Github/littlefs
1
Fork 0
Code Actions Activity

More bikeshedding around mdir mids

Browse Source 
- Use mroot address to determine if we follow mroot during splits
- Prefer bid == -1/bid != -1 for now
- Use u.m when copying mdir internals

I also looked at dropping the bid == -1 representation of inlined mroots,
but it's just too convenient for now. We can leverage address
comparisons to see if we are committing to the actual mroot, and we can
(expensively) compare the mdir blocks for other mroot checks, but we
also use bid == -1 to indicate if we're on the mroot chain in both
lfsr_mdir_commit and lfsr_mtree_traverse...

This is a bit of a shame, since reserving -1 either limits these bids to
15-bits, which is concerning, or requires special handling to cut off
bids at 2^16-1.
This commit is contained in:
Christopher Haster
2023-08-10 22:29:13 -05:00
parent e34665723c
commit 99b83a4ef1
1 changed files with 14 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lfs.c
+14 -15
View File
@@ -5170,18 +5170,17 @@ static inline lfs_size_t lfsr_mtree_weight(lfs_t *lfs) {
}
static int lfsr_mtree_lookup(lfs_t *lfs, lfsr_mid_t mid, lfsr_mdir_t *mdir_) {
// TODO should we really allow -1=>mroot lookup?
LFS_ASSERT(mid.bid >= -1);
LFS_ASSERT(mid.bid < (lfs_ssize_t)lfsr_mtree_weight(lfs));
// looking up mroot?
if (mid.bid < 0) {
if (lfsr_mtree_isinlined(lfs)) {
LFS_ASSERT(mid.bid == -1);
mdir_->mid = mid;
mdir_->u = lfs->mroot.u;
mdir_->u.m = lfs->mroot.u.m;
return 0;
// look up mdir in actual mtree
} else {
LFS_ASSERT(mid.bid >= 0);
LFS_ASSERT(mid.bid < (lfs_ssize_t)lfsr_mtree_weight(lfs));
lfsr_tag_t tag;
lfsr_data_t data;
int err = lfsr_btree_lookup(lfs, &lfs->mtree, mid.bid,
@@ -5394,7 +5393,7 @@ static int lfsr_mdir_compact_(lfs_t *lfs, lfsr_mdir_t *mdir_,
// drop commit if weight goes to zero
if (mdir_->mid.bid >= 0 && mdir_->u.m.weight == 0) {
if (mdir_->mid.bid != -1 && mdir_->u.m.weight == 0) {
// TODO should we just make our pcache not assert?
// drop our pcache, we're not going to complete this commit
lfs_cache_zero(lfs, &lfs->pcache);
@@ -5460,7 +5459,7 @@ static int lfsr_mdir_commit_(lfs_t *lfs, lfsr_mdir_t *mdir,
}
// drop commit if weight goes to zero
if (mdir_.mid.bid >= 0 && mdir_.u.m.weight == 0) {
if (mdir_.mid.bid != -1 && mdir_.u.m.weight == 0) {
// TODO move this up into lfsr_mdir_commit?
// consume gstate so we don't lose any info
int err = lfsr_fs_consumegdelta(lfs, mdir);
@@ -6087,10 +6086,10 @@ static int lfsr_mdir_commit(lfs_t *lfs, lfsr_mdir_t *mdir,
!= lfsr_mtree_weight(lfs));
opened_mdir->mid.bid = msibling_.mid.bid;
opened_mdir->mid.rid -= mdir_.u.m.weight;
opened_mdir->u = msibling_.u;
opened_mdir->u.m = msibling_.u.m;
} else {
opened_mdir->mid.bid = mdir_.mid.bid;
opened_mdir->u = mdir_.u;
opened_mdir->u.m = mdir_.u.m;
}
} else if (opened_mdir->mid.bid > mdir->mid.bid) {
opened_mdir->mid.bid += lfsr_btree_weight(&mtree_)
@@ -6103,20 +6102,20 @@ static int lfsr_mdir_commit(lfs_t *lfs, lfsr_mdir_t *mdir,
// update mdir to follow requested rid
LFS_ASSERT(mdir->mid.rid <= (lfs_ssize_t)mdir->u.m.weight);
if (mdir->mid.bid == -1 && mdir->mid.rid == -1) {
mdir->u = mroot_.u;
if (mdir == &lfs->mroot) {
// do nothing, we update mroot later
} else if (!lfsr_mdir_isdropped(&msibling_)
&& mdir->mid.rid >= (lfs_ssize_t)mdir_.u.m.weight) {
LFS_ASSERT(lfsr_btree_weight(&mtree_) != lfsr_mtree_weight(lfs));
mdir->mid.bid = msibling_.mid.bid;
mdir->mid.rid -= mdir_.u.m.weight;
mdir->u = msibling_.u;
mdir->u.m = msibling_.u.m;
} else {
mdir->u = mdir_.u;
mdir->u.m = mdir_.u.m;
}
// update our mroot and mtree
lfs->mroot.u = mroot_.u;
lfs->mroot.u.m = mroot_.u.m;
lfs->mtree = mtree_;
return 0;