sweep_bh_for_rgrps(9)

💅 sweep_bh_for_rgrps(9) — a Kernel API Manual moment

okay so bestie, this is from the Kernel API Manual, and honestly? it's giving "cleanup era" and i'm obsessed 💖

---

✨ NAME ✨

sweep_bh_for_rgrps — find an rgrp in a meta buffer and free blocks therein 🧹

---

🌸 SYNOPSIS 🌸

int sweep_bh_for_rgrps (struct gfs2_inode *ip , struct gfs2_holder
*rd_gh , struct buffer_head *bh , __be64 *start , __be64 *end , bool
meta , u32 *btotal );

---

🎀 ARGUMENTS 🎀

• ip — the inode, bestie. that's the whole vibe.
• rd_gh — holder of resource group glock 🔒
• bh — buffer head to sweep (yes, we're sweeping it, periodt)
• start — starting point in bh
• end — end point in bh
• meta — true if bh points to metadata (rather than data) 🫶
• btotal — place to keep count of total blocks freed 🔥

---

🦋 DESCRIPTION 🦋

so here's the tea ☕ — we sweep a metadata buffer (the one provided by the metapath) for blocks we need to free, and we free them all. no cap.

but!! we do it one rgrp at a time. the way that works: if this block has references to multiple rgrps, we break it into individual transactions. this lets other processes use the rgrps while we're focused on a single one — better concurrency / performance, iykyk 💅

and at every transaction boundary, we rewrite the inode into the journal ✍️. that way the bitmaps are kept consistent with the inode and we can recover if we're interrupted by power-outages 😭⚡ (yes, genuinely — that's the whole safety net, don't skip it).

---

🌈 RETURN 🌈

0, or return code if an error occurred. and *btotal has the total number of blocks freed 🍓

---

👑 SEE ALSO 👑

Kernel file ./fs/gfs2/bmap.c — and the whole friend group:

• gfs2_unstuffer_folio(9)
• gfs2_unstuff_dinode(9)
• find_metapath(9)
• metaptr1(9)
• metapointer(9)
• lookup_metapath(9)
• fillup_metapath(9)
• gfs2_extent_length(9)
• gfs2_hole_size(9)
• __gfs2_iomap_alloc(9)
• gfs2_alloc_size(9)
• __gfs2_iomap_get(9)
• gfs2_block_map(9)
• gfs2_journaled_truncate(9)
• find_nonnull_ptr(9)
• punch_hole(9)
• do_shrink(9)
• do_grow(9)
• gfs2_setattr_size(9)
• gfs2_free_journal_extents(9)
• gfs2_add_jextent(9)
• gfs2_map_journal_extents(9)
• gfs2_write_alloc_required(9)

---

🎀 that's the whole page, bestie — thanks for reading, mwah 💖

sign-off time: this one lives at ./fs/gfs2, and the date on it is October 2026 ⭐ go forth and sweep responsibly, fr fr 🧚