This one really has me stumped...
I'm trying to trace through some of the kernel code to understand why I can't access eMMC.
I put a dump_stack() call at the top of the function sdhci_add_host() (in drivers/mmc/host/sdhci.c kernel=3.11.4).
When I boot up, I get a stack trace that looks something like this...
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.11.4 #22
Backtrace:
[<80011ddc>] (dump_backtrace+0x0/0x10c) from [<80011f78>] (show_stack+0x18/0x1c)
r6:bfba82c0 r5:bfba82c0 r4:00000000 r3:00000000
[<80011f60>] (show_stack+0x0/0x1c) from [<80548398>] (dump_stack+0x80/0x9c)
[<80548318>] (dump_stack+0x0/0x9c) from [<8040d388>] (sdhci_add_host+0x24/0xbb4)
r4:bfb21d10 r3:00000000
[<8040d364>] (sdhci_add_host+0x0/0xbb4) from [<8040ef4c>] (sdhci_esdhc_imx_probe+0x28c/0x560)
[<8040ecc0>] (sdhci_esdhc_imx_probe+0x0/0x560) from [<802f8740>] (platform_drv_probe+0x20/0x24)
r8:80705e38 r7:80756c38 r6:00000000 r5:bf981e10 r4:807b2178
[<802f8720>] (platform_drv_probe+0x0/0x24) from [<802f739c>] (driver_probe_device+0x10c/0x224)
[<802f7290>] (driver_probe_device+0x0/0x224) from [<802f7548>] (__driver_attach+0x94/0x98)
r7:00000000 r6:bf981e44 r5:80756c38 r4:bf981e10
[<802f74b4>] (__driver_attach+0x0/0x98) from [<802f58a8>] (bus_for_each_dev+0x5c/0x90)
r6:802f74b4 r5:80756c38 r4:00000000 r3:bf83c85c
[<802f584c>] (bus_for_each_dev+0x0/0x90) from [<802f6e90>] (driver_attach+0x20/0x28)
r6:80746068 r5:bfb21c80 r4:80756c38
[<802f6e70>] (driver_attach+0x0/0x28) from [<802f6a08>] (bus_add_driver+0xdc/0x238)
[<802f692c>] (bus_add_driver+0x0/0x238) from [<802f7a6c>] (driver_register+0x80/0x154)
r8:80705e38 r7:806e746c r6:80762800 r5:00000006 r4:80756c38
[<802f79ec>] (driver_register+0x0/0x154) from [<802f8964>] (__platform_driver_register+0x50/0x64)
[<802f8914>] (__platform_driver_register+0x0/0x64) from [<80705e50>] (sdhci_esdhc_imx_driver_init+0x18/0x20)
[<80705e38>] (sdhci_esdhc_imx_driver_init+0x0/0x20) from [<80008798>] (do_one_initcall+0xf4/0x154)
[<800086a4>] (do_one_initcall+0x0/0x154) from [<806e7c24>] (kernel_init_freeable+0x104/0x1d0)
[<806e7b20>] (kernel_init_freeable+0x0/0x1d0) from [<805441f8>] (kernel_init+0x10/0xec)
[<805441e8>] (kernel_init+0x0/0xec) from [<8000e8b8>] (ret_from_fork+0x14/0x3c)
r4:00000000 r3:ffffffff
Referring to the stack trace, I see that the function sdhci_esdhc_imx_driver_init is called, so I tried to look at that code in the kernel...
Its not there; yet it does show up in the System.map (and several other .o) files. Then I thought maybe my browser wasn't showing
me the function, so I did a brute force find/grep on the whole tree...
First I did it on the tree that I had built and it showed me that the function name was in several .o files and also in the System.map file...
find . -type f -exec grep -l sdhci_esdhc_imx_driver_init {} \;
./drivers/built-in.o
./drivers/mmc/host/built-in.o
./drivers/mmc/host/sdhci-esdhc-imx.o
./drivers/mmc/built-in.o
./System.map
./.tmp_vmlinux1
./.tmp_System.map
./vmlinux.o
./.tmp_vmlinux2
./vmlinux
then I did it on a clean source tree and it came back empty.
Whats up with that? How can the function exist symbolically but not be in the source code somewhere?
Ed
