamic Clock Power Management disabled). Yes, you told me. I tested it. I've tested it again and it just ocurrs like you say, it shows the message: "Dynamic Clock Power Management disabled". However, the pause still going on even if I pass that option on the kernel command line. It's weird because with the kernels that are working right for me (2.6.9 and 2.6.10), the pause doesn't ocurrs. One more time, thanks for your work. :-) Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Fri, 2005-04-08 at 01:56 +0200, Esteban Martinez wrote: > > Well, there is something weird going on with the radeon chip, I have no > > idea what exactly at this point. I don't remember if I already asked > > you, but did you try booting with that line on the kernel command line: > > "radeonfb.default_dynclk=0" ? (That should turn the above message into > > Dynamic Clock Power Management disabled). > > Yes, you told me. I tested it. I've tested it again and it just ocurrs like > you say, it shows the message: "Dynamic Clock Power Management disabled". > However, the pause still going on even if I pass that option on the kernel > command line. It's weird because with the kernels that are working right > for me (2.6.9 and 2.6.10), the pause doesn't ocurrs. > > One more time, thanks for your work. :-) The pause is expected, but shouldn't be _that_ long, it seems some problem is triggered by the errata workaround for those chips. If you edit drivers/video/radeonfb.h, and comment out that bit of code: /* we can't deal with posted writes here ... */ _radeon_msleep(rinfo, 5); In the definition of radeon_pll_errata_after_data(), the pause will go away. Does it have any effect on your crashes ? Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Fri, 2005-04-08 at 01:20 +0200, Andreas Schwab wrote: > Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes: > > > I know, but this oneliner is incorrect. It can't be correct, no way. > > That is very strange, and that's what I'm trying to figure out. Can you > > try instead changing the start address of the flush from 0xfff0 to 0 > > by changing the L1 flush bit initalisation from lis r4,0xfff0 to li > > r4,0 ? > > This fixes the problem. Ok, I still don't have anything like a good explanation of why this fixes it, but it's a better workaround for now. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes: > I know, but this oneliner is incorrect. It can't be correct, no way. > That is very strange, and that's what I'm trying to figure out. Can you > try instead changing the start address of the flush from 0xfff0 to 0 > by changing the L1 flush bit initalisation from lis r4,0xfff0 to li > r4,0 ? This fixes the problem. Andreas. -- Andreas Schwab, SuSE Labs, [EMAIL PROTECTED] SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different." -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Thu, 2005-04-07 at 09:42 +0200, Kristian Peters wrote: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Hello. > > First, where can I get the latest bk snapshot of the kernel ? Or is the > latest -rc2 enough ? rc2 is fine. > Can anyone send me a working .config to test the patches ? Somehow any > of my self-compiled 2.6 kernels won't boot. Don't know the reason... There should be a pmac_defconfig in arch/ppc/configs. Not very up to date but should provide a good basis. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> The same test but with DRI _disabled_: > 1) Boot in console mode. > 2) Start X and goes well. Normal exit, but with green vertical slashes > for a few seconds (like the other times). > > I've seen every time it crashes, at the beginning of boot time it pauses > for 4 or 5 seconds just at the line: > "radefonfb: Dynamic Power Clock Management enabled" > > I wish this information can help you. :-) If you need something else, > tell me. Well, there is something weird going on with the radeon chip, I have no idea what exactly at this point. I don't remember if I already asked you, but did you try booting with that line on the kernel command line: "radeonfb.default_dynclk=0" ? (That should turn the above message into Dynamic Clock Power Management disabled). Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Fri, 2005-04-08 at 00:05 +0200, Andreas Schwab wrote: > Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes: > > > http://gate.crashing.org/~benh/ppc32-pmac-sleep-fix.diff > > This patch does not work on the 750fx, I'm getting random crashes on > wakeup. When I replace it with this one-liner, all is well again: I know, but this oneliner is incorrect. It can't be correct, no way. That is very strange, and that's what I'm trying to figure out. Can you try instead changing the start address of the flush from 0xfff0 to 0 by changing the L1 flush bit initalisation from lis r4,0xfff0 to li r4,0 ? Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Benjamin Herrenschmidt <[EMAIL PROTECTED]> writes: > http://gate.crashing.org/~benh/ppc32-pmac-sleep-fix.diff This patch does not work on the 750fx, I'm getting random crashes on wakeup. When I replace it with this one-liner, all is well again: --- linux-2.6.12-rc2.orig/arch/ppc/platforms/pmac_cache.S 2005-04-04 23:05:41.0 +0200 +++ linux-2.6.12-rc2/arch/ppc/platforms/pmac_cache.S2005-04-07 23:22:04.0 +0200 @@ -73,7 +73,7 @@ /* disable / invalidate / enable L1 data */ mfspr r3,SPRN_HID0 - rlwinm r0,r0,0,~HID0_DCE + rlwinm r3,r3,0,~(HID0_DCE | HID0_ICE) mtspr SPRN_HID0,r3 sync isync Andreas. -- Andreas Schwab, SuSE Labs, [EMAIL PROTECTED] SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different." -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> You haven't tried: > 1) Boot in console mode > 2) Test X > (with no sleep wkaeup cycle and with DRI enabled). Can you try and let > me know ? Ok. Tested it with plain 2.6.12-rc1-bk6 kernel. It crashes (black screen, no keyboard) but the hard disk continues spinning. >> Test with DRI _disabled_: >> 1) Boot in console mode. >> 2) Just after I've logged, I start X and it works. After that, I come >> back to console mode and the screens stays for a few seconds with >> vertical green slashes, and then appears the console. I can contine >> working properly. >> 3) After that I do a sleep/wakeup cycle and it crashes when it tries >> to wake up (black screen, no keyboard), but this time the hard disk >> continues spinning. > > Ok. Can you also try the attached patch ? Test 2.6.12-rc1-bk6 kernel plus this last patch and without the others: 1) Boot in console mode. 2) Start X (DRI enabled). It crashes (blank screen, no keyboard, hdd on) The same test but with DRI _disabled_: 1) Boot in console mode. 2) Start X and goes well. Normal exit, but with green vertical slashes for a few seconds (like the other times). I've seen every time it crashes, at the beginning of boot time it pauses for 4 or 5 seconds just at the line: "radefonfb: Dynamic Power Clock Management enabled" I wish this information can help you. :-) If you need something else, tell me. Thanks. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
hi, here is the info related to the processor from proc/cpuinfo processor : 0 cpu : 750FX temperature : 52 C (uncalibrated) clock : 900MHz revision: 2.3 (pvr 7000 0203) bogomips: 1785.85 machine : PowerBook4,3 motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh detected as : 257 (iBook 2 rev. 2) pmac flags : 001b L2 cache: 512K unified memory : 384MB pmac-generation : NewWorld I've worked all days with your patched 2.6.12-rc2 using sleep/wakeups without any problems (in both X and console sessions, with usb devices plugged in). concerning the power savings, I've done some rough statistics, and it seems that I consume 60-100 mA less (compared to a debianized 2.6.8.1 kernel ) when working with battery. I haven't yet checked the influence on the total autonomy (~5h with 2.6.8 and low lcd brightness). I hope that it will help, and thanks again for your great job. hubert Benjamin Herrenschmidt a écrit : On Wed, 2005-04-06 at 10:15 +0200, klein wrote: Hi, I've applied your patches to a 2.6.12-rc2 tree, and all works fine with my ibook 2.2. sleep / wakeup (to ram) works smoothly, no problems with hotplugging... cpufreq works also correctly I will check power savings today What is the CPU model & revision on this iBook ? Thanks, Ben. -- == Hubert Klein <[EMAIL PROTECTED]> CRMC-N UPR 7251 CNRS Faculte des Sciences de Luminy, case 901, F-13288 Marseille cedex 09 tel : +33(0)491 829 302 / fax : +33(0)491 829 176 == -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello. First, where can I get the latest bk snapshot of the kernel ? Or is the latest -rc2 enough ? Can anyone send me a working .config to test the patches ? Somehow any of my self-compiled 2.6 kernels won't boot. Don't know the reason... *Kristian -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.0 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFCVORRkngZ5psz1ewRArquAJ9Nw3HxLCbGQ0Q1GJ9p/mPr0BflTgCg2N8+ p6lV/36WqS1CqHj4S4cV958= =2Ect -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Wed, 2005-04-06 at 19:40 +0200, Esteban Martinez wrote: > Hi Ben! Here I'm again. > > > What about with none of my patches ? (Just plain 2.6.12-rc1-bk6). > Ok. Those tests are with plain 2.6.12-rc1-bk6 kernel. I summarize you: > 1) Boot in console mode. > 2) Sleep/wakeup cycle correctly. > 3) After wake up, I start X and it crashes (black screen, no keyboard, but hdd >continues spinning) You haven't tried: 1) Boot in console mode 2) Test X (with no sleep wkaeup cycle and with DRI enabled). Can you try and let me know ? > Test with DRI _disabled_: > 1) Boot in console mode. > 2) Just after I've logged, I start X and it works. After that, I come back to >console mode and the screens stays for a few seconds with vertical green >slashes, and then appears the console. I can contine working properly. > 3) After that I do a sleep/wakeup cycle and it crashes when it tries to wake >up (black screen, no keyboard), but this time the hard disk continues >spinning. Ok. Can you also try the attached patch ? > > > I have no idea what's wrong for now, the log looks normal. Can you try > > not enabling DRI see if it makes a difference ? > This test is with 2.6.12-rc1-bk6 kernel plus your patches and the last > change (lis -> li...). I've disabled DRI and it works. The test is: > 1) Boot in console mode. > 2) Start X and it works. I can work properly. After that I return to the >console and the screens shows vertical green slashes for a few seconds, and >then returns to console. I can continue working. All ok. > 3) I repeat the point 2 and it ocurrs the same. > > > > If it still crashes, I would appreciate some regression testing, that is > > going backward in the bk snapshots to point out which one precisely > > introduced the problem. > After all, I only left to test the kernel 2.6.12-rc1-bk5 and below, and the > 2.6.12-rc2 kernel. Let me know which one do you want I test. > > I wait news from you. Thanks a lot. :-) > > Esteban. > > > -- > Keep it soulful and spread love! > "Ask and it will given to you; seek and you will find; > knock and the door will be opened to you." (Mathew 7:7) > > -- Benjamin Herrenschmidt <[EMAIL PROTECTED]> Index: linux-work/drivers/char/agp/uninorth-agp.c === --- linux-work.orig/drivers/char/agp/uninorth-agp.c 2005-03-15 11:57:17.0 +1100 +++ linux-work/drivers/char/agp/uninorth-agp.c 2005-04-05 15:20:29.0 +1000 @@ -10,6 +10,7 @@ #include #include #include +#include #include "agp.h" /* @@ -26,6 +27,7 @@ static int uninorth_rev; static int is_u3; + static int uninorth_fetch_size(void) { int i; @@ -264,7 +266,8 @@ &scratch); } while ((scratch & PCI_AGP_COMMAND_AGP) == 0 && ++timeout < 1000); if ((scratch & PCI_AGP_COMMAND_AGP) == 0) - printk(KERN_ERR PFX "failed to write UniNorth AGP command reg\n"); + printk(KERN_ERR PFX "failed to write UniNorth AGP" + " command register\n"); if (uninorth_rev >= 0x30) { /* This is an AGP V3 */ @@ -278,13 +281,24 @@ } #ifdef CONFIG_PM -static int agp_uninorth_suspend(struct pci_dev *pdev, pm_message_t state) +/* + * These Power Management routines are _not_ called by the normal PCI PM layer, + * but directly by the video driver through function pointers in the device + * tree. + */ +static int agp_uninorth_suspend(struct pci_dev *pdev) { + struct agp_bridge_data *bridge; u32 cmd; u8 agp; struct pci_dev *device = NULL; - if (state != PMSG_SUSPEND) + bridge = agp_find_bridge(pdev); + if (bridge == NULL) + return -ENODEV; + + /* Only one suspend supported */ + if (bridge->dev_private_data) return 0; /* turn off AGP on the video chip, if it was enabled */ @@ -309,12 +323,13 @@ printk("uninorth-agp: disabling AGP on device %s\n", pci_name(device)); cmd &= ~PCI_AGP_COMMAND_AGP; - pci_write_config_dword(device, agp + PCI_AGP_COMMAND, cmd); + pci_write_config_dword(device, agp + PCI_AGP_COMMAND, cmd); } /* turn off AGP on the bridge */ agp = pci_find_capability(pdev, PCI_CAP_ID_AGP); pci_read_config_dword(pdev, agp + PCI_AGP_COMMAND, &cmd); + bridge->dev_private_data = (void *)cmd; if (cmd & PCI_AGP_COMMAND_AGP) { printk("uninorth-agp: disabling AGP on bridge %s\n", pci_name(pdev)); @@ -329,9 +344,23 @@ static int agp_uninorth_resume(struct pci_dev *pdev) { + struct agp_bridge_data *bridge; + u32 command; + + bridge = agp_find_bridge(pdev); + if (bridge == NULL) + return -ENODEV; + + command = (u32)bridge->dev_private_data; + bridge->dev_private_data = NULL; + if (!(command & PCI_AGP_COMMAND_AGP)) + return 0; + + uninorth_agp_enable(bridge, command); + return 0; } -#endif +#endif /* CONFIG_PM */ static int uninorth_create_gatt_table(struct agp_bridge_data *bridge) { @@ -575,6 +604,12 @@ of_node_put(uninorth_node); } +#ifdef CONFIG_PM + /* Inform platform of our suspend/resume caps */ + pmac_register_agp_pm(pdev, agp_
Re: iBook G3 owners
> If it still crashes, I would appreciate some regression testing, that > is going backward in the bk snapshots to point out which one precisely > introduced the problem. Hey Ben. I've tested kernel 2.6.12-rc1-bk5 plain and 2.6.12-rc1-bk5 with the your patches (plus the last one) and in both cases it crashes when I try to start X. One of those times I was testing, I forgot to apply your last patch (this one: replace "lis r4,0xfff0" for "li r4,0") and when I tried to wake up after resume, it crashes. After that I've made the change and the cycle resume/wakeup works correctly. I tell you this because maybe it can be useful. By the way, I've tested too the 2.6.10 kernel plain and all works properly (resume/wakeup, X, ...) Now I don't know if it will be more useful for you if I continue testing the 2.6.12-rc1-bk4 and below, or try 2.6.11 plain and then going up. What do you think about it? Thanks. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> If it still crashes, I would appreciate some regression testing, that > is going backward in the bk snapshots to point out which one precisely > introduced the problem. Hi Ben. Just now I've tested the new release candidate: 2.6.12-rc2 (plain, no extra patches). I've tested with cpudyn enabled and the cycles resume/wakeup work correctly. And also, if I start X with DRI _disabled_ it works. The problem comes if I enable DRI and try to start X. It crashes and I have to power off the ibook pressing turn off button for a few seconds. Now I will test the 2.6.12-rc1-bk5 plain and I will let you know. Thanks. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> Indeed, making those functions stubs returns everything to normal. > > My video card is the ATI Radeon Mobility M6 LY, so maybe those errata > need to be disabled for that card? No they need to be here to workaround a very rare problem, but they seem to slow things down a lot more than expected. I need to investigate. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Hi Ben! Here I'm again. > What about with none of my patches ? (Just plain 2.6.12-rc1-bk6). Ok. Those tests are with plain 2.6.12-rc1-bk6 kernel. I summarize you: 1) Boot in console mode. 2) Sleep/wakeup cycle correctly. 3) After wake up, I start X and it crashes (black screen, no keyboard, but hdd continues spinning) Test with DRI _disabled_: 1) Boot in console mode. 2) Just after I've logged, I start X and it works. After that, I come back to console mode and the screens stays for a few seconds with vertical green slashes, and then appears the console. I can contine working properly. 3) After that I do a sleep/wakeup cycle and it crashes when it tries to wake up (black screen, no keyboard), but this time the hard disk continues spinning. > I have no idea what's wrong for now, the log looks normal. Can you try > not enabling DRI see if it makes a difference ? This test is with 2.6.12-rc1-bk6 kernel plus your patches and the last change (lis -> li...). I've disabled DRI and it works. The test is: 1) Boot in console mode. 2) Start X and it works. I can work properly. After that I return to the console and the screens shows vertical green slashes for a few seconds, and then returns to console. I can continue working. All ok. 3) I repeat the point 2 and it ocurrs the same. > If it still crashes, I would appreciate some regression testing, that is > going backward in the bk snapshots to point out which one precisely > introduced the problem. After all, I only left to test the kernel 2.6.12-rc1-bk5 and below, and the 2.6.12-rc2 kernel. Let me know which one do you want I test. I wait news from you. Thanks a lot. :-) Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Wed, Apr 06, 2005 at 08:35:33PM +1000, Benjamin Herrenschmidt wrote: > On Wed, 2005-04-06 at 19:41 +1000, Ian Wienand wrote: > > On my 700Mhz 750FX iBook, with 2.6.12-rc2 and your patches from the > > website I require those extra changes to make it work. Without them, > > the iBook goes to sleep OK, but doesn't seem to wake up (this has been > > a problem since somewhere just after 2.6.10 IIRC). > > Cool ! Can you tell me the revision (PVR value) of the CPU ? It's all > in /proc/cpuinfo For completeness, here is my /proc/cpuinfo processor : 0 cpu : 750FX temperature : 23-25 C (uncalibrated) clock : 700MHz revision: 1.18 (pvr 7000 0112) bogomips: 1388.54 machine : PowerBook4,3 motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh detected as : 257 (iBook 2 rev. 2) pmac flags : 001b L2 cache: 512K unified memory : 256MB pmac-generation : NewWorld > Ok. I suspect the wakeup effect might be caused by the delays in > radeonfb that I added to work around some chip errata... You can try > hacking them out from the errata function in radeonfb.h Indeed, making those functions stubs returns everything to normal. My video card is the ATI Radeon Mobility M6 LY, so maybe those errata need to be disabled for that card? -i signature.asc Description: Digital signature
Re: iBook G3 owners
On Wed, 2005-04-06 at 19:41 +1000, Ian Wienand wrote: > On Wed, Apr 06, 2005 at 09:28:08AM +1000, Benjamin Herrenschmidt wrote: > > Ok, can you try this please: After you have applied my patches, edit the > > file arch/ppc/platforms/pmac_cache.S, and spot those 2 bits of code: > > On my 700Mhz 750FX iBook, with 2.6.12-rc2 and your patches from the > website I require those extra changes to make it work. Without them, > the iBook goes to sleep OK, but doesn't seem to wake up (this has been > a problem since somewhere just after 2.6.10 IIRC). Cool ! Can you tell me the revision (PVR value) of the CPU ? It's all in /proc/cpuinfo > X works OK, and the iBook sleeps OK in X (Radeon Mobility M6) but when > it wakes up the screen is very strange to start with; you can see > "echos" of the terminal up the top half of the screen and the > background seems dithered ... then it slowly seems to add bring back > the colours over a one-two seconds till it comes back to life. It's a > very strange effect that I've never seen before. > > The CPU speed seems to work fine, and it sleeps and wakes in both > 700Mhz and 400Mhz mode. Ok. I suspect the wakeup effect might be caused by the delays in radeonfb that I added to work around some chip errata... You can try hacking them out from the errata function in radeonfb.h Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Wed, Apr 06, 2005 at 09:28:08AM +1000, Benjamin Herrenschmidt wrote: > Ok, can you try this please: After you have applied my patches, edit the > file arch/ppc/platforms/pmac_cache.S, and spot those 2 bits of code: On my 700Mhz 750FX iBook, with 2.6.12-rc2 and your patches from the website I require those extra changes to make it work. Without them, the iBook goes to sleep OK, but doesn't seem to wake up (this has been a problem since somewhere just after 2.6.10 IIRC). X works OK, and the iBook sleeps OK in X (Radeon Mobility M6) but when it wakes up the screen is very strange to start with; you can see "echos" of the terminal up the top half of the screen and the background seems dithered ... then it slowly seems to add bring back the colours over a one-two seconds till it comes back to life. It's a very strange effect that I've never seen before. The CPU speed seems to work fine, and it sleeps and wakes in both 700Mhz and 400Mhz mode. -i [EMAIL PROTECTED] http://www.gelato.unsw.edu.au signature.asc Description: Digital signature
Re: iBook G3 owners
On Wed, 2005-04-06 at 10:15 +0200, klein wrote: > Hi, > I've applied your patches to a 2.6.12-rc2 tree, and all works fine with > my ibook 2.2. > sleep / wakeup (to ram) works smoothly, no problems with hotplugging... > cpufreq works also correctly > I will check power savings today What is the CPU model & revision on this iBook ? Thanks, Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Hi, I've applied your patches to a 2.6.12-rc2 tree, and all works fine with my ibook 2.2. sleep / wakeup (to ram) works smoothly, no problems with hotplugging... cpufreq works also correctly I will check power savings today hubert Benjamin Herrenschmidt a écrit : Hi ! There have been various reports of issues with sleep among others on iBook G3 equiped with the 750FX processor. Also, the cpufreq code on these so far didn't change the CPU voltage, which limited the actual power saving at low frequency. I have uploaded various patches that should help fix these issues. Those patches are all against current linus bk and they should be all applied in the order below. I would really appreciate some tests as I don't have access to any of these machines. I need to know if cpufreq works reliably with those patches and if the new voltage control makes any differnece on battery life (check power consumption in /proc/pmu/battery_*/current when running on battery) and I need to know if the patches are improving reliability of sleep/wakeup. The 4 patches can be found at these URLs. If you had earlier versions of some of these, those patches replace them: http://gate.crahsing.org/~benh/ppc32-750-errata-fix.diff http://gate.crahsing.org/~benh/ppc32-pmac-sleep-fix.diff http://gate.crahsing.org/~benh/cpufreq-add-suspend.diff http://gate.crahsing.org/~benh/ppc32-cpufreq-gpio-off.diff Please, let me know asap, Ben. -- == Hubert Klein <[EMAIL PROTECTED]> CRMC-N UPR 7251 CNRS Faculte des Sciences de Luminy, case 901, F-13288 Marseille cedex 09 tel : +33(0)491 829 302 / fax : +33(0)491 829 176 == -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Wed, 2005-04-06 at 04:50 +0200, Esteban Martinez wrote: > > The X problem is new ? > No. I had this problem before, with your first patch. I'm sorry if I not > mentioned it before. :-/ > > > Does it happen after a sleep/wake cycle or all the time ? > It happens all the time, when I just booted in console mode and I want to > start X, and also after sleep/wake cycle. > > > Does it happen without CPUFREQ ? > Yes, it does. In both cases it crashes: Starting X just booted and after > sleep/wakeup cycle. > > > Can you also try booting with radeonfb.default_dynclk=0 on the kernel > > command line and tell me if that makes a difference ? > Yes, I tried it and also it crashes. CPUFREQ was disabled. > > I attach you a fragment from my XFree86.0.log with the 2.6.12-rc1-bk6 kernel, > just when it crashes. I hope it could be useful. What about with none of my patches ? (Just plain 2.6.12-rc1-bk6). If it still crashes, I would appreciate some regression testing, that is going backward in the bk snapshots to point out which one precisely introduced the problem. I have no idea what's wrong for now, the log looks normal. Can you try not enabling DRI see if it makes a difference ? Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> The X problem is new ? No. I had this problem before, with your first patch. I'm sorry if I not mentioned it before. :-/ > Does it happen after a sleep/wake cycle or all the time ? It happens all the time, when I just booted in console mode and I want to start X, and also after sleep/wake cycle. > Does it happen without CPUFREQ ? Yes, it does. In both cases it crashes: Starting X just booted and after sleep/wakeup cycle. > Can you also try booting with radeonfb.default_dynclk=0 on the kernel > command line and tell me if that makes a difference ? Yes, I tried it and also it crashes. CPUFREQ was disabled. I attach you a fragment from my XFree86.0.log with the 2.6.12-rc1-bk6 kernel, just when it crashes. I hope it could be useful. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) XFree86.0.log.gz Description: Binary data
Re: iBook G3 owners
On Wed, 2005-04-06 at 02:54 +0200, Esteban Martinez wrote: > > Ok, can you try this please: After you have applied my patches, edit the > > file arch/ppc/platforms/pmac_cache.S, and spot those 2 bits of code: > > [...] > > In those bits of code, replace the statements > > lis r4,0xfff0 > > with > > li r4,0 > > [...] > > And let me know if that helps. > > Ok, I've made the 3 changes, as you said me, and recompiled the kernel. > And it WORKS! :-) Ben, you are the master. :-) I've tested with no cpudyn > and manually changing the cpu frequency, in both cases (powersave and > performance) it works properly. I've attached a log file with information > about "/proc/cpuinfo" and "/proc/pmu/battery_0" in every case. > > Just I've one problem. If I try to start X, the system hangs up, with no > response through the keyboard, black lcd, but the hard disk continues > spinning, and then I've to power off the computer pressing the turn-off > button. The X problem is new ? Does it happen after a sleep/wake cycle or all the time ? Does it happen without CPUFREQ ? Can you also try booting with radeonfb.default_dynclk=0 on the kernel command line and tell me if that makes a difference ? > As always, if you need something else, tell me. Thanks for your work > and keep it up. :-) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> Ok, can you try this please: After you have applied my patches, edit the > file arch/ppc/platforms/pmac_cache.S, and spot those 2 bits of code: > [...] > In those bits of code, replace the statements > lis r4,0xfff0 > with > li r4,0 > [...] > And let me know if that helps. Ok, I've made the 3 changes, as you said me, and recompiled the kernel. And it WORKS! :-) Ben, you are the master. :-) I've tested with no cpudyn and manually changing the cpu frequency, in both cases (powersave and performance) it works properly. I've attached a log file with information about "/proc/cpuinfo" and "/proc/pmu/battery_0" in every case. Just I've one problem. If I try to start X, the system hangs up, with no response through the keyboard, black lcd, but the hard disk continues spinning, and then I've to power off the computer pressing the turn-off button. As always, if you need something else, tell me. Thanks for your work and keep it up. :-) Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) test.tar.gz Description: Binary data
Re: [Fwd: Re: iBook G3 owners]
> I have the same problem. I'm also seeing the same hang when the KDED > Media Manager is running (independent of hal/dbus). I believe the cause > for this is that the ide subsystem wakes up too late, and the cdrom > polling of hal or kded is resuming too early. Yes, but that should not cause a hang... the ioctl should just be blocked until the driver is resumed, though there must be some kind of bug preventing that from working... Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Tue, 2005-04-05 at 15:10 +0200, Esteban Martinez wrote: > > Ok, before those, just download the patches again. I found and fixed a > > couple of bugs that can explain the problem. > Hi Ben! Here we go again: > > I've done the same tests with your new patches and in all cases it happens the > same. Resume correctly, but hang up when wakeup. I got removed "cpudyn" and > I've changed manually the cpu frecuency through: echo "powersave"... Until I > don't have tested with CPU_FREQ disabled in kernel. > > As I said in other email, just at the beginning of start booting, when it's > loading the kernel, the computer just stop a few seconds (4 or 5) and the last > message line it shows me is: > "radeonbf: Dynamic Power Clock Management enabled" > After that, it continues booting correctly. It's weird because with my old > kernel (2.6.9) this didn't happen. > > I've read in Sebastian's email he has installed "dbus". I also have it > installed. I don't know if the problem is there. I've stopped the "dbus-1" > daemon and then resume. And crash at wakep too. I attach you my config > kernel fot 2.6.12-rc1-bk6, the programs I start in "/etc/rcS.d/", the programs > I have in "/etc/init.d", and syslog from my old kernel (2.6.9) and the new > kernel (2.6.12-rc1-bk6). > > Perhaps there is an option wrong or not needed in my kernel configuration. Ok, can you try this please: After you have applied my patches, edit the file arch/ppc/platforms/pmac_cache.S, and spot those 2 bits of code: /* disp-flush L1 */ li r4,0x4000 mtctr r4 lis r4,0xfff0 1: lwz r0,0(r4) addir4,r4,32 bdnz1b sync isync and 1: /* disp-flush L2. The interesting thing here is that the L2 can be * up to 2Mb ... so using the ROM, we'll end up wrapping back to memory * but that is probbaly fine. We disp-flush over 4Mb to be safe */ lis r4,2 mtctr r4 lis r4,0xfff0 1: lwz r0,0(r4) addir4,r4,32 bdnz1b ../.. In those bits of code, replace the statements lis r4,0xfff0 with li r4,0 (There are 3 occurences, you don't have to care about what happens after the label flush_disable_745x). And let me know if that helps. Thanks ! Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> Ok, before those, just download the patches again. I found and fixed a > couple of bugs that can explain the problem. Hi Ben! Here we go again: I've done the same tests with your new patches and in all cases it happens the same. Resume correctly, but hang up when wakeup. I got removed "cpudyn" and I've changed manually the cpu frecuency through: echo "powersave"... Until I don't have tested with CPU_FREQ disabled in kernel. As I said in other email, just at the beginning of start booting, when it's loading the kernel, the computer just stop a few seconds (4 or 5) and the last message line it shows me is: "radeonbf: Dynamic Power Clock Management enabled" After that, it continues booting correctly. It's weird because with my old kernel (2.6.9) this didn't happen. I've read in Sebastian's email he has installed "dbus". I also have it installed. I don't know if the problem is there. I've stopped the "dbus-1" daemon and then resume. And crash at wakep too. I attach you my config kernel fot 2.6.12-rc1-bk6, the programs I start in "/etc/rcS.d/", the programs I have in "/etc/init.d", and syslog from my old kernel (2.6.9) and the new kernel (2.6.12-rc1-bk6). Perhaps there is an option wrong or not needed in my kernel configuration. I wait for your orders. :-) Thanks. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) info.tar.gz Description: Binary data
Re: [Fwd: Re: iBook G3 owners]
Sebastian Heutling <[EMAIL PROTECTED]> writes: > Sleep/wakeup works with a modified arch/ppc/platform/pmac_sleep.S (bl > reloc_offset was missing which you send later). I still have a strange > problem with some programs started in init-Scripts. For example if I run > dbus on bootup and directly after that go to sleep and wake-up again my > ibook doesn't resume harddisks leaving the following last messages: > > eth0 resuming > PHY ID: 4061e4, addr: 0 > eth0: Link is up at 100 Mbps, full duplex. > eth0: Pause is disabled > > If I disable dbus (stop it) and go to sleep, wakeup - no problem. I have the same problem. I'm also seeing the same hang when the KDED Media Manager is running (independent of hal/dbus). I believe the cause for this is that the ide subsystem wakes up too late, and the cdrom polling of hal or kded is resuming too early. Andreas. -- Andreas Schwab, SuSE Labs, [EMAIL PROTECTED] SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5 "And now for something completely different." -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
[Fwd: Re: iBook G3 owners]
Bah! Different mailing list, different behaviour if you click on "Reply to Sender only"! ;-) Sebasitna --- Begin Message --- Benjamin Herrenschmidt wrote: Hi ! There have been various reports of issues with sleep among others on iBook G3 equiped with the 750FX processor. Also, the cpufreq code on these so far didn't change the CPU voltage, which limited the actual power saving at low frequency. I have uploaded various patches that should help fix these issues. Those patches are all against current linus bk and they should be all applied in the order below. I would really appreciate some tests as I don't have access to any of these machines. I need to know if cpufreq works reliably with those patches and if the new voltage control makes any differnece on battery life (check power consumption in /proc/pmu/battery_*/current when running on battery) and I need to know if the patches are improving reliability of sleep/wakeup. The 4 patches can be found at these URLs. If you had earlier versions of some of these, those patches replace them: http://gate.crahsing.org/~benh/ppc32-750-errata-fix.diff http://gate.crahsing.org/~benh/ppc32-pmac-sleep-fix.diff http://gate.crahsing.org/~benh/cpufreq-add-suspend.diff http://gate.crahsing.org/~benh/ppc32-cpufreq-gpio-off.diff Please, let me know asap, Sleep/wakeup works with a modified arch/ppc/platform/pmac_sleep.S (bl reloc_offset was missing which you send later). I still have a strange problem with some programs started in init-Scripts. For example if I run dbus on bootup and directly after that go to sleep and wake-up again my ibook doesn't resume harddisks leaving the following last messages: eth0 resuming PHY ID: 4061e4, addr: 0 eth0: Link is up at 100 Mbps, full duplex. eth0: Pause is disabled If I disable dbus (stop it) and go to sleep, wakeup - no problem. Sometimes it is even enough to just only restart the daemon. I'm not sure about cpufreq. I used to run the cpufreq daemon. With this program I realised three possible frequencies on my ibook: 400, 462, 700 MHz. Using powernowd or directly the userspace governour gives only 400 and 700MHz. So maybe there is something wrong here, too or this specific CPU/ibook does only support 2 states. Details about the ibook: iBook, Dual USB blueberry:~# cat /proc/cpuinfo processor : 0 cpu : 750FX temperature : 1-76 C (uncalibrated) clock : 700MHz revision: 2.2 (pvr 7000 0202) bogomips: 1388.54 machine : PowerBook4,3 motherboard : PowerBook4,3 MacRISC2 MacRISC Power Macintosh detected as : 257 (iBook 2 rev. 2) pmac flags : 001b L2 cache: 512K unified memory : 384MB pmac-generation : NewWorld Sebastian --- End Message ---
Re: iBook G3 owners
First of all, the tests are done with all USB devices unplugged, before turn on the iBook and with the kernel 2.6.12-rc1-bk6. > Ok, several things to test here: > - First, did sleep/wakeup work previously ? Yes, it did. For my working kernel (2.6.9 with no patches, take it from www.kernel.org directly) sleep/wakeup works properly. The only problems there are the known bugs derived from USB plugged devices. But I unplugged all USB devices (usb-wireless, usb-mouse) before sleep, then wakeup, all works fine and I can plug the USB devices again and continue working with no problems. > - Try without every going to X (boot in console mode) and tell me if > sleep/wakeup works I booted in console mode and it doesn't work. Furthermore, if I try go to X from the console, the system crash exactly as if it had woken up after resume. > - Can you disable cpudyn and manually change the CPU speed (doing echo > "powersave" or "performance" >/sys/devices/system/cpu/cpu0/scaling_governor) > and try sleep/wakeup in both low and high speed and tell me if both are > crashing In my system I have: "/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor". Well, both are crashing. I've removed cpudyn and manually I've checked out those cases: 1) echo "powersave"... then resume/wakeup: crash at wakeup 2) echo "performance"... then resume/wakeup: crash at wakeup 3) I just boot and in console mode resume/wakeup: crash at wakeup > - Can you disable CONFIG_CPU_FREQ and tell me if it sleep/wakeup works I've disabled CPU_FREQ and recompiled the kernel and it doesn't work. It crash too, like in the other cases (black screen, no keyboard...) power off through the turn-off button. :-( > Thanks, > Ben. Thanks to you :-) If you wanna I can test other kernels (original 2.6.11, or 2.6.10, or with other patches) or whatever, no problem, just tell me. Ok, now I've seen your other email. I check this out as soon as possible. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> Ok, several things to test here: > > - First, did sleep/wakeup work previously ? > - Try without every going to X (boot in console mode) and tell me if > sleep/wakeup works > - Can you disable cpudyn and manually change the CPU speed (doing echo > "powersave" or "performance" > >/sys/devices/system/cpu/cpu0/scaling_governor) and try sleep/wakeup in > both low and high speed and tell me if both are crashing > - Can you disable CONFIG_CPU_FREQ and tell me if it sleep/wakeup works Ok, before those, just download the patches again. I found and fixed a couple of bugs that can explain the problem. Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
On Mon, 2005-04-04 at 23:56 +0200, Esteban Martinez wrote: > Hi Ben! > > After downloading kernel 2.6.11 sources, patching 2.6.12-rc1, then patching > 2.6.12-rc1-bk6 and compiling, I have done some test as you asked. > > I've attached a log file with a little explanation. Basically, I've put > information about "/proc/cpuinfo" and "/proc/pmu/battery_0" before and after > to suspend to ram. I got an iBook G3/600Mhz/256MB/20GB/Ati16MB. I've "cpudyn" > installed and for this reason the cpu frecuency is 400Mhz, because it's in > powersave mode. > > I don't know if I've done some wrong (I'm a new kid on the block on this > :-)), but with the 2.6.12-rc1-bk6 kernel, just when starts the boot Debian > Linux, the machine stay a few seconds quiet, just when the kernel is > loaded. With my old kernel this didn't happen. Then it continues, and boot > correctly, as always. Also, when I close the lid for suspend to ram (without > any usb devices plugged), it suspends correctly. The problem comes when I open > the lid and the system tries to turn on. Just at this moment, my system crash, > with a black lcd and no reponse through the keyboard. All the computer stops, > the hard disk power off... I've to push the power-off button about 5 seconds > and then I can turn on the computer again. Ok, several things to test here: - First, did sleep/wakeup work previously ? - Try without every going to X (boot in console mode) and tell me if sleep/wakeup works - Can you disable cpudyn and manually change the CPU speed (doing echo "powersave" or "performance" >/sys/devices/system/cpu/cpu0/scaling_governor) and try sleep/wakeup in both low and high speed and tell me if both are crashing - Can you disable CONFIG_CPU_FREQ and tell me if it sleep/wakeup works Thanks, Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> [...] > I would really appreciate some tests as I don't have access to any of > these machines. I need to know if cpufreq works reliably with those > patches and if the new voltage control makes any differnece on battery > life (check power consumption in /proc/pmu/battery_*/current when > running on battery) and I need to know if the patches are improving > reliability of sleep/wakeup. Hi Ben! After downloading kernel 2.6.11 sources, patching 2.6.12-rc1, then patching 2.6.12-rc1-bk6 and compiling, I have done some test as you asked. I've attached a log file with a little explanation. Basically, I've put information about "/proc/cpuinfo" and "/proc/pmu/battery_0" before and after to suspend to ram. I got an iBook G3/600Mhz/256MB/20GB/Ati16MB. I've "cpudyn" installed and for this reason the cpu frecuency is 400Mhz, because it's in powersave mode. I don't know if I've done some wrong (I'm a new kid on the block on this :-)), but with the 2.6.12-rc1-bk6 kernel, just when starts the boot Debian Linux, the machine stay a few seconds quiet, just when the kernel is loaded. With my old kernel this didn't happen. Then it continues, and boot correctly, as always. Also, when I close the lid for suspend to ram (without any usb devices plugged), it suspends correctly. The problem comes when I open the lid and the system tries to turn on. Just at this moment, my system crash, with a black lcd and no reponse through the keyboard. All the computer stops, the hard disk power off... I've to push the power-off button about 5 seconds and then I can turn on the computer again. If you need other information or I've done some wrong or whatever, tell me. I want to help. :-) Thanks again for your work and your patience. Regards. Esteban. -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) tests.tar.gz Description: Binary data
Re: iBook G3 owners
On Mon, 2005-04-04 at 14:32 +0200, Jean-Paul Vincent wrote: > Hi! > > < Le lundi 04 avril 2005 à 08:06:23 +0200 (CEST), Benjamin Herrenschmidt > écrivit :> > > > Hi ! > > > > There have been various reports of issues with sleep among others on > > iBook G3 equiped with the 750FX processor. Also, the cpufreq code on > > these so far didn't change the CPU voltage, which limited the actual > > power saving at low frequency. > > > > I have uploaded various patches that should help fix these issues. Those > > patches are all against current linus bk and they should be all applied > > in the order below. > > > > Can't compile on 2.6.12-rc1-bk5: > -- > GEN .version > CHK include/linux/compile.h > UPD include/linux/compile.h > CC init/version.o > LD init/built-in.o > LD .tmp_vmlinux1 > arch/ppc/platforms/built-in.o(.pmac.text+0x6a88): In function > `flush_disable_caches': > : undefined reference to `cpufreq_frequency_table_verify' > arch/ppc/platforms/built-in.o(.pmac.text+0x6ac8): In function > `flush_disable_caches': > : undefined reference to `cpufreq_frequency_table_target' > arch/ppc/platforms/built-in.o(.pmac.text+0x6b54): In function > `flush_disable_caches': > : undefined reference to `cpufreq_frequency_table_cpuinfo' > make: *** [.tmp_vmlinux1] Erreur 1 Get the latest bk, a fix was pushed for that. If it's not in your tree, edit arch/ppc/kernel/Kconfig, look for the PMAC_CPU_FREQ option, and add the line: select CPU_FREQ_TABLE Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
Hi! < Le lundi 04 avril 2005 à 08:06:23 +0200 (CEST), Benjamin Herrenschmidt écrivit :> > Hi ! > > There have been various reports of issues with sleep among others on > iBook G3 equiped with the 750FX processor. Also, the cpufreq code on > these so far didn't change the CPU voltage, which limited the actual > power saving at low frequency. > > I have uploaded various patches that should help fix these issues. Those > patches are all against current linus bk and they should be all applied > in the order below. > Can't compile on 2.6.12-rc1-bk5: -- GEN .version CHK include/linux/compile.h UPD include/linux/compile.h CC init/version.o LD init/built-in.o LD .tmp_vmlinux1 arch/ppc/platforms/built-in.o(.pmac.text+0x6a88): In function `flush_disable_caches': : undefined reference to `cpufreq_frequency_table_verify' arch/ppc/platforms/built-in.o(.pmac.text+0x6ac8): In function `flush_disable_caches': : undefined reference to `cpufreq_frequency_table_target' arch/ppc/platforms/built-in.o(.pmac.text+0x6b54): In function `flush_disable_caches': : undefined reference to `cpufreq_frequency_table_cpuinfo' make: *** [.tmp_vmlinux1] Erreur 1 Regards. Jean-Paul Vincent. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> http://gate.crahsing.org/~benh/ppc32-750-errata-fix.diff > http://gate.crahsing.org/~benh/ppc32-pmac-sleep-fix.diff > http://gate.crahsing.org/~benh/cpufreq-add-suspend.diff > http://gate.crahsing.org/~benh/ppc32-cpufreq-gpio-off.diff Of course, without the typo, it gives: http://gate.crashing.org/~benh/ppc32-750-errata-fix.diff http://gate.crashing.org/~benh/ppc32-pmac-sleep-fix.diff http://gate.crashing.org/~benh/cpufreq-add-suspend.diff http://gate.crashing.org/~benh/ppc32-cpufreq-gpio-off.diff -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: iBook G3 owners
> [...] > http://gate.crahsing.org/~benh/ppc32-750-errata-fix.diff > http://gate.crahsing.org/~benh/ppc32-pmac-sleep-fix.diff > http://gate.crahsing.org/~benh/cpufreq-add-suspend.diff > http://gate.crahsing.org/~benh/ppc32-cpufreq-gpio-off.diff There is an error in the URL: replace "crahsing" -> "crashing" I own a G3 ibook. I will test those patches as soon as possible. Thanks for all your work, Ben. Esteban -- Keep it soulful and spread love! "Ask and it will given to you; seek and you will find; knock and the door will be opened to you." (Mathew 7:7) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
iBook G3 owners
Hi ! There have been various reports of issues with sleep among others on iBook G3 equiped with the 750FX processor. Also, the cpufreq code on these so far didn't change the CPU voltage, which limited the actual power saving at low frequency. I have uploaded various patches that should help fix these issues. Those patches are all against current linus bk and they should be all applied in the order below. I would really appreciate some tests as I don't have access to any of these machines. I need to know if cpufreq works reliably with those patches and if the new voltage control makes any differnece on battery life (check power consumption in /proc/pmu/battery_*/current when running on battery) and I need to know if the patches are improving reliability of sleep/wakeup. The 4 patches can be found at these URLs. If you had earlier versions of some of these, those patches replace them: http://gate.crahsing.org/~benh/ppc32-750-errata-fix.diff http://gate.crahsing.org/~benh/ppc32-pmac-sleep-fix.diff http://gate.crahsing.org/~benh/cpufreq-add-suspend.diff http://gate.crahsing.org/~benh/ppc32-cpufreq-gpio-off.diff Please, let me know asap, Ben. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]