bugGNU GRUB - Bugs: bug #54739, "alloc magic is broken"...

 
 

You are not allowed to post comments on this tracker with your current authentication level.

bug #54739: "alloc magic is broken" on NVME SSD

Submitter:  Magnus Lucchese <maglu>
Submitted:  Thu 27 Sep 2018 12:50:31 AM UTC
   
 
Category:  Booting Severity:  Major
Priority:  5 - Normal Item Group:  Software Error
Status:  None Privacy:  Public
Assigned to:  None Originator Name: 
Open/Closed:  Open Release:  2.02
Release:  Reproducibility:  Every Time
Planned Release:  None

Thu 27 Sep 2018 12:50:31 AM UTC, original submission:  

I have a server with multiple NVME SSDs, I noticed that after some tests on these SSDs I got the error below when rebooting (the address change on each run)

alloc magic is broken at 0x34e79f00: ffffffffffffffff

(see full log attached)

The SSD that cause the fail is not the one were GRUB/Linux are installed, when I remove it the system boots correctly.

Some things that I found during debug


  • This SSD format is 4KB + 16 Bytes of metadata, metadata is interleaved with the LBA (NVME Metadata Capabilities (MC))
  • The SSD is completely erased (return all 1's).
  • The maximum allowed size for reads on this SSD is 128KB (NVME MDTS parameter).
  • GRUB is trying to perform some reads greater than 128KB and receives an error from the SSD ("Invalid Field Error").
  • If I change the SSD format to 4KB+0 GRUB works properly and don't issue any reader greater than the maximum size.

 

My Suspicions


 I think that when GRUB calculate the maximum size it can read it is not considering metadata interleaved. The reads that extrapolate the maximum size overflow by the amount of metadata, besides that, it seems that it is not handling the error received from the SSD.

Versions Tested


2.02~beta2-36ubuntu3.18
2.02~beta3-4ubuntu7

Magnus Lucchese <maglu>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attached Files
file #45104:  grub_error_report.txt added by maglu (35KiB - text/plain)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by maglu (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

     

    Follows 1 latest change.

    Date Changed by Updated Field Previous Value => Replaced by
    2018-09-27 maglu Attached File- Added grub_error_report.txt, #45104

    Back to the top

    Powered by Savane 3.13-d3ae.
    Corresponding source code