[Webkit-unassigned] [Bug 79191] [JSC] The end atom of the marked block should be considered to decide if the cell is live

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=79191





--- Comment #10 from hojong.han at samsung.com  2012-02-28 17:19:30 PST ---
Here is some information brought from the core dump.

(JSC::MarkedBlock *) 0x47210000 = {
  <WTF::DoublyLinkedListNode<JSC::MarkedBlock>> = {<No data fields>}, 
  members of JSC::MarkedBlock: 
  static atomSize = 16, 
  static atomShift = 5, 
  static blockSize = 16384, 
  static blockMask = 4294950912, 
  static atomsPerBlock = 1024, 
  static atomMask = 1023, 
  static cardShift = 8, 
  static bytesPerCard = 256, 
  static cardCount = 64, 
  static cardMask = 63, 
  static atomAlignmentMask = 15, 
  m_atomsPerCell = 7, 
  m_endAtom = 1018, 
  m_marks = {
    static wordSize = 32, 
    static words = 32, 
    static one = <optimized out>, 
    bits = {
      m_data = {33818624, 270549121, 2164392968, 135274560, 1082196484, 67637280, 541098242, 33818640, 270549121, 2164392968, 135274560, 1082196484, 67637280, 541098242, 33818640, 
        270549121, 2164392968, 135274560, 1082196484, 67637280, 541098242, 33818640, 270549121, 2164392968, 135274560, 1082196484, 67637280, 541098242, 33818640, 270549121, 
        2164392968, 135274560}
    }
  }, 
  m_state = JSC::MarkedBlock::Marked, 
  m_allocation = {
    <WTF::PageBlock> = {
      m_realBase = 0x47210000, 
      m_base = 0x47210000, 
      m_size = 16384
    }, 
    members of WTF::PageAllocationAligned: 
    m_reservation = {
      m_realBase = 0x4720e000, 
      m_base = 0x4720e000, 
      m_size = 28672
    }
  }, 
  m_heap = 0x5b4be8, 
  m_prev = 0x46508000, 
  m_next = 0x434b4000
}

[Memory dump of the register file]
0x48550700:     0x432533f0    0xfffffffb    0x4640c514    0xfffffffb
0x48550710:     0x00000000    0xfffffffb    0x4640c458    0xfffffffb
0x48550720:     0x026004d0    0xfffffffb    0x428bbeb0    0xfffffffb
0x48550730:     0x428bbeb0    0xfffffffb    0x428bbef0    0xfffffffb
0x48550740:     0x428bbeb0    0xfffffffb    0x00000001    0xffffffff
0x48550750:     0x00000001    0xffffffff    0x41baf3b0    0xfffffffb
0x48550760:     0x426301b0    0xfffffffb    0x428bbe70    0xfffffffb
0x48550770:     0x428bbe30    0xfffffffb    0x00000000    0xfffffffc
0x48550780:     0x426301b0    0xfffffffb    0x00000001    0xffffffff
0x48550790:     0x47213fb0    0xfffffffb    0x00000000    0xfffffffc
0x485507a0:     0x465a88b0    0xfffffffb    0x00000001    0xffffffff
0x485507b0:     0x43eb46b0    0xfffffffb    0x438f2af0    0xfffffffb
0x485507c0:     0x452fd89c    0xfffffffb    0x00c988c8    0xfffffffc
0x485507d0:     0x43eb4670    0xfffffffb    0x43eb4670    0xfffffffb

What I mentioned is 0x47213fb0 in the register file.
This address is fit to the 1019th atom of that MarkedBlock, 
and cannot be filtered by in-middle-of-cell check.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.