uartlite with 2.6.17 kernel and kernel early text messages

uartlite with 2.6.17 kernel and kernel early text messages

[Robert Corley]

I am still trying to get the UARTLITE to work with 2.6 and a plb_temac desi=
gn.  I am using EDK 7.1.  I have generated the edk files and copied xparame=
ters_ml300.h to arch/ppc/platforms/4xx/xparameters.xparameters_ml403.h=0A=
=0AIn an effort to get past the "Rebooting to System ACE Configuration Addr=
ess 6..." message, I have selected "support for early boot texts over seria=
l port" in kernel debugging.=0A=0AFYI, I'm using Peter's patches to create =
the uartlite.c and associated files and have selected the uartlite using ma=
ke menuconfig.=0A=0AThe first error is as follows:=0Aathena startup_network=
 # make ARCH=3Dppc zImage.initrd=0A  CHK     include/linux/version.h=0A  CH=
K     include/linux/compile.h=0Adnsdomainname: Unknown host=0A  CC      arc=
h/ppc/syslib/gen550_dbg.o=0Aarch/ppc/syslib/gen550_dbg.c:36: error: `RS_TAB=
LE_SIZE' undeclared here (not in a function)=0Aarch/ppc/syslib/gen550_dbg.c=
:38: error: empty scalar initializer=0Aarch/ppc/syslib/gen550_dbg.c:38: err=
or: (near initialization for `rs_table')=0Aarch/ppc/syslib/gen550_dbg.c:36:=
 error: storage size of `rs_table' isn't known=0Aarch/ppc/syslib/gen550_dbg=
.c:36: warning: 'rs_table' defined but not used=0Amake[1]: *** [arch/ppc/sy=
slib/gen550_dbg.o] Error 1=0Amake: *** [arch/ppc/syslib] Error 2=0A=0Aso, I=
 modified the gen550_dbg.c file to #include the xparameters.h, where the RS=
_TABLE_SIZE is defined=0A but still get more errors.  Here they are:=0A=0A =
 CHK     include/linux/version.h=0A  CHK     include/linux/compile.h=0Adnsd=
omainname: Unknown host=0A  CC      arch/ppc/syslib/gen550_dbg.o=0Aarch/ppc=
/syslib/gen550_dbg.c:37: error: `XPAR_UARTNS550_0_CLOCK_FREQ_HZ' undeclared=
 here (not in a function)=0Aarch/ppc/syslib/gen550_dbg.c:37: error: initial=
izer element is not constant=0Aarch/ppc/syslib/gen550_dbg.c:37: error: (nea=
r initialization for `rs_table[0].baud_base')=0Aarch/ppc/syslib/gen550_dbg.=
c:37: error: `XPAR_INTC_0_UARTNS550_0_VEC_ID' undeclared here (not in a fun=
ction)=0Aarch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not=
 constant=0Aarch/ppc/syslib/gen550_dbg.c:37: error: (near initialization fo=
r `rs_table[0].irq')=0Aarch/ppc/syslib/gen550_dbg.c:37: error: `XPAR_UARTNS=
550_0_BASEADDR' undeclared here (not in a function)=0Aarch/ppc/syslib/gen55=
0_dbg.c:37: error: initializer element is not constant=0Aarch/ppc/syslib/ge=
n550_dbg.c:37: error: (near initialization for `rs_table[0].iomem_base')=0A=
arch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not constant=
=0Aarch/ppc/syslib/gen550_dbg.c:37: error: (near initialization for `rs_tab=
le[0]')=0Amake[1]: *** [arch/ppc/syslib/gen550_dbg.o] Error 1=0Amake: *** [=
arch/ppc/syslib] Error 2=0A=0AQuestions:=0A=0A1.    Now, is this an issue w=
ith the UARTLITE driver or is it just not supported for early messaging?=0A=
2.   What am I missing w.r.t. getting something out of the serial port?=0A3=
.    I am assuming that the boot args for a initrd boot are:  "console=3Dtt=
yUL0 ip=3Doff root=3D/dev/ram rw", correct?=0A=0A-corley=0A=0A

Eeek! page_mapcount(page) went negative! (-1) -2.6.16.2 kernel

[agnel juni]

[-- Attachment #1: Type: text/plain, Size: 3354 bytes --]

Hello all,
 
 We are working on System ACE driver ported for AMCC44x based custom platform, on linux kernel-2.6.16.2
 
 We have formatted the flash with FAT16 filesystem as System ACE supports only that.
 
 We are able to run all the filessytem commands except that 'mount' shows inconsisten behaviour.
 
 I have turned off the "Use full SLAB allocator" option under kernel configuration.
 
 With this, when mount fails I see Kernel throwing Oops sometimes.
 
 Any suggestions to debug this are greatly appreciated.
 
 ################# Kernel Oops #######################
 
 
 # mount -t msdos /dev/xsysace /mnt/
 Bad pte = df452f3cdf452f3c, process = exe, vm_flags = 75, vaddr = fe11a64
 Call Trace:
 [DFF97DE8] [C000958C] show_stack+0x48/0x194 (unreliable)
 [DFF97E18] [C00470CC] print_bad_pte+0x44/0x54
 [DFF97E20] [C0049694] __handle_mm_fault+0x900/0x930
 [DFF97E78] [C000A174] do_page_fault+0x268/0x4fc
 [DFF97F40] [C0001FE0] handle_page_fault+0xc/0x80
 VM: killing process exe
 swap_free: Bad swap file entry 62696e2f
 swap_free: Bad swap file entry 026f7800
 swap_free: Bad swap file entry 30c070c3
 swap_free: Bad swap file entry b8df44b2
 swap_free: Bad swap file entry c0df44b2
 Bad pte = 5800000001, process = ???, vm_flags = 75, vaddr = fe12000
 Call Trace:
 [DFF97D28] [C000958C] show_stack+0x48/0x194 (unreliable)
 [DFF97D58] [C00470CC] print_bad_pte+0x44/0x54
 [DFF97D60] [C0047168] vm_normal_page+0x8c/0x94
 [DFF97D68] [C004794C] unmap_vmas+0x424/0x5c8
 [DFF97DF8] [C004CAC8] exit_mmap+0x60/0xec
 [DFF97E18] [C0016C3C] mmput+0x50/0xd4
 [DFF97E28] [C001B228] exit_mm+0x120/0x1a0
 [DFF97E40] [C001B864] do_exit+0x11c/0x77c
 [DFF97E78] [C000A2A4] do_page_fault+0x398/0x4fc
 [DFF97F40] [C0001FE0] handle_page_fault+0xc/0x80
 Eeek! page_mapcount(page) went negative! (-1)
   page->flags = 400
   page->count = 1
   page->mapping = 00000000
 kernel BUG in page_remove_rmap at mm/rmap.c:560!
 Oops: Exception in kernel mode, sig: 5 [#1]
 NIP: C004E620 LR: C004E698 CTR: 00000000
 REGS: dff97ca8 TRAP: 0700   Not tainted  (2.6.16.2)
 MSR: 00029000 <EE,ME>  CR: 24022022  XER: 00000000
 TASK = df463000[610] 'exe' THREAD: dff96000
 GPR00: C004E698 DFF97D58 DF463000 0000001E 00000FBC FFFFFFFF C0240000 C01F0C0C
 GPR08: C01F0000 00000001 00000000 C0240000 00000000 10065ABC DFC46D10 0FF08FFF
 GPR16: C0243C18 00000000 FFFFFFFF FFFFFFFF 00000000 0FF09000 DFD2AA50 DF4821FC
 GPR24: 00000000 DFF97D74 A1FF0000 00000001 0FF09000 C0259000 0FE13000 C0259000
 NIP [C004E620] page_remove_rmap+0x68/0xe4
 LR [C004E698] page_remove_rmap+0xe0/0xe4
 Call Trace:
 [DFF97D58] [C004E698] page_remove_rmap+0xe0/0xe4 (unreliable)
 [DFF97D68] [C004799C] unmap_vmas+0x474/0x5c8
 [DFF97DF8] [C004CAC8] exit_mmap+0x60/0xec
 [DFF97E18] [C0016C3C] mmput+0x50/0xd4
 [DFF97E28] [C001B228] exit_mm+0x120/0x1a0
 [DFF97E40] [C001B864] do_exit+0x11c/0x77c
 [DFF97E78] [C000A2A4] do_page_fault+0x398/0x4fc
 [DFF97F40] [C0001FE0] handle_page_fault+0xc/0x80
 Instruction dump:
 83e1000c 38210010 7c0803a6 4e800020 801f0008 3c60c01c 3863d5e0 35200001
 41800034 813f0008 39290001 55290ffe <0f090000> 38600010 3880ffff 4bff07b1
 Fixing recursive fault but reboot is needed!
 
 
Thanks
 Junita

 				
---------------------------------
 Find out what India is talking about on  - Yahoo! Answers India 
 Send FREE SMS to your friend's mobile from Yahoo! Messenger Version 8. Get it NOW

[-- Attachment #2: Type: text/html, Size: 3833 bytes --]

Re: uartlite with 2.6.17 kernel and kernel early text messages

[Michael Galassi]

>In an effort to get past the "Rebooting to System ACE Configuration
>Address 6..." message, I have selected "support for early boot texts
>over serial port" in kernel debugging.

This will not work.  Unfortunately linux here shows its roots as PS OS,
there is a vast amount of hardcoded NS8250/NS16550 knowledge in there.
I'd started modularizing this so we could use a UART lite but the more I
dug the lower my motivation.  This is really ugly code in an area that
is of virtually no interest.  For now get rid of support for early boot
texts over serial port" and use your JTAG to read the buffers if you
panic before the console is up and running.

-michael

Re: uartlite with 2.6.17 kernel and kernel early text messages

[David H. Lynch Jr.]

[-- Attachment #1: Type: text/plain, Size: 2053 bytes --]

Michael Galassi wrote:
>> In an effort to get past the "Rebooting to System ACE Configuration
>> Address 6..." message, I have selected "support for early boot texts
>> over serial port" in kernel debugging.
>>     
>
> This will not work.  Unfortunately linux here shows its roots as PS OS,
> there is a vast amount of hardcoded NS8250/NS16550 knowledge in there.
> I'd started modularizing this so we could use a UART lite but the more I
> dug the lower my motivation.  This is really ugly code in an area that
> is of virtually no interest.  For now get rid of support for early boot
> texts over serial port" and use your JTAG to read the buffers if you
> panic before the console is up and running.
>   
   UartLite early boot text support is fairly trivial. It is also fairly 
standalone.
    You basically need about 5 lines  of code to output a character.
   
    David Bolcsfoldi has posted patches to Peter's driver that should work.
    I have had working early boot text support long before I even had a 
working driver.
    There is atleast one other Uart family besides the 8250 with early 
boot text support.
     You can also fairly trivially fold my early boot text code into 
Peter's driver.
   
    The most disconcerting problem is the #ifdef that results with 
something like 4 or cases.

    The non-8250 support could be more modularized - but it still not a 
major pain.



> -michael
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein


[-- Attachment #2: Type: text/html, Size: 3114 bytes --]

Re: uartlite with 2.6.17 kernel and kernel early text messages

[David H. Lynch Jr.]

Robert Corley wrote:
> I am still trying to get the UARTLITE to work with 2.6 and a plb_temac design.  I am using EDK 7.1.  I have generated the edk files and copied xparameters_ml300.h to arch/ppc/platforms/4xx/xparameters.xparameters_ml403.h
>
> In an effort to get past the "Rebooting to System ACE Configuration Address 6..." message, I have selected "support for early boot texts over serial port" in kernel debugging.
>   
    I do not think Peter's driver actually supports early boot texts. I 
beleive the patches David  Bolcsfoldi posted add early boot support.

    If you can not get that working I posted a driver in January (as 
part of a patchset for the Pico E12) that has early boot support - 
though on inspection David's patches looked like they should work.


> FYI, I'm using Peter's patches to create the uartlite.c and associated files and have selected the uartlite using make menuconfig.
>
> The first error is as follows:
> athena startup_network # make ARCH=ppc zImage.initrd
>   CHK     include/linux/version.h
>   CHK     include/linux/compile.h
> dnsdomainname: Unknown host
>   CC      arch/ppc/syslib/gen550_dbg.o
> arch/ppc/syslib/gen550_dbg.c:36: error: `RS_TABLE_SIZE' undeclared here (not in a function)
> arch/ppc/syslib/gen550_dbg.c:38: error: empty scalar initializer
> arch/ppc/syslib/gen550_dbg.c:38: error: (near initialization for `rs_table')
> arch/ppc/syslib/gen550_dbg.c:36: error: storage size of `rs_table' isn't known
> arch/ppc/syslib/gen550_dbg.c:36: warning: 'rs_table' defined but not used
> make[1]: *** [arch/ppc/syslib/gen550_dbg.o] Error 1
> make: *** [arch/ppc/syslib] Error 2
>   
    Unless you actually have an 8250 based Uart in your system - and you 
are not configured for one, then arch/ppc/syslib/gen550_dbg should NOT 
be getting built.
    Proper early boot text support for the UartLite requires both a 
replacement for this AND changes to use those instead of gen550_dbg.c

> so, I modified the gen550_dbg.c file to #include the xparameters.h, where the RS_TABLE_SIZE is defined
>  but still get more errors.  Here they are:
>   
    You do not want to touch gen550_dbg.c
    You need a new uartlite_dbg.c and Makefile and other changes to use it.

>   CHK     include/linux/version.h
>   CHK     include/linux/compile.h
> dnsdomainname: Unknown host
>   CC      arch/ppc/syslib/gen550_dbg.o
> arch/ppc/syslib/gen550_dbg.c:37: error: `XPAR_UARTNS550_0_CLOCK_FREQ_HZ' undeclared here (not in a function)
> arch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not constant
> arch/ppc/syslib/gen550_dbg.c:37: error: (near initialization for `rs_table[0].baud_base')
> arch/ppc/syslib/gen550_dbg.c:37: error: `XPAR_INTC_0_UARTNS550_0_VEC_ID' undeclared here (not in a function)
> arch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not constant
> arch/ppc/syslib/gen550_dbg.c:37: error: (near initialization for `rs_table[0].irq')
> arch/ppc/syslib/gen550_dbg.c:37: error: `XPAR_UARTNS550_0_BASEADDR' undeclared here (not in a function)
> arch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not constant
> arch/ppc/syslib/gen550_dbg.c:37: error: (near initialization for `rs_table[0].iomem_base')
> arch/ppc/syslib/gen550_dbg.c:37: error: initializer element is not constant
> arch/ppc/syslib/gen550_dbg.c:37: error: (near initialization for `rs_table[0]')
> make[1]: *** [arch/ppc/syslib/gen550_dbg.o] Error 1
> make: *** [arch/ppc/syslib] Error 2
>
> Questions:
>
> 1.    Now, is this an issue with the UARTLITE driver or is it just not supported for early messaging?
>   
    Peters does not support early boot texts.
    Both David's and my drivers do as do David's patches to Peter's driver./
> 2.   What am I missing w.r.t. getting something out of the serial port?
> 3.    I am assuming that the boot args for a initrd boot are:  "console=ttyUL0 ip=off root=/dev/ram rw", correct?
>   
    Presuming you have no other serial device I think you should not 
need any console= argument at all.
> -corley
>
>
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>   


-- 
Dave Lynch 					  	    DLA Systems
Software Development:  				         Embedded Linux
717.627.3770 	       dhlii@dlasys.net 	  http://www.dlasys.net
fax: 1.253.369.9244 			           Cell: 1.717.587.7774
Over 25 years' experience in platforms, languages, and technologies too numerous to list.

"Any intelligent fool can make things bigger and more complex... It takes a touch of genius - and a lot of courage to move in the opposite direction."
Albert Einstein