[Bug 217965] ext4(?) regression since 6.5.0 on sata hdd

[bugzilla-daemon@xxxxxxxxxx]

https://bugzilla.kernel.org/show_bug.cgi?id=217965

--- Comment #37 from Eyal Lebedinsky (bugzilla@xxxxxxxxxxxxxx) ---
Finally got it.

kernel 6.5 was easy. kernel-6.4 debuginfo could not be found with 'dnf' and I
searched long for it. Finally locating it on koji.

Then the third probe was wrong, and after some learning I discovered
    $ sudo perf probe --line "ext4_mb_scan_aligned"
     21         while (i < EXT4_CLUSTERS_PER_GROUP(sb)) {
     22                 if (!mb_test_bit(i, bitmap)) {
     23                         max = mb_find_extent(e4b, i, sbi->s_stripe,
&ex);
     24                         if (max >= sbi->s_stripe) {
     25                                 ac->ac_found++;
     26                                 ex.fe_logical = 0xDEADF00D; /* debug
value */
                                        ac->ac_b_ex = ex;
     28                                 ext4_mb_use_best_found(ac, e4b);
     29                                 break;
                                }
                        }
     32                 i += sbi->s_stripe;
                }

and decided to pick line 24 (not the original 26). Is this right?

I will now upload the reports.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.