Find the answer to your Linux question:
Results 1 to 6 of 6
Recently I have met an error [ 72.643005] ash: Corrupted page table at address ffff88007f294000 [ 72.643005] at ffff88007f294000 [ 72.643005] IP: [<ffffffff8113e775>] copy_page_c+0x5/0x10 [ 72.643005] PGD 160a063 PUD 1fffb067 ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Just Joined!
    Join Date
    Aug 2013
    Posts
    6

    The page table corrupted error


    Recently I have met an error

    [ 72.643005] ash: Corrupted page table at address ffff88007f294000
    [ 72.643005] at ffff88007f294000
    [ 72.643005] IP: [<ffffffff8113e775>] copy_page_c+0x5/0x10
    [ 72.643005] PGD 160a063 PUD 1fffb067 PMD ff7f9
    [ 72.643005] Oops: 0009 [#1] PREEMPT SMP

    with the Oops: 0009 [#1]. With the information from lxr.linux.no/#linux+v3.9.4/arch/x86/mm/fault.c

    Page fault error code bits:
    bit 0 == 0: no page found 1: protection fault
    bit 1 == 0: read access 1: write access
    bit 2 == 0: kernel-mode access 1: user-mode access
    bit 3 == 1: use of reserved bit detected
    bit 4 == 1: fault was an instruction fetch

    The error is due to protection fault and use of reserved bit detected. Are there anybody knowing about this error. What does it exactly mean?

    Thank you

  2. #2
    Linux Guru Rubberman's Avatar
    Join Date
    Apr 2009
    Location
    I can be found either 40 miles west of Chicago, in Chicago, or in a galaxy far, far away.
    Posts
    11,737
    In my opinion, you most likely have a bad memory chip. Reboot into the memcheck86 tool and check your memory. This is one reason why I ALWAYS get ECC RAM capable systems, and fully-buffered ECC RAM to run in them. It has saved my bacon on more than one occasion. Using that, if a memory chip/stick fails, then the OS will automatically map it out of use, resulting in less physical memory, but a still healthy system, allowing you to replace the RAM in a timely manner. The page tables are kept in physical RAM, hence my diagnosis.

    One other possibility is bad cache memory, either external, or internal to the CPU. However, it is more likely system RAM.
    Sometimes, real fast is almost as good as real time.
    Just remember, Semper Gumbi - always be flexible!

  3. #3
    Just Joined!
    Join Date
    Aug 2013
    Posts
    6
    Hi Rubberman,

    Thank you for your reply.

    Could you explain clearly for me about you comment about system RAM that can cause page table corrupted issue?
    It seems you are right, in fact I tried to change the system RAM to map to "MymapRAM" as the following in proc/iomem:

    00100000-12ffffff : System RAM
    01000000-01242043 : Kernel code
    01242044-016455ff : Kernel data
    016b7000-0174dfff : Kernel bss
    02000000-09ffffff : Crash kernel
    13000000-7e0fffff : MymapRAM
    7e100000-7eefafff : System RAM
    7eefb000-7ffeffff : MymapRAM

    The region system RAM for kernel is about 310M, not as the traditional kernel map 1G:3G. Could you please give me some hints that modifying system RAM like in my case for kernel space leads to the page corrupted issue with PF_RSVD. As I understand the PF_RSVD is set when the process accesses the address larger than CPU can access, for example on AMD64, the kernel limits the address to only 47 bit instead of all 64 bit).

    Thank you,

  4. $spacer_open
    $spacer_close
  5. #4
    Linux Guru Rubberman's Avatar
    Join Date
    Apr 2009
    Location
    I can be found either 40 miles west of Chicago, in Chicago, or in a galaxy far, far away.
    Posts
    11,737
    Many systems (usually non-server desktop systems or laptops) do not use error-corrected (ECC) memory. IE, it is possible for memory to be corrupted and the system not detect it until something like you have experienced occurs. So, if a memory stick is faulty, or overheats (this has happened to me), then it can cause issues like what you have experienced. Fortunately, my personal workstation has ECC RAM, so when I had an overheating problem, the OS automatically mapped the memory stick out of use, and did not cause any functional problems. If I didn't have ECC RAM, then "bad" things could have happened!
    Sometimes, real fast is almost as good as real time.
    Just remember, Semper Gumbi - always be flexible!

  6. #5
    Just Joined!
    Join Date
    Aug 2013
    Posts
    6
    Thanks Rubberman. For my case "MymapRAM", do you have any idea about how it can lead to the page corrupted table?. I mean how the issue can occur in the software side since I changed the normal allocation system RAM memory of Linux.

  7. #6
    Linux Guru Rubberman's Avatar
    Join Date
    Apr 2009
    Location
    I can be found either 40 miles west of Chicago, in Chicago, or in a galaxy far, far away.
    Posts
    11,737
    Since you remapped your RAM, it is also possible that you included some invalid addresses, or addresses that are used for other stuff, such as hardware memory mapped I/O. Switch back to the default mappings and see if the problem persists. Also, do run memcheck86 to verify that your RAM is physically ok.
    Sometimes, real fast is almost as good as real time.
    Just remember, Semper Gumbi - always be flexible!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •