wrong definition in cpm_8260.h?
> Joakim Tjernlund wrote: > > > The same error is in commproc.h for 8xx as well. > > It's not an error in either driver since first/last will always be > set in every buffer. There isn't any reason for the driver to test > or take any action based upon these indicators. Yes, the CPM will set these flags, but will it also clear them if the buffer is not first/last in a frame? I think not. I know that the driver is designed to always set them so in practise it won't be problem. However this could change some day. Also, is it not time to remove the code contained in #ifndef final_version macros? > > These are not data transfer error or status indicators in the sense > they indicate some condition on the network itself. These are buffer > management indicators that should be managed as such if they are used. yes I know, but I still think those flags should be included in the RX_STATS macro. Jocke > > Thanks. > > > -- Dan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
where to get ppc_8xx-gdb
In message <885489B3B89FB6449F93E525DF78777F064563 at srvnt506.ALLOPTIC.COM> you wrote: > > My embeded distribution for PPC 860 did not come with the cross-compiled > gdb. See http://www.denx.de/re/ELDK.html Best regards, Wolfgang Denk -- Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd at denx.de "If you want to eat hippopatomus, you've got to pay the freight." - attributed to an IBM guy, about why IBM software uses so much memory ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
wrong definition in cpm_8260.h?
Hi, In cpm_8260.h, BD_ENET_RX_STATS is defined to 0x1ff, which leads to the fcc driver can't clean the BD_ENET_RX_LAST&BD_ENET_RX_FIRST bits with the following statement: bdp->cbd_sc &= ~BD_ENET_RX_STATS; The first&last bits are set by the CPM not the CPU, so they should be cleared after the frame is received. So it should be: #define BD_ENET_RX_STATS ((ushort)0x0fff) /* All status bits */ The same problem in ppcboot. Shenrong ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Antw: Re: Lauterbach ICD In-Circuit Debugger
>> >> A lot of people suggest to use the Abatron BDI2000 as debug system. I used >> the Lauterbach TRACE32-ICD system with ARM processors without MMU (no >> Linux) and it works extremely well. >> The manufacturer claims it fully supports both the MMU and the Linux kernel >> (see here for more details >> http://www.lauterbach.com/cgi-bin/docloc.pl?file=rtoslinux.pdf). Can >> anybody who used it confirm this point? > >I can. You can debug the kernel and e.g. interrupts using virtual >addresses very well, you just have to specify the mapping >(e.g. for ppc map phys 0x to 0xc000) > >I've not looked in detail yet to debug user processes. > We are using the TRACE32-ICD (MPC8xx) with U-Boot and Kernel 2.4.4. Kernel debugging works without any problems. User space debugging is also fine - but only if you don't have to debug shared libraries :-( I think the T32 is a very good choice for kernel debugging, especially if you have experience with the GUI and don t need the documentation or online help ;-) If the problems with the shared libs are solved you have _one_ debugger for kernel and applications. Best regards, Klaus Heydeck -- Klaus Heydeck Kieback & Peter GmbH & Co KG Abt. Softwareentwicklung Tempelhofer Weg 50 12347 Berlin / Germany Tel. +49 30 60095 522 Fax. +49 30 60095 540 ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
wrong definition in cpm_8260.h?
Joakim Tjernlund wrote: > Yes, the CPM will set these flags, but will it also clear them if the > buffer is not first/last in a frame? I think not. It doesn't matter. The flags are always going to be set. Why bother clearing them when you always set them? > I know that the driver is designed to always set them so in practise it > won't be problem. However this could change some day. Well, when it changes we can change the code to set/clear the flags according to the buffer management algorithm. > yes I know, but I still think those flags should be included in the RX_STATS > macro. Why? Like I said in the previous message, they aren't part of any network status. They are buffer control flags and need to properly indicate how the buffers are managed, not cleared as part of network status. -- Dan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Problem with BDI2000 flash programming
Hi Jason, You are correct that your BDI2000 downloads the config file successfully. Otherwise, it should print a message like *** CONFIG: cannot open ep8260.cfg.init all the time once you telnet to BDI. I'd suggest you check permissions on /tftpboot/ppcboot.bin It should be readable by all. Happy hacking, Vladimir Gurevich ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
wrong definition in cpm_8260.h?
Hi The same error is in commproc.h for 8xx as well. > > > Hi, > > In cpm_8260.h, BD_ENET_RX_STATS is defined to 0x1ff, which leads > to the fcc driver can't clean the BD_ENET_RX_LAST&BD_ENET_RX_FIRST bits > with the following statement: > bdp->cbd_sc &= ~BD_ENET_RX_STATS; > > The first&last bits are set by the CPM not the CPU, so they > should be cleared after the frame is received. So it should be: > > #define BD_ENET_RX_STATS ((ushort)0x0fff) /* All status bits */ > > The same problem in ppcboot. > > > Shenrong > > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Lauterbach ICD In-Circuit Debugger
A lot of people suggest to use the Abatron BDI2000 as debug system. I used the Lauterbach TRACE32-ICD system with ARM processors without MMU (no Linux) and it works extremely well. The manufacturer claims it fully supports both the MMU and the Linux kernel (see here for more details http://www.lauterbach.com/cgi-bin/docloc.pl?file=rtoslinux.pdf). Can anybody who used it confirm this point? Many thanks in advance. llandre ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Lauterbach ICD In-Circuit Debugger
> > A lot of people suggest to use the Abatron BDI2000 as debug system. I used > the Lauterbach TRACE32-ICD system with ARM processors without MMU (no > Linux) and it works extremely well. > The manufacturer claims it fully supports both the MMU and the Linux kernel > (see here for more details > http://www.lauterbach.com/cgi-bin/docloc.pl?file=rtoslinux.pdf). Can > anybody who used it confirm this point? I can. You can debug the kernel and e.g. interrupts using virtual addresses very well, you just have to specify the mapping (e.g. for ppc map phys 0x to 0xc000) I've not looked in detail yet to debug user processes. Jaap-Jan > > Many thanks in advance. > > > llandre > > > --- J.G.J. Boor Lucent Technologies Nederland b.v. Room: BE-525 Optical Networking Group TMS Data Phone: +31 (0)35 687 4721 Larenseweg 50, P.O. Box 1168, Fax: +31 (0)35 687 5976 1200 BD, Hilversum, The Netherlands mailto:jjboor at lucent.com ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve('/bin/sh')
Some important informations are missing to help you. Which CPU model and revision are you using? Which kernel version, taken from where? Some newer CPU revs are incompatible with older kernels. Regards, Jean-Denis Boyer, B.Eng., Technical Leader Mediatrix Telecom Inc. 4229 Garlock Street Sherbrooke (Qu?bec) J1L 2C8 CANADA (819)829-8749 x241 > -Original Message- > From: Prakash kanthi [mailto:pkanthi at yahoo.com] > Sent: 22 janvier, 2003 14:50 > To: dnevil at snmc.com > Cc: linuxppc-embedded at lists.linuxppc.org > Subject: Re: Linux Kernel hangs after execve('/bin/sh') > > > > Hurray ... > > Daris, you suggestion seem to work. I Compiled my > hello.c with -Bstatic (size was almost 1MB), added it > to ramdisk and it works. So now the problem is > narrowed down to, being not able to work with the > shared libraries. > > One point is that, i DID NOT get any errors (atleast > on the screen) while execve was loading shared > libraries in the process of executing /bin/sh. But for > sure, my problem is related to it. > > I really appreciate any suggestions related to this > issue. > > Thanks, > Prakash ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve("/bin/sh")
On Wed, Jan 22, 2003 at 01:15:32AM -0800, Prakash kanthi wrote: | I wrote a small standalone program hello.c to test, if | i can dynamically load it from init() instead of | loading /bin/sh. I compiled the program added it to my | ramdisk and called from init(). Following is the | output. You can see that, it did all of it that it did | for /bin/sh and also tried to open /etc/ld.so.preload | and /etc/ld.so.cache and failed. But again towards the | end, it tries to open /lib/libc.so.6 and succeds but | still no output is displayed that hello.c is trying to | print. Quite sometime back, I had run into similar problem. init=/bin/bash would not work, but init=/bin/helloworld would work. Later, I realized that the board did not have FPU, and I had to pass relevant flags to glibc, while compilation. The same might be the case with your board. -- Chirag Kantharia, symonds.net/~chyrag/ ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
ppcboot for mpc8260ads board.
HI, I don't think the vmlinux at 0x0 will work. I tried the same with little success. finally we settled for ppcboot. now I am on the process of making the board up. Some parameters are to be passed from the boot loader to the OS. ( what are they?) Moreover ramdisk loading must be done by boot loader. This is what I know. if I am wrong, please correct me. regards, om. vnair at hss.hns.com wrote: > > Hi, > > Can you pls give some insight what soln worked for you. Was it using BDI or > you > ported ppcBoot. > > I am trying to put ppclinux onto IBM spruce reference board (MPC750). Although > linuxppc-2.5 tree has the support for this board, but i am not able to > finalize > my way for the initial success. > > I have a JTAG (VisionProbe) to download the image. > > Can you please list some steps to proceed. I have the kernel and the build > environment ready. I am thinking of building vmlinux (uncompressed) and > downloading it @ address 0x0 on the board. Will it work ? > > Thanks > - Vikram > > > > > > "Omanakuttan" on 01/22/2003 02:17:52 PM > > To: linuxppc-embedded at lists.linuxppc.org > cc:(bcc: Vikram Nair/HSS) > > Subject: Re: ppcboot for mpc8260ads board. > > > > > > Please ignore my question. I got it up. > > Omanakuttan wrote: > >>Hi, >>I got my bdi2000, thanks to the cogent arguments supplied by WD. >> >>Facing a problem. >>1. Is it required to install ppcboot in an mpc8260ads board? If I make a >>make bzImage , can the resultant image be directly loaded to flash and >>start running it? the bdi2000 guys says that it is not possible. Is it so? >>Does it mean that we have to port ppcboot and put it in out end product >>for loading linux? we have only 8MB of flash. >> >>2. I tried to coross compile ppcboot for mpc8260ads board. >>make mpc8260ads_config does not work. In the readme, no such target is >>supported. but in the boards directory, there is a file for mpc8260ads >>boards. I am confused and reading the makefile. But seems like a little >>complex. if anybody can throw some light on compiling ppcboot for >>mpc8260ads, it would be a great help. >> >>Thanks >>Om. >> >> >> >> > > > > > > > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
ppcboot for mpc8260ads board.
Please ignore my question. I got it up. Omanakuttan wrote: > > Hi, > I got my bdi2000, thanks to the cogent arguments supplied by WD. > > Facing a problem. > 1. Is it required to install ppcboot in an mpc8260ads board? If I make a > make bzImage , can the resultant image be directly loaded to flash and > start running it? the bdi2000 guys says that it is not possible. Is it so? > Does it mean that we have to port ppcboot and put it in out end product > for loading linux? we have only 8MB of flash. > > 2. I tried to coross compile ppcboot for mpc8260ads board. > make mpc8260ads_config does not work. In the readme, no such target is > supported. but in the boards directory, there is a file for mpc8260ads > boards. I am confused and reading the makefile. But seems like a little > complex. if anybody can throw some light on compiling ppcboot for > mpc8260ads, it would be a great help. > > Thanks > Om. > > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
where to get ppc_8xx-gdb
At 12:08 PM 1/22/2003 -0800, Muaddi, Cecilia wrote: > >Hello all, > >My embeded distribution for PPC 860 did not come with the cross-compiled >gdb. > >Can someone point me to where I can get the ppc_8xx-gdb which will allow >me to using GDB on the host=pc and target=power-pc-860 Try ELDK at denx.de. ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
My Memory map ... how do i pick the virtual address (Aside fr om i mmr and sdram)
I can tell you now the mapping between physical and virtual not as important as you make out. I have Linux running on a board I described to you and there is no RAM mapped at 0xc000. I suggest you read your chip manual more carefully. In my 860 PowerQUICC Users manual chapters 13,14 and 16. I also remember reading somewhere at IO should be mapped above 0x8000 but I have not been able to find where I read it in the last two years so that could be another urban myth that need the cold light of day on it. Regards, Rod Boyce -Original Message- From: jgdon [mailto:[EMAIL PROTECTED] Sent: Wednesday, 22 January 2003 2:56 p.m. To: Rod Boyce Cc: 'James Don'; linuxppc-embedded at lists.linuxppc.org Subject:RE: My Memory map ... how do i pick the virtual address (Aside fr om i mmr and sdram) Hi, I think the the mapping of virtual to physical address may be important ... i believe ram should be at 0xc000 for a number of reasons ... there is lots of posts regarding this ... What I am worried about is "rules of thumb" for example i have an ASIC on cs4 ... at BR4 0x0800 ... I would like to map it to 0x0800 in the MMU ... i would just like to be sure this is an ok address to check ... or if there is some generally accepts address range for "non memeory" device stuff ... I assume 'no' ... but i also assume the immr address and sdram address didn't matter at one time ... Jim On Wed, 22 Jan 2003, Rod Boyce wrote: > > I think you have the wrong end of the stick. Our boards have 32MB of RAM > located at 0. We have placed IMMR at 0xfa20. The address of 0xc000 > is a virtual address it does not have to have any relation to a physical > address this is done by the memory manager. The only constant is that the > boot flash MUST be connected to chip select 0. Everything else is mostly > software configurable. > > Regards, > Rod Boyce. > > -Original Message- > From: James Don [mailto:JDon at spacebridge.com] > Sent: Wednesday, 22 January 2003 12:58 p.m. > To: linuxppc-embedded at lists.linuxppc.org > Subject: My Memory map ... how do i pick the virtual address (Aside > from i mmr and sdram) > > > Hi, > > I am just looking at adding my other mem mapped devices to Linux for my > custom MPC860 board ... > > I know RAM has to be at 0xc000 and immr at 0xff00 ... is there any > rule of thumb for anything else? > > Jim > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
My Memory map ... how do i pick the virtual address (Aside fr om i mmr and sdram)
I think you have the wrong end of the stick. Our boards have 32MB of RAM located at 0. We have placed IMMR at 0xfa20. The address of 0xc000 is a virtual address it does not have to have any relation to a physical address this is done by the memory manager. The only constant is that the boot flash MUST be connected to chip select 0. Everything else is mostly software configurable. Regards, Rod Boyce. -Original Message- From: James Don [mailto:[EMAIL PROTECTED] Sent: Wednesday, 22 January 2003 12:58 p.m. To: linuxppc-embedded at lists.linuxppc.org Subject:My Memory map ... how do i pick the virtual address (Aside from i mmr and sdram) Hi, I am just looking at adding my other mem mapped devices to Linux for my custom MPC860 board ... I know RAM has to be at 0xc000 and immr at 0xff00 ... is there any rule of thumb for anything else? Jim ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve('/bin/sh')
Prakash, I had the same trouble -- in my case, a couple of different ways. try compiling without -Bstatic, but do 2 different trials: compile your test -msoft-float and -mhard-float -- if either of these works then the libraries you've got on your ram disk are either floating point emulation, or hardware FP, respectively. the compile must match the libs in terms of floating point.l the other things to look at are which libs you've got in your ram disk. if you've cut it down you might have cut it down too far. I'm running w/ the ELDK /lib and /usr/lib w/o anything removed. try that until you get things working and then you can explore which packages need what libs. Mark Prakash kanthi wrote: > > Hurray ... > > Daris, you suggestion seem to work. I Compiled my > hello.c with -Bstatic (size was almost 1MB), added it > to ramdisk and it works. So now the problem is > narrowed down to, being not able to work with the > shared libraries. > > One point is that, i DID NOT get any errors (atleast > on the screen) while execve was loading shared > libraries in the process of executing /bin/sh. But for > sure, my problem is related to it. > > I really appreciate any suggestions related to this > issue. > > Thanks, > Prakash > > --- Daris A Nevil wrote: > > > > Prakash, > > > > Try compiling your hello.c program with the -Bstatic > > flag. That will > > eliminate the need to load libraries, and should > > show if there are other > > problems. If that works OK then we can focus on why > > your library loading > > is failing. > > > > Regards, > > Daris Nevil > > > > > > Prakash kanthi said: > > > > > > Hi FOlks, > > > > > > I finally reached the last step (hopefully) of > > linux > > > kernel booting process. Kernel gets stuck after > > doing > > > a 'execve("/bin/sh")'. I tried using different > > > ramdisks, still no difference. > > > > > > I added few printk's in the source, following the > > > function call path from 'execve' call. Following > > is > > > the output. You can see that, execve goes through > > > without any file permission problems and > > subsequently > > > it tries to open /lib/ld.so.1 flawlessly. After > > that i > > > see nothing on the terminal. > > > > > > - > > > init:Trying to bring /bin/sh up. > > > do_execve:enter for /bin/sh > > > open_exec:enter for /bin/sh. > > > open_exec:Permission error=0. > > > dentry_open:enter. > > > dentry_open: after file->f_op->open. 0 > > > open_exec:enter for /lib/ld.so.1. > > > open_exec:Permission error=0. > > > dentry_open:enter. > > > dentry_open: after file->f_op->open. 0 > > > do_execve: after search_binary_handler. retval=0 > > > --- > > > > > > > > > I wrote a small standalone program hello.c to > > test, if > > > i can dynamically load it from init() instead of > > > loading /bin/sh. I compiled the program added it > > to my > > > ramdisk and called from init(). Following is the > > > output. You can see that, it did all of it that it > > did > > > for /bin/sh and also tried to open > > /etc/ld.so.preload > > > and /etc/ld.so.cache and failed. But again towards > > the > > > end, it tries to open /lib/libc.so.6 and succeds > > but > > > still no output is displayed that hello.c is > > trying to > > > print. > > > > > > - > > > init:Trying to bring /bin/hello up. > > > do_execve:enter for /bin/hello > > > open_exec:enter for /bin/hello. > > > open_exec:Permission error=0. > > > dentry_open:enter. > > > dentry_open: after file->f_op->open. 0 > > > open_exec:enter for /lib/ld.so.1. > > > open_exec:Permission error=0. > > > dentry_open:enter. > > > dentry_open: after file->f_op->open. 0 > > > do_execve: after search_binary_handler. retval=0 > > > sys_open:enter for /etc/ld.so.preload. > > > sys_open:unused_fd=3. > > > filp_open:enter for /etc/ld.so.preload. > > > filp_open:open_namei error -2. > > > sys_open:enter for /etc/ld.so.cache. > > > sys_open:unused_fd=3. > > > filp_open:enter for /etc/ld.so.cache. > > > filp_open:open_namei error -2. > > > sys_open:enter for /lib/libc.so.6. > > > sys_open:unused_fd=3. > > > filp_open:enter for /lib/libc.so.6. > > > filp_open:open_namei error 0. > > > dentry_open:enter. > > > dentry_open: after file->f_op->open. 0 > > > --- > > > > > > Can someone please throw some light on this? > > > > > > Thanks, > > > Prakash > > > > > > > > > > > > Daris A Nevil > > Simple Network Magic Corporation > > 401 Kentucky > > McKinney, Texas 75069 > > http://www.snmc.com > > 972-562-3751 > > > > > > > -- Mark Pilon Minolta-QMS P.O. Box 37 325 Dawson County Road 227 Fallon, MT. 59326-0037 1-406-486-5539 (primary voice line) 1-406-853-0433 (cell) ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
wrong definition in cpm_8260.h?
Joakim Tjernlund wrote: > The same error is in commproc.h for 8xx as well. It's not an error in either driver since first/last will always be set in every buffer. There isn't any reason for the driver to test or take any action based upon these indicators. These are not data transfer error or status indicators in the sense they indicate some condition on the network itself. These are buffer management indicators that should be managed as such if they are used. Thanks. -- Dan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Problem Exporting Symbol
OK, the source module was pointed out to me (arch/ppc/mm/fault.c). The problem is that Shiv is targeting the 8260, but va_to_pte() is only defined for 8xx or CONFIG_GT64260_ETH. Shiv doesn't say if he has CONFIG_GT64260_ETH defined, I'm guessing not. The code in question: #if defined(CONFIG_8xx) || defined(CONFIG_GT64260_ETH) /* The pgtable.h claims some functions generically exist, but I * can't find them.. */ pte_t *va_to_pte(unsigned long address) gvb At 07:25 PM 1/21/2003 -0500, shivkumarb at hotmail.com wrote: >Hi All, > > I am trying to write SPI driver for 8260, based on the SPI driver for >the MPC860 that I got from this list. I am trying to export the symbol >'va_to_pte', for the driver module, in the kernel. > > I included this in arch/ppc/kernel/ppc_ksyms.c, by using the >EXPORT_SYMTAB(va_to_pte) function. But when I tried to rebuild the kernel >(make zImage), it keeps coming back with the following error: > >arch/ppc/kernel/kernel.o(__ksymtab+0x0370): undefined reference to >'va_to_pte' > > I have exported other symbols, by following similar procedure. Is there >something that I am missing. > > I am running Linux kernel 2.4.2. Any inputs will be very helpful. > > Thanks and Regards, > Shiv > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
where to get ppc_8xx-gdb
Hello all, My embeded distribution for PPC 860 did not come with the cross-compiled gdb. Can someone point me to where I can get the ppc_8xx-gdb which will allow me to using GDB on the host=pc and target=power-pc-860 Thanks Cecilia Muaddi ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve('/bin/sh')
Hurray ... Daris, you suggestion seem to work. I Compiled my hello.c with -Bstatic (size was almost 1MB), added it to ramdisk and it works. So now the problem is narrowed down to, being not able to work with the shared libraries. One point is that, i DID NOT get any errors (atleast on the screen) while execve was loading shared libraries in the process of executing /bin/sh. But for sure, my problem is related to it. I really appreciate any suggestions related to this issue. Thanks, Prakash --- Daris A Nevil wrote: > > Prakash, > > Try compiling your hello.c program with the -Bstatic > flag. That will > eliminate the need to load libraries, and should > show if there are other > problems. If that works OK then we can focus on why > your library loading > is failing. > > Regards, > Daris Nevil > > > Prakash kanthi said: > > > > Hi FOlks, > > > > I finally reached the last step (hopefully) of > linux > > kernel booting process. Kernel gets stuck after > doing > > a 'execve("/bin/sh")'. I tried using different > > ramdisks, still no difference. > > > > I added few printk's in the source, following the > > function call path from 'execve' call. Following > is > > the output. You can see that, execve goes through > > without any file permission problems and > subsequently > > it tries to open /lib/ld.so.1 flawlessly. After > that i > > see nothing on the terminal. > > > > - > > init:Trying to bring /bin/sh up. > > do_execve:enter for /bin/sh > > open_exec:enter for /bin/sh. > > open_exec:Permission error=0. > > dentry_open:enter. > > dentry_open: after file->f_op->open. 0 > > open_exec:enter for /lib/ld.so.1. > > open_exec:Permission error=0. > > dentry_open:enter. > > dentry_open: after file->f_op->open. 0 > > do_execve: after search_binary_handler. retval=0 > > --- > > > > > > I wrote a small standalone program hello.c to > test, if > > i can dynamically load it from init() instead of > > loading /bin/sh. I compiled the program added it > to my > > ramdisk and called from init(). Following is the > > output. You can see that, it did all of it that it > did > > for /bin/sh and also tried to open > /etc/ld.so.preload > > and /etc/ld.so.cache and failed. But again towards > the > > end, it tries to open /lib/libc.so.6 and succeds > but > > still no output is displayed that hello.c is > trying to > > print. > > > > - > > init:Trying to bring /bin/hello up. > > do_execve:enter for /bin/hello > > open_exec:enter for /bin/hello. > > open_exec:Permission error=0. > > dentry_open:enter. > > dentry_open: after file->f_op->open. 0 > > open_exec:enter for /lib/ld.so.1. > > open_exec:Permission error=0. > > dentry_open:enter. > > dentry_open: after file->f_op->open. 0 > > do_execve: after search_binary_handler. retval=0 > > sys_open:enter for /etc/ld.so.preload. > > sys_open:unused_fd=3. > > filp_open:enter for /etc/ld.so.preload. > > filp_open:open_namei error -2. > > sys_open:enter for /etc/ld.so.cache. > > sys_open:unused_fd=3. > > filp_open:enter for /etc/ld.so.cache. > > filp_open:open_namei error -2. > > sys_open:enter for /lib/libc.so.6. > > sys_open:unused_fd=3. > > filp_open:enter for /lib/libc.so.6. > > filp_open:open_namei error 0. > > dentry_open:enter. > > dentry_open: after file->f_op->open. 0 > > --- > > > > Can someone please throw some light on this? > > > > Thanks, > > Prakash > > > > > > > Daris A Nevil > Simple Network Magic Corporation > 401 Kentucky > McKinney, Texas 75069 > http://www.snmc.com > 972-562-3751 > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Need IDE for embedded Linux project
There is much more administration to a development machine than a server. You also have to know how to restart dead tasks. Find out why you keep running out of message queues. Cope with funny sockets. I'm a device driver developer who also looks after the administration. We were a window shop in the beginning but many of us have Linux experience from home and other jobs. I would check around the developers and IT staff you might be surprised the depth of Linux knowledge you actually have. Also I cannot recommend enough the RedHat system administration courses. I went from a talented armature to a good sysadmin. I also have the sysadmin down to about 4 hours a week. Cron is your friend you will end up using it in unusual ways to solve little querks. Regards, Rod Boyce -Original Message- From: Kerl, John [mailto:[EMAIL PROTECTED] Sent: Wednesday, 22 January 2003 11:14 a.m. To: 'brian.auld at adic.com'; 'linuxppc-embedded at lists.linuxppc.org' Subject:RE: Need IDE for embedded Linux project An important issue to address (perhaps the most important, if you're new to Linux) is system administration. When you need help, are you going to call your company's IT department? Is your company willing to send you to training? Are you going to self-educate? This may be your key point, if you're new to Linux. This wasn't an issue here since I've been using Unix for many years -- so I do the administration. Maybe someone else on the list can speak to first-Linux-box-in-a-Windows-shop issues. Unix is different -- no registry, more text-based config files, etc. There are plenty of admin guides at the local bookstore; also, for what it's worth, I think Unix is easier to administer & troubleshoot than Windows, simply because you have more visiblity. strace alone is a lifesaver ... If all you want to do is cross-development, there shouldn't be too much trouble. But you -- or someone -- should know how to add a user, partition and mount disks, set up NFS exports, install sshd/telnetd, etc. Sort of "admin lite". -Original Message- From: brian.auld at adic.com [mailto:[EMAIL PROTECTED] Sent: Tuesday, January 21, 2003 1:10 PM To: linuxppc-embedded at lists.linuxppc.org Subject: Need IDE for embedded Linux project Hello all, I've just joined the mailing list. I apologize as I write this from a windows machine. I just spent some frustrating minutes trying to tweek Microsoft Outlook to go plain text 72 chars and I hope it worked. This is the unfortunate side effect of the company I work for providing a windows based dev environment for their engineers (nauseous). Also I hope I'm not too off-topic, and if I am, I humbly ask for the name of a different mailing list. My group is just kicking off their first embedded linux project. The bean counters finally bit the bullet and decided to move from VxWorks to Linux on this project. Our hardware will be a custom board running the PPC440GP. I am truly physched!!! As a first step I am trying to determine what IDE makes sense given our Department's infrastructure ... which is all windows (barf). I am aware of the following options: Option 1: All Native Linux. Option 2: Native windows cross development environment. Option 3: Vmware (linux guest on windows) with a Linux based IDE. Option 4: Linux development server <--> PC-X server software (hummingbird) on windows development hosts. Starting from a clean slate, option (1) would be my choice, but I don't think it's practical given our installed base of Win2K workstations. Regarding option (2), I get the impression that there are packages available that do this, but I haven't got any real concrete hits when searching on the web. Regarding option (3), I just got the Vmware 30-day demo and am running RH7.2 as a Virtual Machine on my Win2K pukebox to evaluate. Regarding option (4), it's another idea I had but haven't done much investigation other than having thought about the possibility. So my basic questions are as follows: Question 1: Can anyone first confirm that option (2) is doable, and second recommend or point me to some of these options. Question 2: If I go with (1) or (3), based on investigation so far, I would expect to use Denx's IDE that I've seen mentioned on this list's archives before. Some feedback about this choice or others as a Linux based IDE would be appreciated. Thanks, Brian Auld Firmware Engineer Brian Auld Embedded Software Engineer ADIC 10 Brown Rd. Ithaca, NY 14850 (607) 241-4845 ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
ppcboot for mpc8260ads board.
Hi, I got my bdi2000, thanks to the cogent arguments supplied by WD. Facing a problem. 1. Is it required to install ppcboot in an mpc8260ads board? If I make a make bzImage , can the resultant image be directly loaded to flash and start running it? the bdi2000 guys says that it is not possible. Is it so? Does it mean that we have to port ppcboot and put it in out end product for loading linux? we have only 8MB of flash. 2. I tried to coross compile ppcboot for mpc8260ads board. make mpc8260ads_config does not work. In the readme, no such target is supported. but in the boards directory, there is a file for mpc8260ads boards. I am confused and reading the makefile. But seems like a little complex. if anybody can throw some light on compiling ppcboot for mpc8260ads, it would be a great help. Thanks Om. ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Need IDE for embedded Linux project
I agree with what hs been said except for one thing use tight VNC I have found it to be faster than the original VNC. It is well supported and works very well the web site is www.tightvnc.com it is still open source. -Original Message- From: Jerry Van Baren [mailto:[EMAIL PROTECTED] Sent: Wednesday, 22 January 2003 10:08 a.m. To: linuxppc-embedded at lists.linuxppc.org Subject:Re: Need IDE for embedded Linux project Check out VNC http://www.uk.research.att.com/vnc/ -- it is an open software "PCanywhere" type program that works very well under linux and quite well under Windows. Exceed (X windows on Microsoft Windows) is a nice program but pretty expensive. gvb ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Problem with data cache on MPC823E
> > >This pheomenon is typical for bad SDRAM initialization. Read the chip >manufacturer's manual and follow the steps of the init sequence >_to_the_letter_. It is _not_ sufficient to set up the UPM tables. > >Best regards, > >Wolfgang Denk > > > But we dont use SDRAM, but DRAM EDO SIMM module. I think these RAM types dont require special initialization. Thanks, Jan Damborsky ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Problem with data cache on MPC823E
Wells, Charles wrote: >Jan, > >The following question assumes that your DRAM is physically organized as >32-bits wide and connected to CS1. > >Do you have the PS field of BR1 set to 00 (32-bit mode) or is it set to 01 >(8-bit mode)? If it's set to 8-bit mode, a burst access becomes 16 bus >cycles, which looks a lot like 4 bursts. > >Regards, >Charlie > > Yes, DRAM is organized as 32-bits wide and connected to CS2 (PS field of BR2 is set to 00 - 32 bit mode). The problem is that burst access becomes 16 bus cycles only from time to time. And this problem occurs only when accessing data cache. Instruction cache use always correct 4 bus cycle bursts. Thanks, Jan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
busybox source
Hello, in ftp://ftp.lineo.com/pub/busybox/, you can get busybox source all versions. Cheers - Original Message - From: "Prakash kanthi" <[EMAIL PROTECTED]> To: "PPC-Linux list (E-mail)" Sent: Wednesday, January 22, 2003 9:05 AM Subject: busybox source > > Hi Folks, > > Does anyone know where i can download the source for > busybox used in denx ramdisk-image? > > thanks, > Prakash > > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve('/bin/sh')
Prakash, Try compiling your hello.c program with the -Bstatic flag. That will eliminate the need to load libraries, and should show if there are other problems. If that works OK then we can focus on why your library loading is failing. Regards, Daris Nevil Prakash kanthi said: > > Hi FOlks, > > I finally reached the last step (hopefully) of linux > kernel booting process. Kernel gets stuck after doing > a 'execve("/bin/sh")'. I tried using different > ramdisks, still no difference. > > I added few printk's in the source, following the > function call path from 'execve' call. Following is > the output. You can see that, execve goes through > without any file permission problems and subsequently > it tries to open /lib/ld.so.1 flawlessly. After that i > see nothing on the terminal. > > - > init:Trying to bring /bin/sh up. > do_execve:enter for /bin/sh > open_exec:enter for /bin/sh. > open_exec:Permission error=0. > dentry_open:enter. > dentry_open: after file->f_op->open. 0 > open_exec:enter for /lib/ld.so.1. > open_exec:Permission error=0. > dentry_open:enter. > dentry_open: after file->f_op->open. 0 > do_execve: after search_binary_handler. retval=0 > --- > > > I wrote a small standalone program hello.c to test, if > i can dynamically load it from init() instead of > loading /bin/sh. I compiled the program added it to my > ramdisk and called from init(). Following is the > output. You can see that, it did all of it that it did > for /bin/sh and also tried to open /etc/ld.so.preload > and /etc/ld.so.cache and failed. But again towards the > end, it tries to open /lib/libc.so.6 and succeds but > still no output is displayed that hello.c is trying to > print. > > - > init:Trying to bring /bin/hello up. > do_execve:enter for /bin/hello > open_exec:enter for /bin/hello. > open_exec:Permission error=0. > dentry_open:enter. > dentry_open: after file->f_op->open. 0 > open_exec:enter for /lib/ld.so.1. > open_exec:Permission error=0. > dentry_open:enter. > dentry_open: after file->f_op->open. 0 > do_execve: after search_binary_handler. retval=0 > sys_open:enter for /etc/ld.so.preload. > sys_open:unused_fd=3. > filp_open:enter for /etc/ld.so.preload. > filp_open:open_namei error -2. > sys_open:enter for /etc/ld.so.cache. > sys_open:unused_fd=3. > filp_open:enter for /etc/ld.so.cache. > filp_open:open_namei error -2. > sys_open:enter for /lib/libc.so.6. > sys_open:unused_fd=3. > filp_open:enter for /lib/libc.so.6. > filp_open:open_namei error 0. > dentry_open:enter. > dentry_open: after file->f_op->open. 0 > --- > > Can someone please throw some light on this? > > Thanks, > Prakash > > Daris A Nevil Simple Network Magic Corporation 401 Kentucky McKinney, Texas 75069 http://www.snmc.com 972-562-3751 ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
busybox source
In message <20030122000548.26943.qmail at web41210.mail.yahoo.com> you wrote: > > Does anyone know where i can download the source for > busybox used in denx ramdisk-image? The ELDK documentation at http://www.denx.de/re/ELDK.html explains where you can download _all_ the ELDK sources from our CVS server. And of course you can always get busybox directly at ftp://ftp.lineo.com/pub/busybox/ Best regards, Wolfgang Denk -- Software Engineering: Embedded and Realtime Systems, Embedded Linux Phone: (+49)-8142-4596-87 Fax: (+49)-8142-4596-88 Email: wd at denx.de Drun'? 'm not drun'! You woudn' dare call m' drun' if I was sober! - Terry Pratchett, _Men at Arms_ ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Problem Exporting Symbol
I haven't checked the actual sources, but I bet 'va_to_pte' is _not_ a symbol... it is probably either an "inline" function so the compiler inserts the function's code everywhere it is "called" (saves the overhead of the call and return with associated stack pushing and popping). When you try to export it, it isn't found. Solutions: 1) Write your own non-inline "my_va_to_pte" in the kernel and export it. 2) Figure out how to avoid "va_to_pte" in your driver? 3) Don't make your SPI driver a loadable module? 3) Other? 4) Don't inline "va_to_pte" (BAD SOLUTION). gvb At 07:25 PM 1/21/2003 -0500, shivkumarb at hotmail.com wrote: >Hi All, > > I am trying to write SPI driver for 8260, based on the SPI driver for >the MPC860 that I got from this list. I am trying to export the symbol >'va_to_pte', for the driver module, in the kernel. > > I included this in arch/ppc/kernel/ppc_ksyms.c, by using the >EXPORT_SYMTAB(va_to_pte) function. But when I tried to rebuild the kernel >(make zImage), it keeps coming back with the following error: > >arch/ppc/kernel/kernel.o(__ksymtab+0x0370): undefined reference to >'va_to_pte' > > I have exported other symbols, by following similar procedure. Is there >something that I am missing. > > I am running Linux kernel 2.4.2. Any inputs will be very helpful. > > Thanks and Regards, > Shiv > > ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
Linux Kernel hangs after execve("/bin/sh")
Hi FOlks, I finally reached the last step (hopefully) of linux kernel booting process. Kernel gets stuck after doing a 'execve("/bin/sh")'. I tried using different ramdisks, still no difference. I added few printk's in the source, following the function call path from 'execve' call. Following is the output. You can see that, execve goes through without any file permission problems and subsequently it tries to open /lib/ld.so.1 flawlessly. After that i see nothing on the terminal. - init:Trying to bring /bin/sh up. do_execve:enter for /bin/sh open_exec:enter for /bin/sh. open_exec:Permission error=0. dentry_open:enter. dentry_open: after file->f_op->open. 0 open_exec:enter for /lib/ld.so.1. open_exec:Permission error=0. dentry_open:enter. dentry_open: after file->f_op->open. 0 do_execve: after search_binary_handler. retval=0 --- I wrote a small standalone program hello.c to test, if i can dynamically load it from init() instead of loading /bin/sh. I compiled the program added it to my ramdisk and called from init(). Following is the output. You can see that, it did all of it that it did for /bin/sh and also tried to open /etc/ld.so.preload and /etc/ld.so.cache and failed. But again towards the end, it tries to open /lib/libc.so.6 and succeds but still no output is displayed that hello.c is trying to print. - init:Trying to bring /bin/hello up. do_execve:enter for /bin/hello open_exec:enter for /bin/hello. open_exec:Permission error=0. dentry_open:enter. dentry_open: after file->f_op->open. 0 open_exec:enter for /lib/ld.so.1. open_exec:Permission error=0. dentry_open:enter. dentry_open: after file->f_op->open. 0 do_execve: after search_binary_handler. retval=0 sys_open:enter for /etc/ld.so.preload. sys_open:unused_fd=3. filp_open:enter for /etc/ld.so.preload. filp_open:open_namei error -2. sys_open:enter for /etc/ld.so.cache. sys_open:unused_fd=3. filp_open:enter for /etc/ld.so.cache. filp_open:open_namei error -2. sys_open:enter for /lib/libc.so.6. sys_open:unused_fd=3. filp_open:enter for /lib/libc.so.6. filp_open:open_namei error 0. dentry_open:enter. dentry_open: after file->f_op->open. 0 --- Can someone please throw some light on this? Thanks, Prakash ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/