Kuzco
July 10, 2019, 8:34am
1
I want to start a m4 application from ocram to communicate with a7 over rpmsg.rpmsg_str_echo_freertos_example with tcm scatter file works fine and runs with linux image 2.8b6.ocram scatter file works and I can start the app.# modprobe imx_rpmsg_tty is ok.
root@colibri-imx7:~# ls /dev/ | grep RPMSG
ttyRPMSG0
I followed this tutorial.
#define m_interrupts_start 0x00000000
#define m_interrupts_size 0x00000240
#define m_text_start 0x00910240
#define m_text_size 0x00007DC0
/* m_data_size includes stack and heap */
#define m_data_start 0x20200000
#define m_data_size 0x7800
/* Sizes */
#if (defined(__stack_size__))
#define Stack_Size __stack_size__
#else
#define Stack_Size 0x0200
#endif
#if (defined(__heap_size__))
#define Heap_Size __heap_size__
#else
#define Heap_Size 0x0200
#endif
Also tried a mix scatter file with tcm and ocram:
#define m_interrupts_start 0x00000000
#define m_interrupts_size 0x00000240
#define m_text_start 0x1FFF8240
#define m_text_size 0x0000FDB0
/* m_data_size includes stack and heap */
#define m_data_start 0x20200000
#define m_data_size 0x00020000
/* Sizes */
#if (defined(__stack_size__))
#define Stack_Size __stack_size__
#else
#define Stack_Size 0x8000
#endif
#if (defined(__heap_size__))
#define Heap_Size __heap_size__
#else
#define Heap_Size 0xE000
#endif
On WinCE, application runs with all three scatter files.
Kind regards,
HI @Kuzco
Regarding your issue: What are you trying to do?
Have you done any changes to the Bsp? If yes could you share your changes.
Best regards,
Kuzco
July 15, 2019, 2:15pm
3
Hi @jaski.tx ,str_echo_freertos in ocram respectively in my mix scatter file using Arm Compilers 5.tcm memory (scatter file) with Arm Compilers 5 → works fine.ocram scatter file → no communication over rpmsg between the applications.ocram. But there is no communication between cores.
Regards,
Kuzco
July 16, 2019, 6:54am
4
Dear @jaski.tx ,TCM:
Colibri iMX7 # fatload mmc 0:1 ${loadaddr} rpmsg_tcm.elf
reading rpmsg_tcm.elf
4955108 bytes read in 234 ms (20.2 MiB/s)
Colibri iMX7 # bootaux ${loadaddr}
## Starting auxiliary core at 0x1FFF83D9 ...
Colibri iMX7 # boot
[..]
colibri-imx7 login: root
Last login: Mon Apr 1 17:34:31 UTC 2019 on ttymxc0
root@colibri-imx7:~# stty -F /dev/ttyRPMSG0 -echo
root@colibri-imx7:~# exec 3<> /dev/ttyRPMSG0
root@colibri-imx7:~# echo Test >&3
root@colibri-imx7:~# cat <&3
Test
^C
root@colibri-imx7:~# exec 3>&-
root@colibri-imx7:~#
Exactly what i expected.
OCRAM:
Colibri iMX7 # fatload mmc 0:1 ${loadaddr} rpmsg_ocram.elf
reading rpmsg_ocram.elf
4955112 bytes read in 234 ms (20.2 MiB/s)
Colibri iMX7 # bootaux ${loadaddr}
## Starting auxiliary core at 0x009103D9 ...
Colibri iMX7 # boot
colibri-imx7 login: root
Last login: Mon Apr 1 17:34:49 UTC 2019 on ttymxc0
root@colibri-imx7:~# stty -F /dev/ttyRPMSG0 -echo
root@colibri-imx7:~# exec 3<> /dev/ttyRPMSG0
root@colibri-imx7:~# echo Test >&3
[ 73.025169] INFO: rcu_sched self-detected stall on CPU
[ 73.030349] 0-...: (2099 ticks this GP) idle=e99/140000000000001/0 softirq=3597/3597 fqs=1050
[ 73.039047] (t=2100 jiffies g=757 c=756 q=149)
[ 73.043680] Task dump for CPU 0:
[ 73.046908] sh R running task 0 456 385 0x00000002
[ 73.054009] [<8010e12c>] (unwind_backtrace) from [<8010ade0>] (show_stack+0x10/0x14)
[ 73.061762] [<8010ade0>] (show_stack) from [<801b684c>] (rcu_dump_cpu_stacks+0xa8/0xc0)
[ 73.069777] [<801b684c>] (rcu_dump_cpu_stacks) from [<80177064>] (rcu_check_callbacks+0x5a4/0x81c)
[ 73.078745] [<80177064>] (rcu_check_callbacks) from [<8017a17c>] (update_process_times+0x34/0x5c)
[ 73.087624] [<8017a17c>] (update_process_times) from [<8018b30c>] (tick_sched_timer+0x48/0x8c)
[ 73.096242] [<8018b30c>] (tick_sched_timer) from [<8017b1d0>] (__hrtimer_run_queues+0x11c/0x1a8)
[ 73.105034] [<8017b1d0>] (__hrtimer_run_queues) from [<8017b414>] (hrtimer_interrupt+0xa8/0x204)
[ 73.113827] [<8017b414>] (hrtimer_interrupt) from [<805f5430>] (arch_timer_handler_phys+0x28/0x30)
[ 73.122793] [<805f5430>] (arch_timer_handler_phys) from [<8016de3c>] (handle_percpu_devid_irq+0x78/0x138)
[ 73.132366] [<8016de3c>] (handle_percpu_devid_irq) from [<8016931c>] (generic_handle_irq+0x24/0x34)
[ 73.141418] [<8016931c>] (generic_handle_irq) from [<80169820>] (__handle_domain_irq+0x5c/0xb4)
[ 73.150123] [<80169820>] (__handle_domain_irq) from [<80101474>] (gic_handle_irq+0x48/0x8c)
[ 73.158478] [<80101474>] (gic_handle_irq) from [<8010b8cc>] (__irq_svc+0x6c/0x90)
[ 73.165962] Exception stack(0x8d49bd78 to 0x8d49bdc0)
[ 73.171016] bd60: f5aa0000 00000001
[ 73.179197] bd80: 00010000 00300200 8c247f00 00010000 90060400 8c245400 8c24a180 8083235c
[ 73.187379] bda0: 8c245400 0000000a 00400000 8d49bdcc f5aa0020 8047acc0 40000013 ffffffff
[ 73.195566] [<8010b8cc>] (__irq_svc) from [<8047acc0>] (MU_SendMessage+0x28/0x5c)
[ 73.203057] [<8047acc0>] (MU_SendMessage) from [<8060d628>] (imx_rpmsg_notify+0x30/0x44)
[ 73.211154] [<8060d628>] (imx_rpmsg_notify) from [<8047c248>] (virtqueue_kick+0x2c/0x50)
[ 73.219250] [<8047c248>] (virtqueue_kick) from [<8060cc64>] (rpmsg_send_offchannel_raw+0x264/0x318)
[ 73.228300] [<8060cc64>] (rpmsg_send_offchannel_raw) from [<8060ce30>] (virtio_rpmsg_send+0x2c/0x34)
[ 73.237446] [<8060ce30>] (virtio_rpmsg_send) from [<7f020130>] (rpmsgtty_write+0x48/0x8c [imx_rpmsg_tty])
[ 73.247025] [<7f020130>] (rpmsgtty_write [imx_rpmsg_tty]) from [<8048a75c>] (do_output_char+0x160/0x1e4)
[ 73.256510] [<8048a75c>] (do_output_char) from [<8048b75c>] (n_tty_write+0x204/0x440)
[ 73.264345] [<8048b75c>] (n_tty_write) from [<80486f1c>] (tty_write+0x180/0x358)
[ 73.271749] [<80486f1c>] (tty_write) from [<801fda8c>] (__vfs_write+0x1c/0x110)
[ 73.279064] [<801fda8c>] (__vfs_write) from [<801fe90c>] (vfs_write+0xa4/0x1b0)
[ 73.286380] [<801fe90c>] (vfs_write) from [<801ff9e4>] (SyS_write+0x3c/0x90)
[ 73.293436] [<801ff9e4>] (SyS_write) from [<80107580>] (ret_fast_syscall+0x0/0x48)
[ 100.005174] NMI watchdog: BUG: soft lockup - CPU#0 stuck for 23s! [sh:456]
[ 100.012064] Modules linked in: bluetooth(O) compat(O) usb_f_rndis u_ether imx_rpmsg_tty libcomposite configfs
[ 100.022094] CPU: 0 PID: 456 Comm: sh Tainted: G O 4.9.166-2.8.6+gd899927728be #1
[ 100.030617] Hardware name: Freescale i.MX7 Dual (Device Tree)
[ 100.036365] task: 8c51f300 task.stack: 8d49a000
[ 100.040907] PC is at MU_SendMessage+0x28/0x5c
[ 100.045266] LR is at 0xf5aa0020
[ 100.048410] pc : [<8047acc0>] lr : [<f5aa0020>] psr: 40000013
[ 100.048410] sp : 8d49bdcc ip : 00400000 fp : 0000000a
[ 100.059888] r10: 8c245400 r9 : 8083235c r8 : 8c24a180
[ 100.065114] r7 : 8c245400 r6 : 90060400 r5 : 00010000 r4 : 8c247f00
[ 100.071642] r3 : 00300200 r2 : 00010000 r1 : 00000001 r0 : f5aa0000
[ 100.078173] Flags: nZcv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
[ 100.085309] Control: 10c5387d Table: 8d4ac06a DAC: 00000051
[ 100.091057] CPU: 0 PID: 456 Comm: sh Tainted: G O 4.9.166-2.8.6+gd899927728be #1
[ 100.099580] Hardware name: Freescale i.MX7 Dual (Device Tree)
[ 100.105344] [<8010e12c>] (unwind_backtrace) from [<8010ade0>] (show_stack+0x10/0x14)
[ 100.113094] [<8010ade0>] (show_stack) from [<8040dc8c>] (dump_stack+0x88/0x9c)
[ 100.120325] [<8040dc8c>] (dump_stack) from [<801a03b8>] (watchdog_timer_fn+0x260/0x2d8)
[ 100.128337] [<801a03b8>] (watchdog_timer_fn) from [<8017b1d0>] (__hrtimer_run_queues+0x11c/0x1a8)
[ 100.137216] [<8017b1d0>] (__hrtimer_run_queues) from [<8017b414>] (hrtimer_interrupt+0xa8/0x204)
[ 100.146009] [<8017b414>] (hrtimer_interrupt) from [<805f5430>] (arch_timer_handler_phys+0x28/0x30)
[ 100.154976] [<805f5430>] (arch_timer_handler_phys) from [<8016de3c>] (handle_percpu_devid_irq+0x78/0x138)
[ 100.164551] [<8016de3c>] (handle_percpu_devid_irq) from [<8016931c>] (generic_handle_irq+0x24/0x34)
[ 100.173603] [<8016931c>] (generic_handle_irq) from [<80169820>] (__handle_domain_irq+0x5c/0xb4)
[ 100.182307] [<80169820>] (__handle_domain_irq) from [<80101474>] (gic_handle_irq+0x48/0x8c)
[ 100.190663] [<80101474>] (gic_handle_irq) from [<8010b8cc>] (__irq_svc+0x6c/0x90)
[ 100.198146] Exception stack(0x8d49bd78 to 0x8d49bdc0)
[ 100.203200] bd60: f5aa0000 00000001
[ 100.211382] bd80: 00010000 00300200 8c247f00 00010000 90060400 8c245400 8c24a180 8083235c
[ 100.219565] bda0: 8c245400 0000000a 00400000 8d49bdcc f5aa0020 8047acc0 40000013 ffffffff
[ 100.227749] [<8010b8cc>] (__irq_svc) from [<8047acc0>] (MU_SendMessage+0x28/0x5c)
[ 100.235240] [<8047acc0>] (MU_SendMessage) from [<8060d628>] (imx_rpmsg_notify+0x30/0x44)
[ 100.243337] [<8060d628>] (imx_rpmsg_notify) from [<8047c248>] (virtqueue_kick+0x2c/0x50)
[ 100.251433] [<8047c248>] (virtqueue_kick) from [<8060cc64>] (rpmsg_send_offchannel_raw+0x264/0x318)
[ 100.260482] [<8060cc64>] (rpmsg_send_offchannel_raw) from [<8060ce30>] (virtio_rpmsg_send+0x2c/0x34)
[ 100.269627] [<8060ce30>] (virtio_rpmsg_send) from [<7f020130>] (rpmsgtty_write+0x48/0x8c [imx_rpmsg_tty])
[ 100.279204] [<7f020130>] (rpmsgtty_write [imx_rpmsg_tty]) from [<8048a75c>] (do_output_char+0x160/0x1e4)
[ 100.288689] [<8048a75c>] (do_output_char) from [<8048b75c>] (n_tty_write+0x204/0x440)
[ 100.296524] [<8048b75c>] (n_tty_write) from [<80486f1c>] (tty_write+0x180/0x358)
[ 100.303927] [<80486f1c>] (tty_write) from [<801fda8c>] (__vfs_write+0x1c/0x110)
[ 100.311244] [<801fda8c>] (__vfs_write) from [<801fe90c>] (vfs_write+0xa4/0x1b0)
[ 100.318560] [<801fe90c>] (vfs_write) from [<801ff9e4>] (SyS_write+0x3c/0x90)
[ 100.325616] [<801ff9e4>] (SyS_write) from [<80107580>] (ret_fast_syscall+0x0/0x48)
Don’t know what all the messages mean.
Hope this helps.
Dear @Kuzco
Linux seems to somewhat use two OCRAM ranges:
4kB at the start of OCRAM0x2020_0000 - 0x2020_0FFF / 0x0090_0000 - 0x0090_0FFF
60kB at offset 0x10000x2020_1000 - 0x2020_FFFF / 0x0090_1000 - 0x0090_FFFF
I verified that the first 4kB get overwritten during Linux boot, the other 60kB seem to be left untouched.
Can you try to modify your scatter file to not use the first 4kB?
Regards, Andy
Kuzco
August 1, 2019, 1:26pm
6
Dear @andy.tx ,str_echo_freertos example and my own M4 application are running with ocram and my mix scatter file.
Kind regards,
