Date: Fri, 11 Dec 2009 21:27:26 +0700 From: Yohanes Nugroho <yohanes@gmail.com> To: freebsd-arm@freebsd.org Subject: CNS11XX FreeBSD port completed Message-ID: <260bb65e0912110627o6b67b399vabaae57477b91023@mail.gmail.com>
next in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] Hi, Today I have "completed" the CNS11XX (Cavium Econa ARM formerly known as StarSemi STR9104) FreeBSD ARM Port. By "complete" I mean I have written all the drivers (usb, network, and flash). I haven't checked the correctness of the implementation. The device is a network attached storage, Emprex NSD-100, sent to me by Bruce Simpson. This is my first FreeBSD port, so I think there will be mistakes on my code. So I am requesting anyone who have time to have a look at the code, or if you have the same device, may be you can try it. You will need to solder a serial port to the device to try it. >From the usability point of view, the port is usable. I can put the FreeBSD kernel to the flash, and boot it from there with the root filesystem on a USB disk (the boot loader also supports booting from tftp, and during the development I boot using tftp). The device can enter multi user mode, and I can ssh/ftp to the device, and compile some things on the device itself. The only problem left (at least what is visible from user space) is the slow network speed (2 megabyte per second, while the linux version is around 3.8 megabyte per second). Pyun YongHyeon have helped me a lot with the network driver (I am currently asking him to have a final look at the driver). I have tried following all of his suggestions to make a better (more correct driver), and I suspect the problem is not on the network driver itself. I am suspecting that the problem is in pmap. There has been several pmap related problems lately (see http://lists.freebsd.org/pipermail/freebsd-arm/2009-October/002030.html). It is also possible that the pmap code is now fine, and I have made errors when porting the NetBSD CPU code for FA526 to FreeBSD. If I compile my current code with the latest code from HEAD, and activated WITNESS and DIAGNOSTIC i will get the error at err-1.txt, and If I activated WITNESS, DIAGNOSTIC and INVARIANT i got this the message in err-2.txt. With the same options I didn't get this error several months ago . My code is available at http://p4db.freebsd.org/depotTreeBrowser.cgi?FSPC=//depot/projects/str91xx/src/sys/arm/econa&HIDEDEL=NO plus additional code from to support FA526 CPU adapted from NetBSD: http://p4db.freebsd.org/fileLogView.cgi?FSPC=//depot/projects/str91xx/src/sys/arm/arm/cpufunc_asm_fa526.S http://p4db.freebsd.org/fileViewer.cgi?FSPC=//depot/projects/str91xx/src/sys/arm/arm/cpufunc.c Instruction to compile, and write to flash using dd to cfi0 is available at: http://tinyhack.com/2009/09/28/cnx11xxstr91xx-freebsd-progress/ http://tinyhack.com/2009/12/11/cns11xx-freebsd-port-completed/ -- Regards Yohanes http://yohan.es/ [-- Attachment #2 --] ------ start_init: trying /sbin/init lock order reversal: 1st 0xc19e00b0 pmap (pmap) @ arm/arm/pmap.c:929 2nd 0xc140a6c0 vm page queue mutex (vm page queue mutex) @ arm/arm/pmap.c:1656 KDB: stack backtrace: _end() at 0xe3e01000 scp=0xe3e01000 rlv=0xe59b0000 (0xe59b0000) rsp=0xe24cb004 rfp=0xe92dd800lock order reversal: 1st 0xc19e00b0 pmap (pmap) @ arm/arm/pmap.c:1916 2nd 0xc14a405c system map (system map) @ vm/vm_map.c:3532 KDB: stack backtrace: _end() at 0xe3e01000 scp=0xe3e01000 rlv=0xe59b0000 (0xe59b0000) rsp=0xe24cb004 rfp=0xe92dd800 vm_fault(0xc14a4000, e3e00000, 1, 0) -> 1 Fatal kernel mode data abort: 'Translation Fault (S)' trapframe: 0xc67932d4 FSR=00000005, FAR=e3e00ff4, spsr=60000013 r0 =0000001e, r1 =00000000, r2 =00000000, r3 =00000000 r4 =e3e01000, r5 =c11f21f0, r6 =ffffffff, r7 =c199c110 r8 =fffffffe, r9 =fffffff4, r10=c11f21e4, r11=c6793358 r12=00000000, ssp=c6793320, slr=c1201658, pc =c11f20c8 [thread pid 1 tid 100001 ] Stopped at db_md_set_watchpoint+0xb8: ldr r7, [r9, r4] db> ------ [-- Attachment #3 --] ugen0.1: <Cavium> at usbus0 uhub0: <Cavium OHCI root HUB, class 9/0, rev 1.00/1.00, addr 1> on usbus0 ugen1.1: <Cavium> at usbus1 uhub1: <Cavium EHCI root HUB, class 9/0, rev 2.00/1.00, addr 1> on usbus1 uhub0: 2 ports with 2 removable, self powered uhub1: 2 ports with 2 removable, self powered lock order reversal: 1st 0xc144c8d4 pmap (pmap) @ arm/arm/pmap.c:2941 2nd 0xc14de608 PV ENTRY (UMA zone) @ vm/uma_core.c:2526 KDB: stack backtrace: _end() at 0xe3e01000 scp=0xe3e01000 rlv=0xe59b0000 (0xe59b0000) rsp=0xe24cb004 rfp=0xe92dd800lock order reversal: 1st 0xc144c8d4 pmap (pmap) @ arm/arm/pmap.c:2941 2nd 0xc14dc05c system map (system map) @ vm/vm_map.c:3532 KDB: stack backtrace: _end() at 0xe3e01000 scp=0xe3e01000 rlv=0xe59b0000 (0xe59b0000) rsp=0xe24cb004 rfp=0xe92dd800 vm_fault(0xc14dc000, e3e00000, 1, 0) -> 1 Fatal kernel mode data abort: 'Translation Fault (S)' trapframe: 0xc6c19270 FSR=00000005, FAR=e3e00ff4, spsr=60000013 r0 =0000001e, r1 =00000000, r2 =00000000, r3 =00000000 r4 =e3e01000, r5 =c121a92c, r6 =c19d3110, r7 =ffffffff r8 =fffffffe, r9 =fffffff4, r10=c121a920, r11=c6c192f4 r12=00000000, ssp=c6c192bc, slr=c122a830, pc =c121a804 [thread pid 13 tid 100027 ] Stopped at db_md_set_watchpoint+0xb8:panic: _mtx_lock_sleep: recursed on non-recursive mutex pmap @ ../../../arm/arm/pmap.c:3719 Uptime: 3s Cannot dump. Device not defined or unavailable. Automatic reboot in 15 seconds - press a key on the console to abort
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?260bb65e0912110627o6b67b399vabaae57477b91023>