Linux 7.1 Revamps T10 PI Data Integrity Handling For Better Read Performance
Merged yesterday for the Linux 7.1 kernel is overhauling the T10 PI code for generating and verifying data integrity information. In turn the new code is cleaner while also allowing for better read storage performance.
For those interested in T10 PI for end-to-end data integrity, including data in transit, Linux 7.1 has reworked its implementation. Christoph Hellwig led the effort to support generating and verifying integrity information in the file-system rather than the "automatic below the covers" support that has existed. In turn this increases the protection covered and the file-system then has knowledge of it and the ability to act upon integrity errors.
The new code is also more efficient particularly for read workloads. Read performance has improved by around 15% for 4K I/O and also enjoying lower QD1 latency for small I/O sizes.
So far with the revamped code, the XFS file-system is wired up for supporting the T10 protection information.
More details on this revised T10 data integrity handling for the Linux kernel can be found via this pull that landed on Monday in Git.
For those interested in T10 PI for end-to-end data integrity, including data in transit, Linux 7.1 has reworked its implementation. Christoph Hellwig led the effort to support generating and verifying integrity information in the file-system rather than the "automatic below the covers" support that has existed. In turn this increases the protection covered and the file-system then has knowledge of it and the ability to act upon integrity errors.
The new code is also more efficient particularly for read workloads. Read performance has improved by around 15% for 4K I/O and also enjoying lower QD1 latency for small I/O sizes.
So far with the revamped code, the XFS file-system is wired up for supporting the T10 protection information.
More details on this revised T10 data integrity handling for the Linux kernel can be found via this pull that landed on Monday in Git.