Re: [OOPS] hugetlbfs tests with 2.6.30-rc8-git1

Michael Ellerman wrote:

Hi Sachin,

Rather than "-git7" can you tell us the actual SHA, I don't know what
git7 is.
  

Corresponding SHA is 44b7532b8b464f606053562400719c9c21276037.

quoted

Here is the supporting data.

 cpu 0x1: Vector: 300 (Data Access) at [c0000000fe9b3220]
    pc: c00000000003d620: .hpte_need_flush+0x1bc/0x2d8
    lr: c00000000003d4d0: .hpte_need_flush+0x6c/0x2d8
    sp: c0000000fe9b34a0
   msr: 8000000000009032
   dar: c0000000283b0d78
    

This address looks pretty innocuous, but I notice you have
DEBUG_PAGEALLOC=y, so presumably that's why it's triggering.

I can't see from the snippet of disassembly you posted where in the C
code we are, can you work it out?
  

The corresponding C code is :

278:   2f 80 00 00     cmpwi   cr7,r0,0
       } else {
               vsid = get_kernel_vsid(addr, mmu_kernel_ssize);
               ssize = mmu_kernel_ssize;
       }
       vaddr = hpt_va(addr, vsid, ssize);
       rpte = __real_pte(__pte(pte), ptep);
27c:   eb 89 80 00     ld      r28,-32768(r9)  ^^^^ %pc points to the above line. ^^^^^
        * Check if we have an active batch on this CPU. If not, just
        * flush now and return. For now, we don global invalidates
        * in that case, might be worth testing the mm cpu mask though
        * and decide to use local invalidates instead...

I have attached the objdump o/p for tlb_hash64.o.

I could not recreate this issue with git8 kernel
(45e3e1935e2857c54783291107d33323b3ef33c8).

Thanks
-Sachin

-- 

---------------------------------
Sachin Sant
IBM Linux Technology Center
India Systems and Technology Labs
Bangalore, India
---------------------------------