Active Topics
-
Is there a section to talk about Java ME? Apps, etc. (7)
to General by teroyk - 1 hr, 30 mins ago - more...
| The Following User Says Thank You to R-R For This Useful Post: | ||
|
2010-01-26
, 20:09
|
|
Posts: 86 |
Thanked: 362 times |
Joined on Dec 2007
@ Paris / France
|
#102
|
Originally Posted by eber42
My kernel doesn't register anything in /dev/mtd2 (i still see the same very old dump), I can see the oopses flying by on the framebuffer, but nothing is recorded.
'll try to reproduce the problem tonight and post some oops log.
Syslog can't record anything too (even with panic_on_oops to 0)
Is there anything to do to have kernel log its oopses ?
Originally Posted by R-R
The article has been updated and talks about a 3 weeks delay, we can still beat them, hurry up guys !
In the mean time ...
http://thenokiablog.com/2010/01/26/s...ot-nokia-n900/
PS: btw, i won't buy a high speed camera to record oopses.
|
|
2010-01-26
, 21:07
|
|
Posts: 540 |
Thanked: 288 times |
Joined on Sep 2009
|
#103
|
Originally Posted by eber42
Which reminds me, my camera can do 420FPS (even 1000FPS but the resolution is *really* small then).
PS: btw, i won't buy a high speed camera to record oopses.
But that's beside the point, I don't have time for playing around with custom kernels before the weekend.
|
|
2010-01-26
, 22:05
|
|
Posts: 173 |
Thanked: 160 times |
Joined on Jan 2010
@ London, UK
|
#104
|
Originally Posted by qole
I'm 90% certain you can override system calls with a module, rootkits do it. Furhermore even a plain replaced module doesn't need to be called the same filename so deb dependancies need not be much of an issue and it could just be that the nat starting app just does a rmmod;insmod. Ugly but trivial for an end user.
So you're saying we could have a kernel module that we could insert that would patch the kernel and add the missing functionality? That would be really nice. It wouldn't be so nice if you have to disable or replace stock modules, though, because, yeah, that leads back to the problems of flashing a new kernel.
But this is all academic until someone figures out why the device reboots under load.
Do you get panics for less hetfy non http traffic. Almost every device under the sun can use an http proxy, so use one and use nat just for im/ssh/voip etc?
Damion
|
|
2010-01-26
, 22:17
|
|
Posts: 1,224 |
Thanked: 1,763 times |
Joined on Jul 2007
|
#105
|
Enabling NAT changes the size (and layout) of some structures, so while it is theoretically possible for a module to patch every function in the kernel to work with the new structure, it is not going to happen. For kernel space NAT you need a different kernel.
__________________
My repository
"N900 community support for the MeeGo-Harmattan" Is the new "Mer is Fremantle for N810".
No more Nokia devices for me.
My repository
"N900 community support for the MeeGo-Harmattan" Is the new "Mer is Fremantle for N810".
No more Nokia devices for me.
| The Following 2 Users Say Thank You to Matan For This Useful Post: | ||
|
|
2010-01-27
, 20:59
|
|
Posts: 86 |
Thanked: 362 times |
Joined on Dec 2007
@ Paris / France
|
#106
|
I had some issue with my CMDLINE (i should have read mtdoops.c). So i'm back in business, and you can read my oops dump below (it's also consistent with the one i had with Jebba's kernel).
So what's next ?
I'd be grateful if someone gave a confirmation of the issue (maybe it only happens on my n900 because i badly hacked the device or because I used duck fat to grease the slider). I gave an easy way to reproduce this in a post above, please try it.
Now, how can we debug this oops ? (i'm not a kernel hacker)
If kernel NAT is not possible, we can look at user-space solutions (as damion said) :
- transparent proxy could handle all outgoing TCP connections (does this work with standard kernel?) and N900 could act as a DNS proxy with dnsmasq
- maybe some tunneling solution (but it would require a server on the internet to do the NAT thing)
But for now, i still think NAT is the best solution.
btw, I can provide with my config/kernel/modules if someone is interested.
Ooops dump follows ...
So what's next ?
I'd be grateful if someone gave a confirmation of the issue (maybe it only happens on my n900 because i badly hacked the device or because I used duck fat to grease the slider). I gave an easy way to reproduce this in a post above, please try it.
Now, how can we debug this oops ? (i'm not a kernel hacker)
If kernel NAT is not possible, we can look at user-space solutions (as damion said) :
- transparent proxy could handle all outgoing TCP connections (does this work with standard kernel?) and N900 could act as a DNS proxy with dnsmasq
- maybe some tunneling solution (but it would require a server on the internet to do the NAT thing)
But for now, i still think NAT is the best solution.
btw, I can provide with my config/kernel/modules if someone is interested.
Ooops dump follows ...
Code:
Log Entry 9 (at position 8) [90575.336761] Unable to handle kernel NULL pointer dereference at virtual address 000002e4 [90575.336791] pgd = c0004000 [90575.336822] [000002e4] *pgd=00000000 [90575.336822] Internal error: Oops: 17 [#1] PREEMPT [90575.336853] Modules linked in: ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables vfat fat sd_mod scsi_mod iphb rfcomm sco bridge stp llc bnep l2cap unionfs ext3 jbd omaplfb pvrsrvkm bridgedriver g_file_storage uinput board_rx51_camera omap_previewer_hack omap34xxcam_mod isp_mod iovmm videobuf_dma_sg videobuf_core omap3_iommu iommu2 iommu dspbridge ssi_mcsaab_imp cmt_speech phonet smc91x mii wl12xx mmc_block omap_wdt omap_ssi mac80211 crc7 tsc2005 omap_hsmmc nokia_av hci_h4p mmc_core bluetooth fmtx_si4713 et8ek8 ad5820 lis302dl videodev v4l1_compat compat_ioctl32 leds_lp5523 adp1653 tsl2563 smia_sensor smiaregs v4l2_int_device rtc_twl4030 rtc_core twl4030_wdt leds_twl4030_vibra led_class fuse [90575.337188] CPU: 0 Not tainted (2.6.28-omap1 #3) [90575.337249] PC is at net_out_hook+0xdc/0x198 [iphb] [90575.337280] LR is at 0x0 [90575.337280] pc : [<bf2d6404>] lr : [<00000000>] psr: 60000113 [90575.337310] sp : cf091b98 ip : 0000001b fp : cf091bac [90575.337310] r10: 80000000 r9 : c023ed5c r8 : c03bd310 [90575.337341] r7 : cc93f300 r6 : 00000004 r5 : cc93f300 r4 : cf091bfc [90575.337341] r3 : 00404c70 r2 : 4c004c30 r1 : 00000000 r0 : c3ced650 [90575.337371] Flags: nZCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment kernel [90575.337402] Control: 10c5387d Table: 835e8018 DAC: 00000017 [90575.337402] Process omap2_mcspi (pid: 10, stack limit = 0xcf0902e0) [90575.337432] Stack: (0xcf091b98 to 0xcf092000) [90575.337432] 1b80: cf091bfc 00000000 [90575.337493] 1ba0: cf091be4 cf091bb0 c0233f94 bf2d6334 c023ed5c cf091bc0 bf3561f8 cc93f300 [90575.337524] 1bc0: 00000004 c815bf00 c03bd310 00000000 c023ed5c 00000002 cf091c2c cf091be8 [90575.337554] 1be0: c02342cc c0233f3c cf1bb000 cf091bfc c023ed5c 80000000 00000000 bf2d6e1c [90575.337615] 1c00: cf091c3c cc93f300 cf1bb000 c815bf00 cc93f324 cc93f300 cf090000 00000000 [90575.337646] 1c20: cf091c54 cf091c30 c023f3c4 c023425c cf1bb000 c023ed5c 80000000 c02186e8 [90575.337707] 1c40: cc93f300 c3ced650 cf091c6c cf091c58 c023be1c c023f30c cc93f300 c3ced650 [90575.337738] 1c60: cf091c9c cf091c70 c023c148 c023bde4 cf01dd40 c3cede30 cc93f300 c03bcc3c [90575.337799] 1c80: c3cede30 cc93f300 c03bcc3c cf1bb000 cf091ce4 cf091ca0 c023a844 c023be30 [90575.337829] 1ca0: cf1bb000 cf091cb4 c023a544 80000000 c005274c c03bd2f0 c00515dc cc93f300 [90575.337860] 1cc0: cc93f300 cc93f300 cc93f300 c03bcc3c cf1bb000 c03bcf20 cf091d14 cf091ce8 [90575.337921] 1ce0: c023ae88 c023a550 00000000 c023a544 80000000 a0000193 cc93f300 c03bcf08 [90575.337951] 1d00: cf1bb000 00000008 cf091d44 cf091d18 c021ec20 c023ac3c c00515dc c004f324 [90575.338012] 1d20: c03bce64 00000001 000099be 00000040 0000012c 000099be cf091d64 cf091d48 [90575.338043] 1d40: c0221560 c021e938 c03bce64 00000040 c0392dc0 00000001 cf091d8c cf091d68 [90575.338073] 1d60: c0220e8c c02214f0 00000104 cf090000 c0392dc0 00000001 0000000c 0000000a [90575.338134] 1d80: cf091dbc cf091d90 c005c0dc c0220e48 c036d020 0000000c 0000000c 00000000 [90575.338165] 1da0: cf090000 00000003 00000002 cf01deb4 cf091dd4 cf091dc0 c005c4e4 c005c084 [90575.338226] 1dc0: 0000000c 00000000 cf091dec cf091dd8 c002c088 c005c4a4 ffffffff d8200000 [90575.338256] 1de0: cf091e64 cf091df0 c028d204 c002c00c 00000001 00000000 0009dde9 00000002 [90575.338287] 1e00: c0292928 cf01dd40 cf090000 7fffffff c035d8c0 c028b5a4 cf01deb4 cf091e64 [90575.338348] 1e20: c038d880 cf091e38 c028ade4 c028ae2c 00000013 ffffffff 00000000 cf091ec8 [90575.338378] 1e40: cf01dd40 cf090000 7fffffff cf08b31c 00000008 00000002 cf091eb4 cf091e68 [90575.338439] 1e60: c028b5a4 c028ac70 cf091e8c cf091e78 c028f3d4 c028f360 00000000 cf01dd40 [90575.338470] 1e80: cf091ebc cf091e90 c028af18 c0050880 40000000 cf090000 cf091ec8 cf01dd40 [90575.338500] 1ea0: cf090 Log Entry 10 (at position 9) <7>mtdoops: Ready 9, 10 (no erase) [90575.339202] 7fffffff cf091ef4 cf091eb8 c028b374 c028b590 c028b0e0 00000001 [90575.339263] 1ec0: cf01dd40 c00527a8 cf08b320 cf08b320 00000000 00000001 cf08a600 ceac5ea4 [90575.339294] 1ee0: cf08b300 00000002 cf091f04 cf091ef8 c028b4bc c028b28c cf091f7c cf091f08 [90575.339355] 1f00: c01c9978 c028b4b0 00000000 00000000 00000001 c0049fdc cf090000 cf08a540 [90575.339385] 1f20: ceac5eec 00000001 cf08b4c0 00000000 00000000 00000002 480ba000 00000000 [90575.339416] 1f40: ceaf3118 d80ba03c d80ba038 00000008 cf091f74 cf08a544 cf090000 cf0876c0 [90575.339477] 1f60: cf08a540 c01c95e0 00000000 00000000 cf091fa4 cf091f80 c0068dfc c01c95ec [90575.339508] 1f80: cf091fb8 cf01dd40 cf090000 cf0876c0 00000000 00000000 cf091fdc cf091fa8 [90575.339538] 1fa0: c0069b9c c0068d34 cf0876c0 00000000 cf01dd40 c006d1a4 cf091fb8 cf091fb8 [90575.339599] 1fc0: cf0876c0 c0069aac 00000000 00000000 cf091ff4 cf091fe0 c006ce14 c0069ab8 [90575.339630] 1fe0: 00000000 00000000 00000000 cf091ff8 c0059d6c c006cdcc ffdffeef bdffefff [90575.339691] Backtrace: [90575.339691] [<bf2d6328>] (net_out_hook+0x0/0x198 [iphb]) from [<c0233f94>] (nf_iterate+0x64/0xac) [90575.339752] r5:00000000 r4:cf091bfc [90575.339782] [<c0233f30>] (nf_iterate+0x0/0xac) from [<c02342cc>] (nf_hook_slow+0x7c/0x12c) [90575.339813] [<c0234250>] (nf_hook_slow+0x0/0x12c) from [<c023f3c4>] (ip_output+0xc4/0xe8) [90575.339843] [<c023f300>] (ip_output+0x0/0xe8) from [<c023be1c>] (ip_forward_finish+0x44/0x4c) [90575.339904] r5:c3ced650 r4:cc93f300 [90575.339904] [<c023bdd8>] (ip_forward_finish+0x0/0x4c) from [<c023c148>] (ip_forward+0x324/0x3a0) [90575.339935] r5:c3ced650 r4:cc93f300 [90575.339965] [<c023be24>] (ip_forward+0x0/0x3a0) from [<c023a844>] (ip_rcv_finish+0x300/0x320) [90575.339996] r7:cf1bb000 r6:c03bcc3c r5:cc93f300 r4:c3cede30 [90575.340026] [<c023a544>] (ip_rcv_finish+0x0/0x320) from [<c023ae88>] (ip_rcv+0x258/0x290) [90575.340057] r8:c03bcf20 r7:cf1bb000 r6:c03bcc3c r5:cc93f300 r4:cc93f300 [90575.340087] [<c023ac30>] (ip_rcv+0x0/0x290) from [<c021ec20>] (netif_receive_skb+0x2f4/0x344) [90575.340148] r7:00000008 r6:cf1bb000 r5:c03bcf08 r4:cc93f300 [90575.340179] [<c021e92c>] (netif_receive_skb+0x0/0x344) from [<c0221560>] (process_backlog+0x7c/0x110) [90575.340209] [<c02214e4>] (process_backlog+0x0/0x110) from [<c0220e8c>] (net_rx_action+0x50/0x19c) [90575.340240] r7:00000001 r6:c0392dc0 r5:00000040 r4:c03bce64 [90575.340270] [<c0220e3c>] (net_rx_action+0x0/0x19c) from [<c005c0dc>] (__do_softirq+0x64/0xf4) [90575.340332] [<c005c078>] (__do_softirq+0x0/0xf4) from [<c005c4e4>] (irq_exit+0x4c/0xa8) [90575.340362] [<c005c498>] (irq_exit+0x0/0xa8) from [<c002c088>] (__exception_text_start+0x88/0xa8) [90575.340393] r5:00000000 r4:0000000c [90575.340423] [<c002c000>] (__exception_text_start+0x0/0xa8) from [<c028d204>] (__irq_svc+0x44/0xa4) [90575.340454] Exception stack(0xcf091df0 to 0xcf091e38) [90575.340484] 1de0: 00000001 00000000 0009dde9 00000002 [90575.340515] 1e00: c0292928 cf01dd40 cf090000 7fffffff c035d8c0 c028b5a4 cf01deb4 cf091e64 [90575.340545] 1e20: c038d880 cf091e38 c028ade4 c028ae2c 00000013 ffffffff [90575.340606] r5:d8200000 r4:ffffffff [90575.340606] [<c028ac64>] (schedule+0x0/0x328) from [<c028b5a4>] (schedule_timeout+0x20/0xb8) [90575.340667] [<c028b584>] (schedule_timeout+0x0/0xb8) from [<c028b374>] (wait_for_common+0xf4/0x1ac) [90575.340698] r7:7fffffff r6:cf090000 r5:cf01dd40 r4:cf091ec8 [90575.340728] [<c028b280>] (wait_for_common+0x0/0x1ac) from [<c028b4bc>] (wait_for_completion+0x18/0x1c) [90575.340789] [<c028b4a4>] (wait_for_completion+0x0/0x1c) from [<c01c9978>] (omap2_mcspi_work+0x398/0x728) [90575.340820] [<c01c95e0>] (omap2_mcspi_work+0x0/0x728) from [<c0068dfc>] (run_workqueue+0xd4/0x198) [90575.340881] [<c0068d28>] (run_workqueue+0x0/0x198) from [<c0069b9c>] (worker_thread+0xf0/0x104) [90575.340911] r9:00000000 r8:00000000 r7:cf0876c0 r6:cf090000 r5:cf01dd40 [90575.340942] r4:cf091fb8 [90575.340972] [<c0069aac>] (worker_thread+0x0/0x1 Log Entry 11 (at position 10) <7>mtdoops: Ready 10, 11 (no erase) [90575.341644] from [<c006ce14>] (kthread+0x54/0x80) [90575.341674] r7:00000000 r6:00000000 r5:c0069aac r4:cf0876c0 [90575.341705] [<c006cdc0>] (kthread+0x0/0x80) from [<c0059d6c>] (do_exit+0x0/0x7b4) [90575.341735] r5:00000000 r4:00000000 [90575.341766] Code: 1a00000a e5903004 e0232863 e3c228ff (e59e12e4)
|
|
2010-01-27
, 21:07
|
|
Posts: 1,224 |
Thanked: 1,763 times |
Joined on Jul 2007
|
#107
|
Just to make sure: did you replace all modules, together with the kernel?
__________________
My repository
"N900 community support for the MeeGo-Harmattan" Is the new "Mer is Fremantle for N810".
No more Nokia devices for me.
My repository
"N900 community support for the MeeGo-Harmattan" Is the new "Mer is Fremantle for N810".
No more Nokia devices for me.
|
|
2010-01-27
, 21:31
|
|
Posts: 86 |
Thanked: 362 times |
Joined on Dec 2007
@ Paris / France
|
#108
|
Originally Posted by Matan
I usually do so, but for CMDLINE change, I just built & flashed the kernel. I can do it again with redeploying all modules (but it'll be tomorrow or later).
Just to make sure: did you replace all modules, together with the kernel?
|
|
2010-01-27
, 21:56
|
|
Posts: 61 |
Thanked: 60 times |
Joined on Jul 2007
@ Erlangen
|
#109
|
Originally Posted by rambo
Thanks for the great summary.
Please do.
To summarize:
1. N900 stock kernel does not have module support for nf_conntrack et co so just building the modules is not going to help, one needs to flash a kernel that has support for inserting said modules.
About the module thing, I only can contribute what I did on the n800 when I sat that up. The modules I built and loaded were these:
ip_conntrack_amanda.ko
ip_conntrack_ftp.ko
ip_conntrack_h323.ko
ip_conntrack_irc.ko
ip_conntrack.ko
ip_conntrack_netbios_ns.ko
ip_conntrack_pptp.ko
ip_conntrack_sip.ko
ip_conntrack_tftp.ko
ip_nat_amanda.ko
ip_nat_ftp.ko
ip_nat_h323.ko
ip_nat_irc.ko
ip_nat.ko
ip_nat_pptp.ko
ip_nat_sip.ko
ip_nat_snmp_basic.ko
ip_nat_tftp.ko
ip_queue.ko
iptable_nat.ko
ipt_MASQUERADE.ko
I have no possibility to check right now but I think that some was build into the kernel in the original config. I hope to find time to try that out over the weekend with the n900.
For the reboot issue, could it have maybe have something to do with the buffers set for nat and how overflows are handled. I just remembered that having many connections in parallel sometimes caused some buffer to fillup and once it was full the effects were some severe hangs on the router. If I recall correctly increasing the buffer size via /proc solved the problems. Just some idea
__________________
things we learned from movies
38) No matter how badly a spaceship is attacked, its internal gravity system is never damaged.
things we learned from movies
38) No matter how badly a spaceship is attacked, its internal gravity system is never damaged.
Last edited by greygoo; 2010-01-27 at 21:58.
|
|
2010-01-27
, 23:19
|
|
Posts: 107 |
Thanked: 74 times |
Joined on Jan 2010
|
#110
|
I continue to have no problem with "normal" usage, tryied the 20 firefox tab reloading, it may be that UMTS is slower so the problem does not came out, I did not have the time to test NAT with wifi at higher bitrate.
Hope I will do during the weekend.
Note: I'm using usb to share connection to my laptop, not ad-hoc networking.
Going a bit OT, is there a way to flash the kernel with mtd-utils directly from the device? I'd like to swap kernels even if I do not have my laptop with me.
I hope this is possible so we may workaround the missing fiasco-gen closed binary, his resolution seems to be set for the next milestone, what's the problem to add the binary to the sdk feed just *now*?
Hope I will do during the weekend.
Note: I'm using usb to share connection to my laptop, not ad-hoc networking.
Going a bit OT, is there a way to flash the kernel with mtd-utils directly from the device? I'd like to swap kernels even if I do not have my laptop with me.
I hope this is possible so we may workaround the missing fiasco-gen closed binary, his resolution seems to be set for the next milestone, what's the problem to add the binary to the sdk feed just *now*?
| The Following User Says Thank You to nicola.mfb For This Useful Post: | ||
http://thenokiablog.com/2010/01/26/s...ot-nokia-n900/
Please vote for the following Bugs: