Messages in this thread | | | Date | Fri, 7 Nov 2014 10:40:22 +0100 | Subject | Re: [LKP] [dmi] PANIC: early exception 0e rip 10:ffffffff81899e6b error 9 cr2 ffffffffff240000 | From | Ard Biesheuvel <> |
| |
On 7 November 2014 10:36, Matt Fleming <matt.fleming@intel.com> wrote: > On Fri, 2014-11-07 at 17:26 +0800, Yuanhan Liu wrote: >> >> Here is the output: >> >> [ 0.000000] NX (Execute Disable) protection: active >> [ 0.000000] ====== Decoding _DMI_ header: >> [ 0.000000] 5f 44 4d 49 5f 48 a3 0b 00 20 60 8f 3e 00 25 00 >> [ 0.000000] ====== Remapped SMBIOS table 0xffffffff8f602000 at ffffffffff240000, size 0xba3, num 0x3e > > Smells like a sign extension issue. Previously how could dmi_base (u32) > hold 0xffffffff8f602000? >
Exactly. And note that we already found (and fixed, or so we thought) this issue.
I.e, on the ARM you get
>>> ====== Remapped SMBIOS table 0xfe8ac000 at ffffff800001e000, size 0x262, num 0xc
which has the top bit set as well, but is handled correctly, whereas with the original code (i.e., before adding the get_unaligned_le32()), we were hitting the same problem.
-- Ard.
| |