> From: Kundapura, Ganapati [mailto:ganapati.kundap...@intel.com] > Sent: Thursday, 30 May 2024 13.42 > > Hi, > > > From: Morten Brørup <m...@smartsharesystems.com> > > Sent: Thursday, May 30, 2024 4:44 PM > > > > > From: Akhil Goyal [mailto:gak...@marvell.com] > > > Sent: Thursday, 30 May 2024 13.02 > > > > > > ++ Morten for comment on #if vs #ifdef > > > > > > > > Subject: [EXTERNAL] [PATCH v2 1/2] crypto: fix build issues on > > > > > unsetting > > > crypto > > > > > callbacks macro > > > > > > > > > > Crypto callbacks macro is defined with value 1 and being used with > > > > > ifdef, on config value is changed to 0 to disable, crypto callback > > > > > changes still being compiled. > > > > > > > > > > Used #if instead of #ifdef and also wrapped crypto callback > > > > > changes under RTE_CRYPTO_CALLBACKS macro to fix build issues when > > > > > macro is unset. > > > > > > > > > > Fixes: 1c3ffb95595e ("cryptodev: add enqueue and dequeue > > > > > callbacks") > > > > > Fixes: 5523a75af539 ("test/crypto: add case for enqueue/dequeue > > > callbacks") > > > > > > > > > > Signed-off-by: Ganapati Kundapura <ganapati.kundap...@intel.com> > > > > > --- > > > > > v2: > > > > > * Used #if instead of #ifdef and restored macro definition in > > > > > config > > > > > * Split callback registration check in a seperate patch > > > > The DPDK convention is #ifdef, not #if. > > Ethdev also uses #ifdef, e.g.: > > https://elixir.bootlin.com/dpdk/v24.03/source/lib/ethdev/rte_ethdev.h#L61 > > 12 > > > > PS: Personally, I prefer #if too. But we should set personal preferences > aside, > > and follow the existing convention in DPDK and use #ifdef. > Irrespective of RTE_CRYPTO_CALLBACKS value set to 0/1, > #ifdef RTE_CRYPTO_CALLBACKS returns true always.
Yes, because it is defined in both cases. So according to the #ifdef convention, rte_config.h must contain either: #define RTE_CRYPTO_CALLBACKS 1 Or: /* RTE_CRYPTO_CALLBACKS is not set */
