If we happen to alloc a extent buffer and then alloc a page and notice that page is already attached to an extent buffer, we will only unlock it and free our existing eb. Any pages currently attached to that eb will be properly freed, but we don't do the page_cache_release() on the page where we noticed the other extent buffer which can cause us to leak pages and I hope cause the weird issues we've been seeing in this area. Thanks,
Unified · Split
| @@ -4120,6 +4120,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, · 4119 unmodified lines | ||
| 4120 | 4120 | if (atomic_inc_not_zero(&exists->refs)) { |
| 4121 | 4121 | spin_unlock(&mapping->private_lock); |
| 4122 | 4122 | unlock_page(p); |
| 4123 | + page_cache_release(p); | |
| 4123 | 4124 | mark_extent_buffer_accessed(exists); |
| 4124 | 4125 | goto free_eb; |
| 4125 | 4126 | } |
| @@ -4199,8 +4200,7 @@ free_eb: · 73 unmodified lines | ||
| 4199 | 4200 | unlock_page(eb->pages[i]); |
| 4200 | 4201 | } |
| 4201 | 4202 | |
| 4202 | − if (!atomic_dec_and_test(&eb->refs)) | |
| 4203 | − return exists; | |
| 4203 | + WARN_ON(!atomic_dec_and_test(&eb->refs)); | |
| 4204 | 4204 | btrfs_release_extent_buffer(eb); |
| 4205 | 4205 | return exists; |
| 4206 | 4206 | } |