http://blog.gmane.org/gmane.os.netbsd.ports.arm hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3911 <pre>Dear port-arm&lt; at &gt;, I have presented the current status of my porting effort of NetBSD to the Nokia N900 smartphone at BSDCan a couple days ago. The slides can be found online there: http://people.defora.org/~khorben/papers/bsdcan2013/Call%20your%20NetBSD.pdf Pretty picture: https://twitter.com/khorben/status/335771466496868352/photo/1 The presentation was recorded (audio + slides) and should be available publicly, if not already. Short summary: - the keypad is now usable, - I am working on OMAP SPI support and a touchscreen driver, - X is running (the keymap is wrong though) - I am mainly working within the "khorben-n900" branch, - there is a short explanation in the slides on how to get it to boot. Do not hesitate to ask for details if necessary! Cheers, </pre> Pierre Pronchery 2013-05-19T20:39:06 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3910 <pre> </pre> Cabrera 2013-05-19T06:47:31 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3909 <pre>Hello, just to let you know that this snapshot : HEAD/201305161420Z is good enough to build the kernel on a Raspberry Pi now. I managed to compile/boot (see http://www.netbsd.org/docs/guide/en/chap-kernel.html#chap-kernel-installing-kernel-sources with manual compile) smoothly, just adding the UNION FS as it seems quite usefull to save SDCard writes on /var/ stuff. Thanks to the arm and Raspberry port devs ^^ Mat. NetBSD net-rbpi.local.net 6.99.19 NetBSD 6.99.19 (M_RPI) #0: Sat May 18 13:34:27 UTC 2013 mat&lt; at &gt;net-rbpi.local.net:/mnt/sd0d/usr/src/sys/arch/evbarm/compile/M_RPI evbarm </pre> Mathieu Lubrano 2013-05-18T17:21:17 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3908 <pre>Oops. Sorry for the noise. Thanks! Rich On May 14, 2013, at 12:58 PM, Tim Rightnour &lt;root&lt; at &gt;garbled.net&gt; wrote: </pre> Rich Neswold 2013-05-15T01:55:17 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3907 <pre> On 14-May-2013 Rich Neswold wrote: Err, no, thats netbsd-6 release branch. You want -current, so it's under HEAD: http://nyftp.netbsd.org/pub/NetBSD-daily/HEAD/201305121300Z/evbarm/binary/gzimg/ rpi_inst.img.gz --- Tim Rightnour &lt;root&lt; at &gt;garbled.net&gt; NetBSD: Free multi-architecture OS http://www.netbsd.org/ gnhast: NetBSD based home automation tools: http://sourceforge.net/p/gnhast/ </pre> Tim Rightnour 2013-05-14T17:58:59 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3906 <pre> On 12/05/2013, at 2:21 AM, Nick Hudson &lt;skrll&lt; at &gt;netbsd.org&gt; wrote: Does the serial console need to be disabled? Also, when I added KGDB support for the TS7200, I added the following code to arch/arm/ep93xx/epcom.c. +#ifdef KGDB + /* + * Allow kgdb to "take over" this port. If this is + * the kgdb device, it has exclusive use. + */ + if (sc-&gt;sc_iot == kgdb_sc.sc_iot &amp;&amp; + sc-&gt;sc_hwbase == kgdb_sc.sc_hwbase) { + SET(sc-&gt;sc_hwflags, COM_HW_KGDB); + printf("%s: kgdb\n", sc-&gt;sc_dev.dv_xname); + } +#endif I wrote this five years ago and I did a lot of cutting and pasting from other places, so I can't tell you what this does. Unfortunately I have to get back to work&lt; at &gt;job, otherwise I would look into this some more. Cheers, Lloyd </pre> Lloyd Parkes 2013-05-12T21:39:15 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3905 <pre> in rpi.mk, osrelease.sh is invoked with $(HOST_SH), so the bug in your build is that HOST_SH is empty. You should be using the nbmake-evbarm wrapper from the tools build. </pre> Greg Troxel 2013-05-12T10:07:56 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3904 <pre>Hi, First: many thanks for the Raspberry pi port, keep up :-) If you like to know, the plan is to have NetBSD/RaspberryPi handle a tiny mail server. About my current setup : Board: Raspberry Pi B (v2) Release: NetBSD 6.99.19 System: NetBSD net-rbpi.local.net 6.99.19 NetBSD 6.99.19 (RPI) #0: Sat May 11 12 :12:19 UTC 2013 builds&lt; at &gt;b6.netbsd.org:/home/builds/ab/HEAD/evbarm/201305110850Z-o bj/home/builds/ab/HEAD/src/sys/arch/evbarm/compile/RPI evbarm Architecture: arm Machine: evbarm Install sets form: http://nyftp.netbsd.org/pub/NetBSD-daily/HEAD/201305110850Z/evbarm/binary/sets/ Source sets from: ftp -i ftp://ftp.fr.netbsd.org/pub/NetBSD/NetBSD-6.1_RC4/source/sets/ There is a little issue here with kernel building script. Everything from install to unpacking needed sources goes perfect except a chmod +x missing on this file: /usr/src/sys/conf/osrelease.sh See, I have already unpacked, customised my kernel conf file, and moved forward to "make depend" : net-rbpi$ make depend sh: ../../../../conf/osrelease.sh: permission denied make: "../../../../arch/evbarm/conf/../../../../arch/evbarm/conf/mk.rpi" line 8: warning: " ../../../../conf/osrelease.sh" returned non-zero status ^C Fix transcript: net-rbpi$ pwd /usr/src/sys/arch/evbarm/compile/M_RPI_KERNEL net-rbpi$ ls -alh ../../../../conf/osrelease.sh -rw-r--r-- 1 ml ml 2.8K Feb 17 2012 ../../../../conf/osrelease.sh net-rbpi$ chmod +x /usr/src/sys/conf/osrelease.sh net-rbpi$ ls -alh /usr/src/sys/conf/osrelease.sh -rwxr-xr-x 1 ml ml 2.8K Feb 17 2012 /usr/src/sys/conf/osrelease.sh net-rbpi$ make depend depending the kern library objects # create kern/anddi3.d mkdep -f anddi3.d -- -I../../../../../../lib/libkern/arch/arm -march=armv6z -mtune=arm1176jzf-s -mfpu=vfp -I../../. -I../../../../. &lt;...&gt; l.d wsdisplay_util.d wsdisplay_vcons.d wsdisplay_vcons_util.d wsemul_dumb.d wsemul_vt100.d wsemul_vt100_chars.d wsemul_vt100_keys.d wsemul_vt100_subr.d wsemulconf.d wsevent.d wsfont.d wsfontdev.d wskbd.d wskbdutil.d wsmouse.d wsmux.d xc3028.d xc5k.d xform.d xform_ah.d xform_esp.d xform_ipcomp.d xform_ipip.d zlib.d assym.d" | mkdep -D net-rbpi$ echo $? 0 net-rbpi$ After this, it goes fine: net-brpi$ make &lt;... ~1h later&gt; ld -Map netbsd.map --cref -T ldscript -X -o netbsd ${SYSTEM_OBJ} ${EXTRA_OBJ} vers.o text data bss dec hex filename 3350652 960512 118628 4429792 4397e0 netbsd dbsym netbsd objcopy -S -O binary netbsd kernel.img net-rbpi$ echo $? 0 Regards, Mat. </pre> Mathieu Lubrano 2013-05-12T08:43:04 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3903 <pre>I committed a few things that seemed to get this very close to working. Not sure what the last piece of the puzzle is... Thanks, Nick </pre> Nick Hudson 2013-05-11T14:21:38 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3902 <pre> A good reference that helps you with kgdb below. http://gnats.netbsd.org/16407 -- </pre> Sujit K M 2013-05-11T11:16:28 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3901 <pre>Hi folks, On Fri, May 10, 2013 at 09:39:05PM +0200, Reinoud Zandijk wrote: .... As Nick quickly replied to me, this part needs to go since its already done in arm32_boot.c: Now it boots into DDB but i haven't managed to get it into KGDB yet. Well, it shows the colourfull graphics, so maybe its waiting now. Trying to attach from my evbarm NAS doesn't seem to work yet though. Ideas? With regards, Reinoud </pre> Reinoud Zandijk 2013-05-10T21:45:26 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3900 <pre>Hi folks, i've tried to enable KGDB on evbarm. I've managed to get it compiled and i think it *should* work, but when i try to connect to it, it just isn't responding. It might well be that that is due to the wrong architecture compiled gdb, but even on `cu' i see no communication going on. Note that the patch is also adding DDB entry on startup when the RB_KDB flag is set but neither when using the framebuffer nor when using the tty console i see any activity nor dialog. It might thus be related? The current work in progress patch is attached. Any ideas? What am i doing wrong? Is something not initialized correctly? Has anyone managed to get into DDB and have a working keyboard too? With regards, Reinoud ---------- Index: conf/RPI =================================================================== RCS file: /cvsroot/src/sys/arch/evbarm/conf/RPI,v retrieving revision 1.31 diff -u -p -r1.31 RPI --- conf/RPI17 Apr 2013 19:04:19 -00001.31 +++ conf/RPI10 May 2013 19:32:37 -0000 &lt; at &gt;&lt; at &gt; -136,10 +136,10 &lt; at &gt;&lt; at &gt; options DEBUG #options LOCKDEBUG #options IPKDB# remote kernel debugging #options VERBOSE_INIT_ARM # verbose bootstraping messages -options DDB# in-kernel debugger -optionsDDB_ONPANIC=1 -options DDB_HISTORY_SIZE=100# Enable history editing in DDB -#options KGDB +#options DDB# in-kernel debugger +#optionsDDB_ONPANIC=1 +#options DDB_HISTORY_SIZE=100# Enable history editing in DDB +options KGDB,KGDB_PLCOMUNIT=0,KGDB_DEVRATE=115200,KGDB_CONMODE=0xB00 makeoptionsDEBUG="-g"# compile full symbol table options SYMTAB_SPACE=800000 options PLCONSOLE Index: rpi/rpi_machdep.c =================================================================== RCS file: /cvsroot/src/sys/arch/evbarm/rpi/rpi_machdep.c,v retrieving revision 1.35 diff -u -p -r1.35 rpi_machdep.c --- rpi/rpi_machdep.c19 Mar 2013 22:16:55 -00001.35 +++ rpi/rpi_machdep.c10 May 2013 19:32:37 -0000 &lt; at &gt;&lt; at &gt; -139,9 +139,18 &lt; at &gt;&lt; at &gt; int plcomcnspeed = PLCONSPEED; int plcomcnmode = PLCONMODE; #endif +#include "opt_ddb.h" +#ifdef DDB +#include &lt;arm/db_machdep.h&gt; +#endif + #include "opt_kgdb.h" #ifdef KGDB +#if (NPLCOM == 0) +#error Enable plcom0 for KGDB support +#endif #include &lt;sys/kgdb.h&gt; +static void kgdb_port_init(void); #endif /* Smallest amount of RAM start.elf could give us. */ &lt; at &gt;&lt; at &gt; -521,6 +530,9 &lt; at &gt;&lt; at &gt; initarm(void *arg) cpu_reset_address = bcm2835_system_reset; +#ifdef KGDB +kgdb_port_init(); +#endif #ifdef VERBOSE_INIT_ARM printf("done.\n"); #endif &lt; at &gt;&lt; at &gt; -538,6 +550,19 &lt; at &gt;&lt; at &gt; initarm(void *arg) boothowto |= BOOTHOWTO; #endif +#ifdef KGDB +if (boothowto &amp; RB_KDB) { +kgdb_debug_init = 1; +kgdb_connect(1); +} +#endif + +#ifdef DDB +db_machine_init(); +if (boothowto &amp; RB_KDB) +Debugger(); +#endif + /* we've a specific device_register routine */ evbarm_device_register = rpi_device_register; &lt; at &gt;&lt; at &gt; -574,6 +599,37 &lt; at &gt;&lt; at &gt; consinit(void) #endif } +#ifdef KGDB +#if !defined(KGDB_PLCOMUNIT) || !defined(KGDB_DEVRATE) || !defined(KGDB_CONMODE) +#error Specify KGDB_PLCOMUNIT, KGDB_DEVRATE and KGDB_CONMODE for KGDB. +#endif + +void +static kgdb_port_init(void) +{ +static int kgdbsinit_called = 0; +static struct plcom_instance rpi_pi = { +.pi_type = PLCOM_TYPE_PL011, +.pi_flags = PLC_FLAG_32BIT_ACCESS, +.pi_iot = &amp;bcm2835_bs_tag, +.pi_size = BCM2835_UART0_SIZE +}; +int res; + +if (kgdbsinit_called != 0) +return; + +kgdbsinit_called = 1; + +rpi_pi.pi_iobase = consaddr; + +res = plcom_kgdb_attach(&amp;rpi_pi, KGDB_DEVRATE, BCM2835_UART0_CLK, + KGDB_CONMODE, KGDB_PLCOMUNIT); +if (res) +panic("KGDB uart can not be initialized."); +} +#endif + #if NGENFB &gt; 0 static bool rpi_fb_parse_mode(const char *s, uint32_t *pwidth, uint32_t *pheight) ---------- </pre> Reinoud Zandijk 2013-05-10T19:39:05 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3899 <pre>Greetings, I found the specifications to the iMAP x210 SoC a few months ago: http://aitendo3.sakura.ne.jp/aitendo_data/product_img/audio/info210/210-datasheet.pdf Since then, I've lost access to the hardware. This CPU is used in these tiny netbooks from a Spanish company called Airis (for one), which come with Android installed. It would be cool to see NetBSD run on these and make them the real computers that they are. They can be bought for 250 euro or less. Just putting this out there in case anyone has time, interest, and access to the hardware. </pre> rodent< at >NetBSD.org 2013-05-06T11:14:45 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3898 <pre>http://www.flavio-cavaller.com/emailembarrassmentjonathancook/r4.php </pre> Фомуля Гарина 2013-04-18T09:10:35 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3897 <pre> I found the issue; the fix goes in sys/arch/arm/omap/omap2430_intr.h instead of sys/arch/arm/pic/picvar.h as I originally suspected. Patch attached, FWIW. HTH, </pre> Pierre Pronchery 2013-05-03T20:06:21 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3896 <pre> Very Difficult to understand what is meant? Here is what I think is Meant. I searched for size size in the document. But my understanding is below as an example how do we know what is the size or address, moslty elf format(relocatable). This is an example of what I was finding. If it is a stack trace you could as for trace else log of dmesg etc. As far I think what this means MEMSIZE is not to be configurable, I think this will me more lines of plug and play. As far as I think of this, it is more of cross compiler/xen etc. to better start of rather than buying AT91rm9200 . Hope this is help ful. -- </pre> Sujit K M 2013-05-03T12:45:23 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3895 <pre>Hi folks, very hard to read, but maybe someone can make something of this? With regards, Reinoud ----- Forwarded message from Misha G-ka &lt;sewen7&lt; at &gt;mail.ru&gt; ----- Date: Fri, 03 May 2013 13:05:18 +0400 From: Misha G-ka &lt;sewen7&lt; at &gt;mail.ru&gt; To: port-arm-maintainer&lt; at &gt;netbsd.org Subject: at91rm9200 configuration X-Mailer: Mail.Ru Mailer 1.0 hi! I collect NetBSD for the AT91rm9200-qu processor, I read article about http://www.netbsd.org/docs/kernel/porting_netbsd_arm_soc.html, there were questions: 1 . how to learn the size size? # On-chip interrupt controler vx115_pic0 at vx115_apb? addr 0x700C1000 size 0x14c 2 . If there is options MEMSIZE=256, it if the NAND flash card is connected... ? 3 . / arch/arm/at91 this stately NetBSD for a kernel, and everything about a payment of AT91rm9200 to collect and write in source codes for work to arch/evbarm/name_board (work with the flash card, the screen, MicroSD, SDRAm. ? the keyboard, all that is necessary and there is no arch/arm/at91 ?) ----- End forwarded message ----- </pre> Reinoud Zandijk 2013-05-03T09:32:31 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3894 <pre> On Apr 29, 2013, at 3:34 PM, Radoslaw Kujawa &lt;radoslaw.kujawa&lt; at &gt;c0ff33.net&gt; wrote: It is and it should be. it should also report 512MB of RAM and a second SDMMC port. But for now, it seems we are close. The ethernet works so I'm happy to have something small for testing that I don't have to swap SD cards with. tftp is your friend. </pre> Matt Thomas 2013-04-29T22:37:06 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3893 <pre>Hi. On 30 Apr, 2013, at 0:04 , Matt Thomas &lt;matt&lt; at &gt;3am-software.com&gt; wrote: Is it really connected to external power source? If it is then shouldn't CPU run at 1GHz? </pre> Radoslaw Kujawa 2013-04-29T22:34:21 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3892 <pre>[Note this is an earm kernel and userland but that's not really relevant.] deghuboot arg = 0x9fe42fe0, 0, 0x9ffa20c5, 0x9ffa20c5 [ Kernel symbol table missing! ] Loaded initial symtab at 0x80544d70, strtab at 0x8056bd30, # entries 9963 pmap_postinit: Allocated 35 static L1 descriptor tables Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013 The NetBSD Foundation, Inc. All rights reserved. Copyright (c) 1982, 1986, 1989, 1991, 1993 The Regents of the University of California. All rights reserved. NetBSD 6.99.19 (BEAGLEBONE_INSTALL) total memory = 256 MB avail memory = 244 MB cprng kernel: WARNING insufficient entropy at creation. mainbus0 (root) cpu0 at mainbus0 core 0: 550 MHz Cortex-A8 r3p2 (Cortex V7A core) cpu0: DC enabled IC enabled WB disabled EABT branch prediction enabled cpu0: 32KB/64B 4-way L1 Instruction cache cpu0: 32KB/64B 4-way write-back-locking-C L1 Data cache cpu0: 256KB/64B 8-way write-through L2 Unified cache vfp0 at cpu0: NEON MPE (VFP 3.0+) obio0 at mainbus0 base 0x44000000-0x4fffffff: On-Board IO omapicu0 at obio0 addr 0x48200000-0x48200fff intrbase 0 prcm0 at obio0 addr 0x44e00000-0x44e01fff: Power, Reset and Clock Management sitaracm0 at obio0 addr 0x44e10000-0x44e11fff: control module, rev 1.0 gpmc0 at mainbus0 base 0x50000000: General Purpose Memory Controller, rev 6.0 gpmc0: CS#0 valid, addr 0x08000000, size 256MB com0 at obio0 addr 0x44e09000-0x44e09fff intr 72: ns16550a, working fifo com0: console sdhc0 at obio0 addr 0x48060100-0x48060fff intr 64: SDHC controller sdhc0: SD Host Specification 2.0, rev.49 sdmmc0 at sdhc0 slot 0 sdhc0: INIT Procedure timeout tiiic0 at obio0 addr 0x44e0b000-0x44e0bfff intr 70: rev 0.11 iic0 at tiiic0: I2C bus seeprom0 at iic0 addr 0x50: AT24Cxx EEPROM: size 32768 tps65217pmic0 at iic0 addr 0x24: TPS65217C Power Management Multi-Channel IC (rev 1.2) tps65217pmic0: power sources USB max 1300 mA, [AC] max 2500 mA tps65217pmic0: [LDO1: 1800 mV] [LDO2: 3300 mV] [LDO3: 1800 mV] [LDO4: 3300 mV] [DCDC1: 1500 mV] [DCDC2: 950 mV] [DCDC3: 1100 mV] omapdmtimer0 at obio0 addr 0x48040000-0x48040fff intr 68: DMTIMER2 omapdmtimer1 at obio0 addr 0x44e31000-0x44e31fff intr 67: DMTIMER1ms omapdmtimer2 at obio0 addr 0x48044000-0x48044fff intr 92: DMTIMER4 omapwdt32k0 at obio0 addr 0x44e35000-0x44e35fff: rev 0.1 cpsw0 at obio0 addr 0x4a100000-0x4a107fff intrbase 40: TI CPSW Ethernet cpsw0: Ethernet address c8:a0:30:af:16:9b ukphy0 at cpsw0 phy 0WARNING: module error: vfs load failed for `miiverbose', error 45 : OUI 0x00800f, model 0x000f, rev. 1 ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto sdmmc0: couldn't enable card: 60 cprng sysctl: WARNING insufficient entropy at creation. boot device: &lt;unknown&gt; root on md0a dumps on md0b mountroot: trying nfs... mountroot: trying msdos... mountroot: trying ffs... root file system type: ffs WARNING: no TOD clock present WARNING: using filesystem time WARNING: CHECK AND RESET THE DATE! init: copying out flags `-s' 3 init: copying out path `/sbin/init' 11 </pre> Matt Thomas 2013-04-29T22:04:16 http://permalink.gmane.org/gmane.os.netbsd.ports.arm/3891 <pre>Hi port-arm&lt; at &gt;, as some of you may have noticed already, I have made great progress on supporting the Nokia N900 (OMAP 3430). I had to change some code deep down in interrupt handling within the ARM PIC code, with potential impact on every port using it, so it is definitely not ready for being committed to the main tree. It also requires the patch that was discussed earlier for interrupt locators on the GPIO, I2C and SPI buses, yet found not desirable for integration as-is. Anyway, now, for the current progress. - lots of GPIO sensors and buttons are supported: * USB charging status (may need rework) * audio jack insertion * camera button (focus + capture) * camera cover slide * keypad slide * lock button * proximity sensor - basic keypad support (slow, no key combinations) - extra 18 GPIO pins on the TPS65950 companion chip (needs better testing, doesn't support level-triggered interrupts yet) - ST MEMS accelerometer (needs more tests and merging with hpacel) - OMAP SPI bus support (needs better testing) - TSC2005 touchscreen (to be confirmed) Support for the volume keys and the power button should be fairly trivial to add as well. So far I have implemented most of these simple sensors and buttons via sysmon_pswitch(9), as hotkeys - there will certainly be more appropriate ways to expose them to the system. The current code can be found here: https://github.com/khorben/khorben-n900 Just overwrite a recent -current source repository with the files found in the Github project above, and build a N900 evbarm kernel. My device is set to developer mode (see the Maemo wiki for the procedure) and I am booting on an SD card. It is done the same way as when booting Mer but with a NetBSD kernel (with "netbsd.gz.ub" as "uImage") and userland instead (FFS on the first partition is fine). A known issue, among others: the six times 32 plus 18 GPIO pins seem to be too many for the amount of interrupt slots pre-allocated in the ARM PIC code - which is why I disabled omapgpio5 and n900audjck0 in the kernel configuration file for the moment. I have to investigate some more there. I have made it to single-user, although multi-user should work just fine too. I'd like to run X next, and support some more hardware components. HTH! </pre> Pierre Pronchery 2013-04-29T21:40:53 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.os.netbsd.ports.arm