Hi,

During a recent code review, I was confused multiple times by the
toptxn member of ReorderBufferTXN, which is defined only for
sub-transactions.

e.g. txn->toptxn member == NULL means the txn is a top level txn.
e.g. txn->toptxn member != NULL means the txn is not a top level txn

It makes sense if you squint and read it slowly enough, but IMO it's
too easy to accidentally misinterpret the meaning when reading code
that uses this member.

~

Such code can be made easier to read just by introducing some simple macros:

#define isa_toptxn(rbtxn) (rbtxn->toptxn == NULL)
#define isa_subtxn(rbtxn) (rbtxn->toptxn != NULL)
#define get_toptxn(rbtxn) (isa_subtxn(rbtxn) ? rbtxn->toptxn : rbtxn)

~

PSA a small patch that does this.

(Tests OK using make check-world)

Thoughts?

------
Kind Regards,
Peter Smith.
Fujitsu Australia

Attachment: v1-0001-Add-macros-for-ReorderBufferTXN-toptxn.patch
Description: Binary data

Reply via email to