Re: [loongson-dev] as not work well when compile android code from mips
> "李晖" == 李晖 writes: > the assembler just can not identify the multu.g instruction. > why the same assembler with the same march option can identify the > multu.g instruct in one file but can not in another file, did any one > knows the reason, it's so strange! Ar you sure that the file in question does not contain a '.set arch=' directive? See also here: http://sourceware.org/binutils/docs/as/MIPS-ISA.html#MIPS-ISA cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgpLtDAyK7yiE.pgp Description: PGP signature
Re: [loongson-dev] ld error when compile the libwebcore.so
> "李晖" == 李晖 writes: > hi all, I am porting android to loongson based on the mips android, > and the android version is froyo_r8, which can be download from > http://developer.mips.com/android/android-source-code/, my workstation > is ubuntu 9.10 in the vmware, the memory is 512M, free space of > hardware is 20G, > after some configuration and modification, I typed in "make > showcommands", when the compile system compile the webkit module, the > following error printed: [..] > /home/andy/work/mips_android_r8/prebuilt/linux-x86/toolchain/ > mips-4.4.3/bin/../lib/gcc/mips-linux-gnu/4.4.3/../../../../mips-linux- > gnu/bin/ld: final link failed: Bad value collect2: ld returned 1 exit Googeling for the "Bad value" error, it looks like there were multiple mips binutils bugs in the past that caused that error. The most recent report seems to be this one (2007, includes a patch): http://www.linux-mips.org/archives/linux-mips/2007-03/msg00197.html previously (2004) there was this bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=268001 which version of binutils are you using? Maybe it'd help to file a bug-report upstream? cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgpp3lXg8KxPW.pgp Description: PGP signature
Re: [loongson-dev] xulrunner
> "Roman" == Roman Mamedov writes: > On Fri, 11 Nov 2011 14:12:37 +0100 > Lluís Batlle i Rossell wrote: >> Hello, >> >> building firefox, I end up in this: >> ./assembler/assembler/MacroAssembler.h:46:32: fatal error: >> MacroAssemblerMIPS.h: No such file or directory >> >> Anyone with a quick patch on that? (n64 abi) > Hello, > Check here: http://www.anheng.com.cn/loongson2f/wheezy/iceweasel/ Be careful, AFAICS MIPSAsssembler.h.diff only has code for o32 ABI (maybe compatible with o64?). MIPSRegisters may need to be changed for n32/n64 ABIs. cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgpnuUxzPvKVC.pgp Description: PGP signature
[loongson-dev] Fuloong FL6004 case fan
Hi, my FL6004's fan is getting louder and louder, guess it wasn't designed to run 24/7. Anybody knows a good part to replace it? Or is it safe to just remove/unplug it? cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgpiaK8OAUjmU.pgp Description: PGP signature
Re: [loongson-dev] Fuloong FL6004 case fan
> "Roman" == Roman Mamedov writes: > On Thu, 26 Jan 2012 10:21:18 +0100 > Lluís Batlle i Rossell wrote: >> Yes, I noticed that web page, but "450g eCu" frightened me a bit. :) >> I can't imagine what is that. >> >> If fanless, perfect, of course. But changing a 'fan' looks like much >> less effort at least. Do you have any ideas in more detail, to >> achieve a fanless fuloong? > I believe it could be possible to run fanless with the standard > heatsinks. > Open the Fuloong, then: > - unscrew the heatsinks from the board; - check if thermal compound > between the chips and the heatsinks is of good quality and is enough; You forgot one step: tear the heatsinks from the ICs. When I got my fuloong I tried to remove the heat-sinks. However, they seemed to be "glued" so bady to the chips, that I couldn't, even using moderate force... David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgpwuvc72OlkS.pgp Description: PGP signature
Re: [loongson-dev] Fuloong FL6004 case fan
>>>>> "Roman" == Roman Mamedov writes: > On Thu, 26 Jan 2012 01:15:38 +0100 > David Kuehling wrote: >> Hi, >> >> my FL6004's fan is getting louder and louder, guess it wasn't >> designed to run 24/7. Anybody knows a good part to replace it? Or >> is it safe to just remove/unplug it? > What makes work towards attaining this a bit difficult is the lack of > a temperature sensor in the Fuloong. I usually use the temperature sensor of the hard-disk: smartctl --all /dev/sda|grep -i temp Package 'hddtemp' can do the same, I guess. David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40 pgp8d7GBkn5gZ.pgp Description: PGP signature
[loongson-dev] Replacement Cooling Fan for Fuloong 6014
Hi, I recently decided to replace the cooler in my Fuloong 6014, which had been running almost continuously for > 2 years. The original cooler had started to make strange grinding noises and the housing was also getting quite hot recently. I successfully used this part as a replacement: http://www.amazon.de/sunon-GM0503PFV2-8/dp/B006DX0UJ4 Datasheet here (name of part: GM0503PFV2-8) http://www.sunon.com.tw/products/pdf/maglev.pdf As this part came without any plug, I cut off the wires of the original fan and soldered the old plug to the new fan. The yellow wire was not provided by the sunon fan so I left it unconnected. The new fan makes quite some high-pitched noise, sounds like much more than the 20dBA on the product specification (maybe hitting a resonance with the fuloong enclosure?). I won't recommend it for anyone who has to sit near the Fuloong for longer periods. Though the cooling is now very efficient. Harddisk temperature went down from 50 to 41 Celsius with the new fan (recent weather changes may be a contributing factor :). For a less noisy solution maybe it'd be better to purchase the Sunon 12V cooler (e.g. GM1203PFV1-8) which will then run much slower on the 5V provided by the Fuloong's fan connector (but might overheat?). cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpCRoIYSba0L.pgp Description: PGP signature
Re: [loongson-dev] On port boost1.53 to MIPS64(el)
Hi YunQuiang, > "YunQiang" == YunQiang Su writes: > I am working on port Debian to mips64el platform (mips64elr2 indeed) > When working on boost1.54, it has a libboost-context now which is not > exist in boost1.49. > boost-context has some assembly code then mips64(el) is marked as > unsupported for now. Maybe we should port it to mips64(el) Your email is a little unclear to me. I guess the boost-context inline-assembly problem is unrelated to the compiler error you mention? Are you referring to source files ./boost_1_54_0/libs/context/src/asm/make_mips32_o32_elf_gas.S ./boost_1_54_0/libs/context/src/asm/jump_mips32_o32_elf_gas.S ? This should be pretty easy to fix. I could have a look at it, but I'm currently swamped with a lot of work, so may take some time (I didn't even have time to upgrade my fuloong from squeeze to wheezy yet). It would help a lot if you could give me login access to a (very fast :) mips debian machine that has all the required toolchain packages for mips64 builds installed. > There is a gcc internal error like > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=715526 WRT the compiler error, did you try to reproduce the error with the latest 4.8 series GCC built from sources provided by http://gcc.gnu.org/? When tracking down architecture related GCC bugs, the GCC people may be much more ressourceful (helpful) then the Debian people. In case you can reproduce the bug, just create a new bug with the single preprocessed source file that triggers the bug attached (plus GCC command line) here: http://gcc.gnu.org/bugzilla/ Just my two cents. > FYI: mips(el) architectures may be removed from Debian for next > release Yes, bad news. Maybe you should have CCed debian-mips in your email, to raise some more awareness? As long as the mips64 effort becomes usable within the next years, I for one won't miss Debian mips(32)el :) cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpD_kW0cSf4E.pgp Description: PGP signature
Re: [loongson-dev] On port boost1.53 to MIPS64(el)
> "YunQiang" == YunQiang Su writes: >>> boost-context has some assembly code then mips64(el) is marked as >>> unsupported for now. Maybe we should port it to mips64(el) >> >> Are you referring to source files >> >> ./boost_1_54_0/libs/context/src/asm/make_mips32_o32_elf_gas.S >> ./boost_1_54_0/libs/context/src/asm/jump_mips32_o32_elf_gas.S >> >> ? > Yes. [..] > I have not very much sense on assembly. Well, I have done some MIPS assembly and know boost as well, this is why I was offering my help... >> >> It would help a lot if you could give me login access to a (very fast >> :) mips debian machine that has all the required toolchain packages >> for mips64 builds installed. > They are in a internal network, which need a vpn access, which is not > controlled by me. I could give you an SSH account on my server, then you just need to run ssh -R 22:localhost:60022 -N account@my_server to open a reverse tunnel through SSH for login on your machine (maybe your admins won't like that, though). I can also give you a config file for 'monit' that keeps such tunnels open permanently. You can do a similar trick with any server with public IP you have access too, mapping your local port 22 to a publicly accessible IP address. Just in case you're still interested in any help. >>> There is a gcc internal error like >>> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=715526 >> >> WRT the compiler error, did you try to reproduce the error with the >> latest 4.8 series GCC built from sources provided by >> http://gcc.gnu.org/? [..] > The gcc-4.8_4.8-10 in Debian is update to about 2 or 3 weeks ago That's not the point. The point is: GCC developers are capable to track down & fix these bugs. But GCC people will likely ignore a bug report for a version that doesn't match exactly their sources. Don't make it unneccessarily difficult for them to help you. David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpor1IbCRRq9.pgp Description: PGP signature
Re: [loongson-dev] new Loongson-3A laptop, basic questions
> "YunQiang" == YunQiang Su writes: >>> If not, I am thinking of installing that particular kernel on my >>> laptop now, then update the repositories in /etc/apt/sources... to >>> point to offical debian wheezy, then apt-get dist-upgrade. Would >>> that work? > The current debian installer doesn't support loongson 3A. If you have > another mipsel system, which you can create a system by debootstrap, > and edit boot.cfg as description by jjzhong. > I am using the this kernel: > http://people.ubuntu.com/~wzssyqa/linux-loongson3/ It works well > except the DMA cannot work. Not being supported by the debian installer shouldn't be a problem. When I installed debian squeeze on loongson2f, looongson2f wasn't supported either. I just used the installer initrd image from http://ftp.debian.org/debian/dists/squeeze/main/installer-mipsel/current/images/malta/netboot/ and built myself a loongson-2f kernel, then booted kernel+initrd and got a working installer. Here is the script and config that I used to compile a kernel and embed the initrd in it: http://mosquito.selfhost.eu/opensvn/free/trunk/loongson Same or similar approach should work for the loongson3a. cheers and good luck, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpTCpLPD1JP2.pgp Description: PGP signature
Re: [loongson-dev] [llongson-dev] 3a/3b mini-pcs?
Am 04.12.2013 17:39 schrieb Javier Vasquez: On Mon, Jul 22, 2013 at 10:08 PM, Fuxin Zhang wrote: 于 2013/7/21 4:39, Javier Vasquez 写道: | | 于 2013/7/19 12:11, Javier Vasquez 写道: OK, thanks, I'll be waiting... Or let me know if I need to contact someone, :-) She is already in the CC list(betty zha...@lemote.com). OK, I haven't been able to get any further regarding acquiring some motherboards with loongson-3a... I'm looking forward to getting 4 of them, but I have no clue how, neither which formats I could get them (2 NICs, HDMI support, etc). Any help towards acquiring such boards is welcome. I recently ordered a Loongson 3A "desktop" system by just mailing zhangm lemote . I would have preferred a bare mini-ITX board but they don't currently sell these. They sold me the Xinghuo 6100 computer for $600 plus shipment (and customs). Here's the product page: http://www.lemote.com/products/computer/fulong/348.html Payment was in advance via international bank transfer (may add >30 EUR depending on your bank), shipment came within a few days via TNT (had to pay import tax in cash to the TNT delivery person and sign some stuff). Luckily they had my phone number and phoned me to ask when to attempt another delivery where I'd actually be at home (and then delivered saturday morning). Make sure you give your mobile phone number with your order. All in all a very pleasant experience. The computer does have a mainboard that looks like mini-ITX. A x16 PCI-e socket seems to be designed into the board, but has been omitted at assembly. Also two of four SATA sockets have been omitted. Two DDR3(?) slots (2x1GB preinstalled), ATX power connector, one large and one small fan on board (one for the CPU, one for the AMD bridge/graphics chip?). Internal power supply has a fan, plus there's another case fan mounted above the hard disk space. Metal enclosure, pretty much space for a mini-ITX system, looks like it can take at least a 3.5" hdd plus a slim DVD drive, but only a single 2.5" hard disk comes pre-installed. Will post photos as soon as I find the time. The system comes pre-installed with two different chinese linux derivates that can be selected via pmon. Only booted it once to do a quick check. Didn't try to install a more mainstream Linux (e.g. Debian) yet. cheers, David -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at http://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/groups/opt_out.
Re: [loongson-dev] [llongson-dev] 3a/3b mini-pcs?
Am 04.12.2013 19:38 schrieb Javier Vasquez: On Wed, Dec 4, 2013 at 11:03 AM, David Kuehling wrote: I recently ordered a Loongson 3A "desktop" system by just mailing zhangm lemote . I would have preferred a bare mini-ITX board but they don't currently sell these. They sold me the Xinghuo 6100 computer for $600 plus shipment (and customs). Here's the product page: http://www.lemote.com/products/computer/fulong/348.html Thanks David... I had to use google translate: http://translate.google.com/translate?sl=auto&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.lemote.com%2Fproducts%2Fcomputer%2Ffulong%2F348.html I never even bothered to have it translated. This way it's more of a surprise later :) Is it OK that page indicates loongson 2g/3a in the specs? I'm all for 3a. Any ways, every other places indicates 3a, :-) As far as I know, these machines are only sold with 3A. Product name in your invoice should list "3A" as part of the name so you'll be sure what you pay for. Also it comes with only 1 NIC. Do you know if there are other products with 2? The debian linux is having problems with USB NICs... Do you really need the second NIC? In the past I used to demand a second nick on my mini-servers as well but later turned to using VLANs and smart firewalls to just rely on a single NIC (plus a switch). VLANs work stably with the NIC that comes with Fuloong 6004. Looks like nowadays even cheap switches allow you to route VLANs onto physically separate connectors [1]. cheers, David [1] http://www.amazon.com/dp/B00AKRTLXA -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at http://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/groups/opt_out.
[loongson-dev] Loongson 3A kernel source code?
Hi, does anybody know where I can find source code for the linux kernel with support for Loongson 3A systems? I tried looking here: http://dev.lemote.com/cgit/linux-loongson-community.git/ and downloaded: http://dev.lemote.com/cgit/linux-loongson-community.git/snapshot/linux-loongson-community-3.12.4.tar.bz2 But that kernel does not seem to have any Loongson 3A support in it. I'm trying to build a kernel for my Xinghuo 3A 6100 system [1], that I can use to install debian wheezy. cheers, David [1] http://www.lemote.com/products/computer/fulong/348.html -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpdjZmMgF5k5.pgp Description: PGP signature
[loongson-dev] Loongson-3A/PMON problems
Hi, just trying to install Debian on a Xinghuo Desktop machine with Loongson 3A CPU. I built a 3.5 linux kernel that contains a minimum Debian installer as initrd. Here is the build scripts I used: http://mosquito.dyndns.tv/opensvn/free/trunk/loongson/loongson3a-debian-wheezy-installer/ And this is the resulting linux image: http://mosquito.dyndns.tv/david/loongson/vmlinuz-3.5.0-8.lemote-with-wheezy-installer I put this on a usb stick (8GB, vfat partition), under path+name "/installer". Now I boot the Xinghuo, enter the pmon prompt and type: load (usb0,0)/installer It loads for a while, outputs Loading file: (usb0,0)/installer (elf) (elf64) 0x0161/6427652 + 0x81c81404/4202540(z) Then the whole system deadlocks. Is this a pmon problem? I didn't even run the 'boot' command yet. Or could this be some problem with my kernel config (this is almost the same kernel config as the one that I found on the boot partition of the system)? Any ideas how to proceed would be appreciated. Solving these kinds of problems via trial+error can take a lot of time... cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at http://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout.
Re: [loongson-dev] Loongson-3A/PMON problems
Replying to myself: after a little trial+error I made pmon load the kernel by separating out the initrd image. I can load it from PMON via commands load (usb0,0)/installer initrd (usb0,0)/initrd g However, there seems to be some problem with the kernel: the installer runs for about one minute, then the system deadlocks. It does not seem to be the installer's fault: if I switch to a different console and enter the installer's shell, the deadlocks happen, too. Any problems know WRT the official kernel image 3.5 from http://dev.lemote.com/cgit/linux-official.git/ ? cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpl_jakt_TjE.pgp Description: PGP signature
Re: [loongson-dev] Loongson-3A/PMON problems
> "Huacai" == Huacai Chen writes: > Hi, David, Could you please tell me the PMON version? And does 3.5.0-7 > works fine? Hi Huacei, entering 'vers' at the PMON prompt prints PMON2000 2.1 (Bonito) #1: 2013 [XXX]03[XXX]18[XXX]09:26:42 CST [XXX] is nonsensical ASCII garbage, maybe it's some multibyte chinese printed on singlebyte PMON console? I'm also tried version 3.5.0-7 from: http://dev.lemote.com/cgit/linux-official.git/snapshot/linux-official-3.5.0-7.lemote.tar.gz But this has the same problem as the 3.5.0-8 : it deadlocks after a short time of use (maybe related to the radeon framebuffer driver? the last thing I see is when the installer blanks the screen, which taks quite some time, given how slow the framebuffer is). My kernel config is here (minus the CONFIG_INITRAMFS_SOURCE=): http://mosquito.dyndns.tv/opensvn/free/trunk/loongson/loongson3a-debian-wheezy-installer/kconfig I'm using upstream GCC 4.6.4 from ftp://ftp.cs.tu-berlin.de/pub/gnu/gcc/gcc-4.6.4/gcc-core-4.6.4.tar.bz2 to build the kernel. Hopefully we're not also fighting against GCC bugs here? cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpwm6Li3Py9v.pgp Description: PGP signature
Re: [loongson-dev] Loongson-3A/linux kernel problems
I'm now going to try the master branch from git://dev.lemote.com/linux-official.git . I'm wondering which source Aurelien Jarno is using for building http://temp.aurel32.net/loongson3/linux-image-3.13-1-loongson-3_3.13.4-1+loongson3+pv19_mipsel.deb , this cannot come from Linux upstream, as there is no Loongson 3A support there yet. Looking at /boot partition that came pre-installed with the Xinghuo, the boot.cfg only lists kernel versions 2.6 for boot. There is a 3.5 kernel image, but this is commented out, maybe for the same poblems that I'm encountering? title Neokylin linux kernel /dev/fs/ext2@wd0/vmlinuz-2.6 args console=tty8 root=/dev/sda7 quiet splash loglevel=0 title LOonux 3 kernel /dev/fs/ext2@wd0/vmlinuz-2.6 args console=tty1 quiet init=/sbin/LOonux3 splash root=/dev/sda5 #title LOonux 3 (3.5+) # kernel /dev/fs/ext2@wd0/vmlinuz-3.x # args console=tty1 quiet init=/sbin/LOonux3 splash root=/dev/sda5 resume=/dev/sda3 [..] According to /boot/config-2.6, these "known good" kernels are extremely old, listing 2.6.36-21.lemote as source. David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpFL3Dy1o8yx.pgp Description: PGP signature
Re: [loongson-dev] Loongson-3A/linux kernel problems
> "Huacai" == Huacai Chen writes: > Hi, David, Please use "set" command in PMON and the "Version" variable > will give an exact version number. 'set Version' answers 'A1101-3.1.8' > Then, I think the pre-installed should be good, though I don't know > why it is commented. I'm pretty sure the included 3.5 version is not going to work either, though I'll try that tonight just to be sure. It won't be too helpful for installation, as it doesn't have the full dm-crypt support that I need during installation (I never use unencrypted disks for my computers, you never know where these computers/disks end up). I now tried Ralf's master branch, it already has the Loongson3A patches (upstream support coming closer?): git://git.linux-mips.org/pub/scm/ralf/linux It lacks the radeon firmware, had to manually add that from http://people.freedesktop.org/~agd5f/radeon_ucode/R600_rlc.bin The R8168 network driver is not supported on that branch, but googeling around I think the 8169 driver should work as a replacement and I enabled it. I do get a missing firmware warning on the kernel log, though networking seems to work (promiscous mode/DHCP may be broken? had to manually setup networking) The resulting kernel now boots into the installer. But the installation does not proceed beyond "downloading installer components". This is now a debian-specific problem, will try to ask at debian-mips and/or hack the installer. This looks pretty promising. About PMON: the version that came with my Fuloong 6004 also was very picky about which kernels it would load. Finally worked around these problems by installing grub. Hopefully Grub will also work on the Xinghuo. cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpJHCI5Gc04O.pgp Description: PGP signature
[loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
There seems to be something broken with the 'linux-official' [1] line of kernels. I now tested a lot of kernels from various sources and these are the results: * linux-official-3.5.0-7.lemote * linux-official-3.5.0-8.lemote : Loading the Debian wheezy initrd [2], I get a blank screen about half a minute into the installation, then the system looks frozen (maybe just the display is gone?). * linux-official master branch (abab4490cfb4b89) : Same problem as above, plus the fan is spinning very loudly (CONFIG_LOONGSON3_CPUAUTOPLUG=y is set, anything else I need for proper fan control or power management?) Also tried with the newer Debian Jessie installer initrd [5], with the same result. I think that rules out a userspace issue. I.e. linux-official looks unusable for my system :/ * ralf's linux mips branch, master (859a350909b3ee) [3] * torvald's linux branch, master (04535d273ee3e) [4] : Fan spins very loudly, but Debian installer seems to work. I get a segfault somewhere early in the installation when using the Debian Wheezy installer initrd [2], but does not seem kernel related. The Debian Jessie installer [5] works much better, so far I encountered no (unexpected) problems. Remaining questions: - What's wrong with linux-official line of kernels WRT frame buffer becoming unusable? Whatever it is, hopefully the "feature" does not get merged upstream ;) Anything else I can do to help triage/debug the problem? - How do I get proper fan control (and power management) with the more recent kernels? Ralf's and Torvald's branches lack the CONFIG_LOONGSON3_CPUAUTOPLUG option. Any workarounds? cheers, David [1] http://dev.lemote.com/cgit/linux-official.git [2] http://ftp.debian.org/debian/dists/wheezy/main/installer-mipsel/current/images/loongson-2f/netboot/initrd.gz [3] git://git.linux-mips.org/pub/scm/ralf/linux [4] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git [5] ftp://ftp.de.debian.org/debian/dists/jessie/main/installer-mipsel/current/images/loongson-2f/netboot/initrd.gz -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpN7MvebYrTJ.pgp Description: PGP signature
Re: [loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
The v3.5 kernel included in binary form on the boot disks of the Xinghuo
seems to work (file "vmlinuz3.x"). It also includes a kernel config
"config-3.x". Trying to rebuild any of included kernels using Lemote's
"official" source repositories with kernel config taken from the boot
partition does not result in a working kernel though:
* linux-official-3.5.0-1.lemote [1], using original kernel config
("config-3.x") as included on the Xinghuo's boot partition: Fan is
silent, but screen blanks after a minute or so, not usable. Radeon
firmware had to be retrived from [3], isn't included in the tarball.
* linux-official-3.5.0-8.lemote, using original kernel config
("config-3.x") as included on the Xinghuo's boot partition (plus
tweaks required by 'make oldconfig'): Fan is silent, but screen
blanks after a minute or so, not usable. Radeon firmware had to be
retrived from [3], isn't included in the tarball.
* kernel 2.6.36-21.lemote [2] using original kernel config "config-2.6.36+"
from boot disk:
This shows 4 penguins and that's it, no kernel messages, no kernel
console, nothing.
* kernel 2.6.36-23.lemote using original kernel config "config-2.6.36+"
from boot disk (plus 'make oldconfig' tweaks):
This shows 4 penguins and that's it, no kernel messages, no kernel
console, nothing.
This is getting frustrating. Whatever magic Lemote applied to their
kernels doesn't seem to have made it into their repositories.
cheers,
David
[1]
http://dev.lemote.com/cgit/linux-official.git/snapshot/linux-official-3.5.0-1.lemote.tar.bz2
[2]
http://dev.lemote.com/cgit/linux-3A.git/snapshot/linux-3A-2.6.36-21.lemote.tar.bz2
[3] http://people.freedesktop.org/~agd5f/radeon_ucode/R600_rlc.bin
--
GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg
Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F
pgpQg5c6aAd2W.pgp
Description: PGP signature
Re: [loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
>>>>> "Aron" == Aron Xu writes: > On Tue, Apr 8, 2014 at 9:08 AM, David Kuehling wrote: >> >> This is getting frustrating. Whatever magic Lemote applied to their >> kernels doesn't seem to have made it into their repositories. >> > I wonder which version of GCC are you using, I can recall that people > from Lemote said there could be problem if compile their code using a > very new version of GCC. Which is why I went conservatively with gcc-4.6.4 (4.6.3 is the standard compiler on Debian Wheezy mipsel). Also if I built Ralf's linux branch, the resulting kernel works better, which makes a gcc bug seem less likely (but won't rule it out, though). If anybody wants to try, here [1] is the Makefile I used to download and build my toolchain (just checkout via svn and run 'make gcc-4.6.4.build.stamp'. resulting toolchain ./staging/) If anybody can suggest a different compiler version I'll happily recompile and tests the kernels using it. cheers, David [1] http://mosquito.selfhost.eu/opensvn/free/trunk/loongson/loongson3a-debian-wheezy-installer/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpkXrSLjweLM.pgp Description: PGP signature
Re: [loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
Hi Huacei, > "Huacai" == Huacai Chen writes: > Hi, David, Lemote's pre-installed kernel uses exactly the same source > in http://dev.lemote.com/cgit/linux-official.git/, As I wrote before: the kernel 3.5.1 repository tag lacks a copy of the radeon firmware. A difference in the firmware may explain the problems I encountered. Any pointer to the source for the radeon firmware version used (embedded) in the pre-installed kernel images? > the only difference > is GCC, we use the cross-compiler which can be found here: > http://dev.lemote.com/files/resource/toolchain/cross-compile/ Which one? You mean the binary cross compiler in http://dev.lemote.com/files/resource/toolchain/cross-compile/x86_64_cross-loongson-gcc4_6_4-binutils2_22.tar.bz2 ? There is a tarball for gcc 4.5, too. This looks very similar to the GCC I'm using. But it's worth a try. thanks for keeping up with my complaints :) cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpdd9MPTB6gg.pgp Description: PGP signature
Re: [loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
Hi Huacai > Hi, David, Lemote's pre-installed kernel uses exactly the same source > in http://dev.lemote.com/cgit/linux-official.git/, the only difference > is GCC, we use the cross-compiler which can be found here: > http://dev.lemote.com/files/resource/toolchain/cross-compile/ You were right, the toolchain may be the culprit here! Compiling kernel 3.5.0-8.lemote with the lemote gcc-4.5.2 toolchain [1], produces a kernel that works without problems. 3.5.0-2, 3.5.0-3, 3.5.0-5 also work when built with that toolchain. Just to make sure I also built 3.5.0-2, -3, -5 with my gcc-4.6.4 toolchain and all these kernels exhibit the same screen blanking problem I encountered before. Quite unbelievable. Then I tried to build my own toolchain using upstream gcc-4.5.4 from [2] and binutils 2.22 from [3]. Kernels built with this toolchain do work :) I wonder whether Lemote's gcc-4.6.4 [4] toolchain actually works. I cannot easily test it as these binaries require libc 2.14 which is unavailable under Debian Wheezy here. I tried to build my own toolchain matching the exact versioning as lemote (gcc-4.6.4 and binutils 2.22), but kernels compiled with that version do have the blank screen problem. Does lemote use any non-upstream version of GCC (linaro?) or any other patches? Where's the source corresponding to [1,4]? Maybe it's not even a GCC problem but just some interaction between kernel bugs and CPU errata (?) that are triggered by certain machine code sequences or code addresses that just happen to be output by my 4.6.4 toolchain? I think I also need to find out how to run the PMON RAM tester to rule out any issues caused by faulty memory. cheers, David [1] http://dev.lemote.com/files/resource/toolchain/cross-compile/cross-loongson-gcc4_5_2-binutils2_21.tar.bz2 [2] ftp://ftp.cs.tu-berlin.de/pub/gnu/gcc/ [3] ftp://ftp.cs.tu-berlin.de/pub/gnu/binutils/ [4] http://dev.lemote.com/files/resource/toolchain/cross-compile/x86_64_cross-loongson-gcc4_6_4-binutils2_22.tar.bz2 -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpMQHpqrk6Ro.pgp Description: PGP signature
Re: [loongson-dev] Results for various Linux kernels on Loongson 3A/Xinghuo
And a few more distressing data points: * Building kernel 3.5.0-8.lemote natively on a Debian "stable" Wheezy System with Debian's default compiler gcc-4.6.3-14 results in another broken kernel (blank screen). * Building another mipsel cross-toolchain on amd64 using upstream GNU gcc-4.7.3 and binutils 2.23.2 results in a working kernel. I think I'll settle on that toolchain for now. With Debian's gcc-4.6.3 and my gcc-4.6.4 both exhibiting problems here, maybe Lemote should reconsider whether it's a good idea to use gcc-4.6 for any serious mips work. cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpDont4wdaP7.pgp Description: PGP signature
[loongson-dev] Errors with Lemote's 3.5.0 kernels for Loongson3A
Hi, running a Xinghuo "Spark" 3A 6100 [4] from an encrypted root filesystem (dm_crypt+lvm), with kernels 3.5.0-8 and 3.5.0-9 [1] I get plenty of these error messages in the kernel log when doing I/O: [ 933.398437] BUG: using smp_processor_id() in preemptible [] code: kworker/2:2/2763 [ 933.398437] caller is crypt_convert+0x80/0x438 [dm_crypt] [ 933.398437] Call Trace: [ 933.398437] [] dump_stack+0x8/0x34 [ 933.398437] [] debug_smp_processor_id+0xcc/0xf0 [ 933.398437] [] crypt_convert+0x80/0x438 [dm_crypt] [ 933.398437] [] kcryptd_crypt+0xe4/0x524 [dm_crypt] [ 933.398437] [] process_one_work+0x290/0x4dc [ 933.398437] [] worker_thread+0x2b4/0x404 [ 933.398437] [] kthread+0x88/0x90 [ 933.398437] [] kernel_thread_helper+0x10/0x18 I also recently encountered one system freeze directly after starting an I/O intensive operation (tar -xIvf on a kernel source), which may be related (this freeze occured with 3.5.0-8). Maybe these errors can be worked around by setting CONFIG_CRYPTO_WORKQUEUE=n? Kernel config and debian packages containing my kernel can be found here [2,3]. cheers, David [1] http://dev.lemote.com/cgit/linux-official.git [2] http://sourceforge.net/projects/loongson-stuff/files/loongson3a/ [3] http://sourceforge.net/p/loongson-stuff/code/ci/master/tree/ [4] http://www.lemote.com/products/computer/fulong/348.html -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpDIvkm4bM42.pgp Description: PGP signature
[loongson-dev] Lemote's kernels vulnerable to CVE-2014-0196
Hi, the 3.5.0-* kernels from lemote [1] are vulnerable to CVE-2014-0196 [2], which makes them somewhat unusable on multi-user systems (if you don't trust all of your users). AFAIU this patch [3] should fix the issue for 3.5.0. I just put a debian package with a fixed kernel here [4] (only 3.5.0-9.lemote-dk1 or newer, untested, YMMV etc.) cheers, David [1] http://dev.lemote.com/cgit/linux-official.git [2] http://www.linuxsecurity.com/content/view/161438 [3] https://sourceforge.net/p/loongson-stuff/code/ci/master/tree/loongson3a-debian-wheezy-installer/kpatches/010-fix-CVE-2014-0196.patch [4] https://sourceforge.net/projects/loongson-stuff/files/loongson3a/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpyIW099xnR7.pgp Description: PGP signature
[loongson-dev] Lemote's kernel master branch not currently working on Xinghuo
Hi, trying to build the master branch from [1] (5014fb051394) does not produce any kernels that run on Xinghuo 3A 6100. The screen just stays blank when loading them from the boot menu (or when boot from PMON prompt, the screen continues to show the PMON console). I tried two configurations: the one from arch/mips/configs/loongson3_defconfig and a config derived from a known-to-work config for kernel 3.5.0-8. Both configs can be found here [2]. cheers, David [1] http://dev.lemote.com/cgit/linux-official.git [2] https://stech.muecke.pw/david/loongson/5014fb051394/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpRravVk3Bxz.pgp Description: PGP signature
Re: [loongson-dev] Errors with Lemote's 3.5.0 kernels for Loongson3A
>>>>> "David" == David Kuehling writes: > running a Xinghuo "Spark" 3A 6100 [4] from an encrypted root > filesystem (dm_crypt+lvm), with kernels 3.5.0-8 and 3.5.0-9 [1] I get > plenty of these error messages in the kernel log when doing I/O: > > [ 933.398437] BUG: using smp_processor_id() in preemptible [] > code: [..] > Maybe these errors can be worked around by setting > CONFIG_CRYPTO_WORKQUEUE=n? Kernel config and debian packages > containing my kernel can be found here [2,3]. This kind of workaround seems to be impossible. CONFIG_CRYPTO_BLKCIPHER depends on CONFIG_CRYPTO_WORKQUEUE. Then let's just hope these error messages sound more severe than they actually are :/ David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpjoTkt_vcQd.pgp Description: PGP signature
Re: [loongson-dev] Errors with Lemote's 3.5.0 kernels for Loongson3A
>>>>> "David" == David Kuehling writes: > running a Xinghuo "Spark" 3A 6100 [4] from an encrypted root > filesystem (dm_crypt+lvm), with kernels 3.5.0-8 and 3.5.0-9 [1] I get > plenty of these error messages in the kernel log when doing I/O: > [ 933.398437] BUG: using smp_processor_id() in preemptible [] > code: kworker/2:2/2763 These errors go away when I use a kernel compiled with CONFIG_LOONGSON3_CPUAUTOPLUG disabled. This may point to the error being logged by use of smp_processor_id() in do_autoplug_timer(). Unfortunately without cpu autoplugging my system consumes about 8 W more when idle (or 1.50 € in electricity per months when running 24/7). I wonder whether these error messages can be savely ignored? David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpROnroTVKmi.pgp Description: PGP signature
Re: [loongson-dev] Errors with Lemote's 3.5.0 kernels for Loongson3A
> "Huacai" == Huacai Chen writes: > If it is really autoplug's bug, why the message reports error in > crypt_convert()? I'm confusing. I'm somewhat confused, too, but looking at he code, there is not a single call to smp_processor_id() in crypt_convert(). Then do_autoplug_timer() runs from some kind of timer (INIT_DELAYED_WORK_DEFERRABLE), and I assumed that it may run on top of the stack of crypt_convert, interrupting crypt_convert like an IRQ. When doing I/O intensive stuff on an encrypted partition, a _lot_ of time is spent in crypt_convert, so statistically speaking a timer would have a high probability of interrupting crypt_convert() and not any other function. However, I don't know nothing about the linux kernel, never having done any serious kernel programming, so this is second-guessing at best. David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpeOubjgjZIz.pgp Description: PGP signature
Re: [loongson-dev] Errors with Lemote's 3.5.0 kernels for Loongson3A
> "Huacai" == Huacai Chen writes: > You can try to remove this line in > arch/mips/loongson/loongson-3/loongson3_cpuautoplug.c > BUG_ON(smp_processor_id() != 0); Just removed that line, recompiled, and the kernel log still shows the BUG: messages [ 458.898437] BUG: using smp_processor_id() in preemptible [] code: kworker/2:2/248 [ 458.898437] caller is crypt_convert+0x80/0x438 [dm_crypt] [ 458.898437] Call Trace: [ 458.898437] [] dump_stack+0x8/0x34 [ 458.898437] [] debug_smp_processor_id+0xcc/0xf0 [ 458.898437] [] crypt_convert+0x80/0x438 [dm_crypt] [ 458.898437] [] kcryptd_crypt+0xe4/0x524 [dm_crypt] [ 458.898437] [] process_one_work+0x290/0x4dc [ 458.898437] [] worker_thread+0x2b4/0x404 [ 458.898437] [] kthread+0x88/0x90 [ 458.898437] [] kernel_thread_helper+0x10/0x18 So yes you're right it's not directly related to autoplug. However with autoplug disabled, these errors still seem to vanish. Will need to run a few more tests to be sure. David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgp8pe3TtUPhK.pgp Description: PGP signature
Re: [loongson-dev] Re: Lemote's kernel master branch not currently working on Xinghuo
> "Huacai" == Huacai Chen writes: > You can try the master branch again, I have updated the code. Just tried with 7a7d4e7770a (based on 3.15-rc5), and it seems to work (I accidentally configured with CONFIG_SMP disabled, will have to check another time whether SMP works). Unfortunately it doesn't seem to support fan control for the loongson3a mini-itx board (xinghuo). CPU fan keeps noisily spinning at top speed. It does have some kind of loongson3a hwmon code: $ cat /sys/class/hwmon/hwmon0/temp1_input 48000 And the wpce_fan.c seems to be compiled in: $ cat /sys/class/hwmon/hwmon2/name wpce775l-fan But I didn't find out where it would set a fan policy nor how I could do so from user space. Also comparing arch/mips/loongson/common/machtype.c between 3.5.0-9 and master, I see that all the loongson3 based systems have been reduced to 'generic-loongson-machine' (and looking at /proc/cpuinfo confirms that). David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgplrnnxHmMJT.pgp Description: PGP signature
Re: [loongson-dev] Re: Lemote's kernel master branch not currently working on Xinghuo
> "Huacai" == Huacai Chen writes: > master branch need a new version of PMON, you can use this one: > http://dev.lemote.com/files/upload/lm/kernel/testing/pmon-A1101-4.0.1.bin Same upgrade method as for Fuloong? Googeling around I only find chinese documentation for Xinghuo pmon upgrade, looks like a command such as this should work [1]: load -rf bfc0 (usb0,0)/pmon-A1101-2.0.6.bin Don't understand chinese, can you confirm that bfc0 is the correct offset for the PMON firmware? I hope PMON's flash is just plugged not soldered into the mainboard, in case something goes wrong!? cheers, David [1] http://bbs.lemote.com/archiver/?tid-71821.html -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpao4LbUKNaD.pgp Description: PGP signature
Re: [loongson-dev] Re: Lemote's kernel master branch not currently working on Xinghuo
> "Huacai" == Huacai Chen writes: > master branch need a new version of PMON, you can use this one: > http://dev.lemote.com/files/upload/lm/kernel/testing/pmon-A1101-4.0.1.bin Sorry to be nitpicking, but I can't find any source code that corresponds to recent PMON versions. Last commit in pmon.git [1] is From 2011, highest version tag is 1.6.1. I also looked here [2] but found nothing new, and upstream [3] is currently offline. Just my fault for not looking at the right spot? Or did Lemote eventually close-source their PMON development? I'd really like to stay with an open source BIOS for now. cheers, David [1] http://dev.lemote.com/cgit/pmon.git [2] http://dev.lemote.com/code/pmon [3] http://www.pmon2000.com/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpWLJQM8xkPu.pgp Description: PGP signature
[loongson-dev] lemote's 3.15-rc5 kernel reports "page allocation failure"
Hi, compiling a Loongson3a kernel with CONFIG_SMP=y on lemote's official kernel (branch master, 7a7d4e7770) shows error messages during boot up and in kernel log: [ 24.253906] modprobe: page allocation failure: order:0, mode:0x2000d1 [ 24.402343] CPU: 3 PID: 135 Comm: modprobe Not tainted 3.15.0-rc5-dk2+ #3 [..] [ 24.992187] Call Trace: [ 24.992187] [] show_stack+0x48/0x70 [ 24.992187] [] dump_stack+0x70/0x90 [ 24.992187] [] warn_alloc_failed+0x110/0x134 [ 24.992187] [] __alloc_pages_nodemask+0x750/0x864 [ 24.992187] [] new_slab+0xe0/0x348 [..] I did not see these kinds of errors with same config and CONFIG_SMP disabled. This is still with the original PMON that came installed with the Xinghuo. I cross-compile all my kernels with with gcc 4.7.3. Full output of dmesg attached below. You can find the kernels in question, including kernel config here [1] (-dk1: without SMP, -dk2: with SMP). Despite these errors, the system did not immediately crash. However I did not allow it to boot beyond the initrd for concerns about possibly corrupting my system. cheers, David [1] https://stech.muecke.pw/david/loongson/7a7d4e7770/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F [ 24.042968] r8169 :04:00.0 eth0: jumbo features [frames: 9200 bytes, tx checksumming: ko] [ 24.253906] modprobe: page allocation failure: order:0, mode:0x2000d1 [ 24.402343] CPU: 3 PID: 135 Comm: modprobe Not tainted 3.15.0-rc5-dk2+ #3 [ 24.546875] Stack : 7000cce3 0008 003d [ 24.546875]5000c000 804afd98 8131 [ 24.546875]8094e308 80a246b7 813111d0 980179166900 [ 24.546875]0087 0003 812a12a0 [ 24.546875] 807e4e28 98017f7177b8 [ 24.546875]980179166460 802885e4 8094e308 [ 24.546875]0003 0087 [ 24.546875] 98017f717700 807e885c [ 24.546875] [ 24.546875] 8020c378 002000d1 807e885c [ 24.546875]... [ 24.992187] Call Trace: [ 24.992187] [] show_stack+0x48/0x70 [ 24.992187] [] dump_stack+0x70/0x90 [ 24.992187] [] warn_alloc_failed+0x110/0x134 [ 24.992187] [] __alloc_pages_nodemask+0x750/0x864 [ 24.992187] [] new_slab+0xe0/0x348 [ 24.992187] [] __slab_alloc.constprop.49+0x240/0x3b8 [ 24.992187] [] __kmalloc+0x180/0x2c4 [ 24.992187] [] sr_probe+0x230/0x728 [sr_mod] [ 24.992187] [] driver_probe_device+0xdc/0x254 [ 24.992187] [] __driver_attach+0x7c/0xb4 [ 24.992187] [] bus_for_each_dev+0x58/0xac [ 24.992187] [] bus_add_driver+0xf0/0x1f8 [ 24.992187] [] driver_register+0xac/0xfc [ 24.992187] [] init_sr+0x40/0x74 [sr_mod] [ 24.992187] [] do_one_initcall+0x88/0x140 [ 24.992187] [] load_module+0x1c48/0x1f34 [ 24.992187] [] SyS_finit_module+0x70/0x94 [ 24.992187] [] handle_sys+0x128/0x14c [ 24.992187] [ 24.992187] Mem-Info: [ 24.992187] DMA per-cpu: [ 24.992187] CPU0: hi:0, btch: 1 usd: 0 [ 24.992187] CPU1: hi:0, btch: 1 usd: 0 [ 24.992187] CPU2: hi:0, btch: 1 usd: 0 [ 24.992187] CPU3: hi:0, btch: 1 usd: 0 [ 24.992187] DMA32 per-cpu: [ 24.992187] CPU0: hi:6, btch: 1 usd: 2 [ 24.992187] CPU1: hi:6, btch: 1 usd: 5 [ 24.992187] CPU2: hi:6, btch: 1 usd: 2 [ 24.992187] CPU3: hi:6, btch: 1 usd: 5 [ 24.992187] Normal per-cpu: [ 24.992187] CPU0: hi: 42, btch: 7 usd: 32 [ 24.992187] CPU1: hi: 42, btch: 7 usd: 39 [ 24.992187] CPU2: hi: 42, btch: 7 usd: 17 [ 24.992187] CPU3: hi: 42, btch: 7 usd: 29 [ 24.992187] active_anon:403 inactive_anon:9 isolated_anon:0 [ 24.992187] active_file:2782 inactive_file:84 isolated_file:0 [ 24.992187] unevictable:0 dirty:0 writeback:0 unstable:0 [ 24.992187] free:120334 slab_reclaimable:241 slab_unreclaimable:482 [ 24.992187] mapped:255 shmem:10 pagetables:67 bounce:0 [ 24.992187] free_cma:0 [ 24.992187] DMA free:0kB min:0kB low:0kB high:0kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:14336kB managed:0kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB unstable:0kB bounce:0kB free_cma:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? yes [ 24.992187] lowmem_reserve[]: 0 40 488 488 [ 24.992187] DMA32 free:160736kB min:464kB low:576kB high:688kB active_anon:816kB inactive_anon:16kB active_file:0
Re: [loongson-dev] Re: Recover bricked bios
> "Lluís" == Lluís Batlle i Rossell writes: > On Sat, Sep 20, 2014 at 05:19:16AM -0700, stig atle steffensen wrote: >> I booted it back up now, It does show the first 'boot' picture, and >> the text. It's a Lemote Yeeloong 8089_B notebook. But no matter how >> much I press delete or tab nothing happens, and it never get's past >> that splashscreen. >> > I remember as if there wasn't any way to recover the bios other than > changing the flash chip. I remember as if it were a removable chip. > In that case you could also use an external programmer to rewrite it. > I remember this from some readings; I don't have a yeelong myself. There may be some guys reading grub-devel [1] that have experience with Yeeloong reflashing. These people developed a version of grub that can be flashed as Yeeloong bios replacement (or at least they have for Fuloong). Maybe you can find somebody there who'd be willing to reflash your bios flash IC and mail it back to you. Maybe you'll even end up with Grub instead of the crappy pmon boot loader :) cheers, David [1] https://lists.gnu.org/mailman/listinfo/grub-devel -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F pgpET_7DEluZr.pgp Description: PGP signature
Re: [loongson-dev] Possibility to get dual socket motherboard with 3b cpu
> "Matt" == Matt Turner writes: > On Thu, Nov 20, 2014 at 7:57 AM, Alexey Shvetsov wrote: >> Is that somehow possible to get dual socket motherboard with 3b cpu? > As far as I can tell, no. I tried without success for a long time. > They (loongson.cn) would go for months without responding to emails, > and when they finally did they'd say "oh, yes, of course we'll send > you a board." They even took my address at one point, but alas no > board. > I've talked to other distro developers, and they've had the same > experiences. I believe Debian did eventually manage to get a board > though. Did you try asking lemote.com? They are now listing a 3B based server system (not sure whether it is dual-socket, though): http://www.lemote.com/products/computer/hongri/s7.html About a year ago, I bought a loongson-3A based computer [1] directly >From lemote, without any problems. cheers, David [1] http://www.lemote.com/products/computer/fulong/348.html -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at http://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout. pgpy6_QX4Shyu.pgp Description: PGP signature
Re: [loongson-dev] Possibility to get dual socket motherboard with 3b cpu
>>>>> "public" == public writes: > 26.11.2014, 18:31, "David Kuehling" : >> Did you try asking lemote.com? They are now listing a 3B based >> server system (not sure whether it is dual-socket, though): >> > Hello. How do you communicate with lemote? They had a store > http://aliexpress.com/store/801318 and > http://store.taobao.com/shop/view_shop-e31813a865b2abbeb74aaa765561f87a.htm > but now they do not work. Via email. One of their sales people once posted his email here. The relevant thread is: https://groups.google.com/forum/#!searchin/loongson-dev/loongson-3a$20mainboard/loongson-dev/twzd5TBz_1k/fZnJs5Xf_6QJ cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at http://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout. pgpMarpXWJk63.pgp Description: PGP signature
[loongson-dev] Using latest Debian Loongson-3 kernels on a XingHuo 3A "Spark"
Hi, I'm running a Xinghuo 3A Desktop machine, currently with a kernel 3.5.0-9 built from Lemote's sources. I stopped upgrading to newer kernels, as these seemed to require a Bios upgrade which I feared to brick my machine. Now Debian Jessie also has a 3.16 kernel for Loongson-3A in its package repository. How are my chances to make these run on my Xinghuo machine? Any simple-to-follow guides about how to safely upgrade the Bios on these machines? Any known-good kernels that I can safely use for booting into the system after an upgrade? thanks for any info, cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at https://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout.
Re: [loongson-dev] Using latest Debian Loongson-3 kernels on a XingHuo 3A "Spark"
Hi Fuxin, > "张福新" == 张福新 writes: > I think a safe way to try is use a ram version pmon(make tgt=ram) > first. The ram version can be loaded from disk or network at old pmon > command line. If the functionality is ok, the updating of pmon itself > is quite safe. Thanks for the reply. "make tgt=ram" means I'd have to recompile PMON from sources? Isn't the PMON version for Loongson3A/mini-ITX closed-source [1]? All the recent source-links I could find are not working any more [2] (and they may not even have the code for the mini-itx board, if I read the mail correctly). cheers, David [1] https://lists.debian.org/debian-mips/2015/02/msg4.html [2] http://www.loongnix.com:8000/dev/gitweb/pmon-loongson3 -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at https://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout.
Re: [loongson-dev] Using latest Debian Loongson-3 kernels on a XingHuo 3A "Spark"
> "Unknown" == Unknown writes: >> > I think a safe way to try is use a ram version pmon(make tgt=ram) > >> first. The ram version can be loaded from disk or network at old > >> pmon > command line. If the functionality is ok, the updating of pmon >> > itself > is quite safe. >> >> Thanks for the reply. >> >> "make tgt=ram" means I'd have to recompile PMON from sources? Isn't >> the PMON version for Loongson3A/mini-ITX closed-source [1]? All the >> recent source-links I could find are not working any more [2] (and >> they may not even have the code for the mini-itx board, if I read the >> mail correctly). > http://mirrors.ustc.edu.cn/loongson/pmon/updates.lemote.com/files/uploa > d/lm/firmware/pmon/ Thanks, this is exactly what I was looking for. Got myself a copy of the sources and looking forward to playing around with RAM-only version of PMON. cheers, David -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at https://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout.
[loongson-dev] Loongson3A mini-itx chipset cooler repacements?
Hi, the Loogson3A 6100 mini-ITX based system comes with two fans installed on the main-board: a 50 x 50 mm fan on the CPU, plus a 30x30 mm fan on the chipset. Both fans have a 3-pin connector. Is it safe to assume that these are 12 V fans and the 3rd pin is a "standard" RPM monitoring pin? Would this fan [1] work as a repacement for the chipset cooler? Any other potential problems to be aware of? (the 30mm fan in my system is not sounding very healthy any more). Thanks for any hints, cheers, David [1] http://www.amazon.de/bla/dp/B006VWJUNE/ -- GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg Fingerprint: B63B 6AF2 4EEB F033 46F7 7F1D 935E 6F08 E457 205F -- You received this message because you are subscribed to the Google Groups "loongson-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to loongson-dev+unsubscr...@googlegroups.com. To post to this group, send email to loongson-dev@googlegroups.com. Visit this group at https://groups.google.com/group/loongson-dev. For more options, visit https://groups.google.com/d/optout.
