* [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() [not found] <20230315181902.4177819-1-joel@joelfernandes.org> @ 2023-03-15 18:18 ` Joel Fernandes (Google) 2023-03-26 12:34 ` Joel Fernandes 2023-03-15 18:19 ` [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() Joel Fernandes (Google) 1 sibling, 1 reply; 8+ messages in thread From: Joel Fernandes (Google) @ 2023-03-15 18:18 UTC (permalink / raw) To: Saeed Mahameed, Leon Romanovsky, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, Boris Pismenny Cc: Uladzislau Rezki (Sony), Ariel Levkovich, Vlad Buslov, Paul E . McKenney, Joel Fernandes, netdev, linux-rdma, linux-kernel From: "Uladzislau Rezki (Sony)" <urezki@gmail.com> The kfree_rcu() and kvfree_rcu() macros' single-argument forms are deprecated. Therefore switch to the new kfree_rcu_mightsleep() and kvfree_rcu_mightsleep() variants. The goal is to avoid accidental use of the single-argument forms, which can introduce functionality bugs in atomic contexts and latency bugs in non-atomic contexts. Cc: Ariel Levkovich <lariel@nvidia.com> Cc: Saeed Mahameed <saeedm@nvidia.com> Cc: Vlad Buslov <vladbu@nvidia.com> Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> --- drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c | 2 +- drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c index ca834bbcb44f..8afcec0c5d3c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c @@ -242,7 +242,7 @@ mlx5e_int_port_remove(struct mlx5e_tc_int_port_priv *priv, mlx5_del_flow_rules(int_port->rx_rule); mapping_remove(ctx, int_port->mapping); mlx5e_int_port_metadata_free(priv, int_port->match_metadata); - kfree_rcu(int_port); + kfree_rcu_mightsleep(int_port); priv->num_ports--; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c index 08d0929e8260..b811dad7370a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c @@ -670,7 +670,7 @@ static int mlx5e_macsec_del_txsa(struct macsec_context *ctx) mlx5e_macsec_cleanup_sa(macsec, tx_sa, true); mlx5_destroy_encryption_key(macsec->mdev, tx_sa->enc_key_id); - kfree_rcu(tx_sa); + kfree_rcu_mightsleep(tx_sa); macsec_device->tx_sa[assoc_num] = NULL; out: @@ -849,7 +849,7 @@ static void macsec_del_rxsc_ctx(struct mlx5e_macsec *macsec, struct mlx5e_macsec xa_erase(&macsec->sc_xarray, rx_sc->sc_xarray_element->fs_id); metadata_dst_free(rx_sc->md_dst); kfree(rx_sc->sc_xarray_element); - kfree_rcu(rx_sc); + kfree_rcu_mightsleep(rx_sc); } static int mlx5e_macsec_del_rxsc(struct macsec_context *ctx) -- 2.40.0.rc1.284.g88254d51c5-goog ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() 2023-03-15 18:18 ` [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() Joel Fernandes (Google) @ 2023-03-26 12:34 ` Joel Fernandes 2023-03-27 15:09 ` Saeed Mahameed 0 siblings, 1 reply; 8+ messages in thread From: Joel Fernandes @ 2023-03-26 12:34 UTC (permalink / raw) To: Saeed Mahameed, Leon Romanovsky, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, Boris Pismenny Cc: Uladzislau Rezki (Sony), Ariel Levkovich, Vlad Buslov, Paul E . McKenney, netdev, linux-rdma, linux-kernel On Wed, Mar 15, 2023 at 2:19 PM Joel Fernandes (Google) <joel@joelfernandes.org> wrote: > > From: "Uladzislau Rezki (Sony)" <urezki@gmail.com> > > The kfree_rcu() and kvfree_rcu() macros' single-argument forms are > deprecated. Therefore switch to the new kfree_rcu_mightsleep() and > kvfree_rcu_mightsleep() variants. The goal is to avoid accidental use > of the single-argument forms, which can introduce functionality bugs in > atomic contexts and latency bugs in non-atomic contexts. In a world where patches anxiously await their precious Ack, could today be our lucky day on this one? We need Acks to take this in for 6.4. David? Others? - Joel > > Cc: Ariel Levkovich <lariel@nvidia.com> > Cc: Saeed Mahameed <saeedm@nvidia.com> > Cc: Vlad Buslov <vladbu@nvidia.com> > Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> > Signed-off-by: Paul E. McKenney <paulmck@kernel.org> > Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> > --- > drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c | 2 +- > drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c | 4 ++-- > 2 files changed, 3 insertions(+), 3 deletions(-) > > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c > index ca834bbcb44f..8afcec0c5d3c 100644 > --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c > @@ -242,7 +242,7 @@ mlx5e_int_port_remove(struct mlx5e_tc_int_port_priv *priv, > mlx5_del_flow_rules(int_port->rx_rule); > mapping_remove(ctx, int_port->mapping); > mlx5e_int_port_metadata_free(priv, int_port->match_metadata); > - kfree_rcu(int_port); > + kfree_rcu_mightsleep(int_port); > priv->num_ports--; > } > > diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c > index 08d0929e8260..b811dad7370a 100644 > --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c > +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c > @@ -670,7 +670,7 @@ static int mlx5e_macsec_del_txsa(struct macsec_context *ctx) > > mlx5e_macsec_cleanup_sa(macsec, tx_sa, true); > mlx5_destroy_encryption_key(macsec->mdev, tx_sa->enc_key_id); > - kfree_rcu(tx_sa); > + kfree_rcu_mightsleep(tx_sa); > macsec_device->tx_sa[assoc_num] = NULL; > > out: > @@ -849,7 +849,7 @@ static void macsec_del_rxsc_ctx(struct mlx5e_macsec *macsec, struct mlx5e_macsec > xa_erase(&macsec->sc_xarray, rx_sc->sc_xarray_element->fs_id); > metadata_dst_free(rx_sc->md_dst); > kfree(rx_sc->sc_xarray_element); > - kfree_rcu(rx_sc); > + kfree_rcu_mightsleep(rx_sc); > } > > static int mlx5e_macsec_del_rxsc(struct macsec_context *ctx) > -- > 2.40.0.rc1.284.g88254d51c5-goog > ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() 2023-03-26 12:34 ` Joel Fernandes @ 2023-03-27 15:09 ` Saeed Mahameed 0 siblings, 0 replies; 8+ messages in thread From: Saeed Mahameed @ 2023-03-27 15:09 UTC (permalink / raw) To: Joel Fernandes Cc: Saeed Mahameed, Leon Romanovsky, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, Boris Pismenny, Uladzislau Rezki (Sony), Ariel Levkovich, Vlad Buslov, Paul E . McKenney, netdev, linux-rdma, linux-kernel On 26 Mar 08:34, Joel Fernandes wrote: >On Wed, Mar 15, 2023 at 2:19 PM Joel Fernandes (Google) ><joel@joelfernandes.org> wrote: >> >> From: "Uladzislau Rezki (Sony)" <urezki@gmail.com> >> >> The kfree_rcu() and kvfree_rcu() macros' single-argument forms are >> deprecated. Therefore switch to the new kfree_rcu_mightsleep() and >> kvfree_rcu_mightsleep() variants. The goal is to avoid accidental use >> of the single-argument forms, which can introduce functionality bugs in >> atomic contexts and latency bugs in non-atomic contexts. > >In a world where patches anxiously await their precious Ack, could >today be our lucky day on this one? > >We need Acks to take this in for 6.4. David? Others? > For mlx5 usually me, but since this is a larger series that is not mlx5 centric and targeting multiple tree, I really don't know which subsystem you should be targeting.. for netdev submissions you need to specify the targeted branch e.g. [PATCH v2 net-next 06/14] ... FWIW: Reviewed-by: Saeed Mahameed <saeedm@nvidia.com> > - Joel > > >> >> Cc: Ariel Levkovich <lariel@nvidia.com> >> Cc: Saeed Mahameed <saeedm@nvidia.com> >> Cc: Vlad Buslov <vladbu@nvidia.com> >> Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> >> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> >> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> >> --- >> drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c | 2 +- >> drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c | 4 ++-- >> 2 files changed, 3 insertions(+), 3 deletions(-) >> >> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c >> index ca834bbcb44f..8afcec0c5d3c 100644 >> --- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c >> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc/int_port.c >> @@ -242,7 +242,7 @@ mlx5e_int_port_remove(struct mlx5e_tc_int_port_priv *priv, >> mlx5_del_flow_rules(int_port->rx_rule); >> mapping_remove(ctx, int_port->mapping); >> mlx5e_int_port_metadata_free(priv, int_port->match_metadata); >> - kfree_rcu(int_port); >> + kfree_rcu_mightsleep(int_port); >> priv->num_ports--; >> } >> >> diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c >> index 08d0929e8260..b811dad7370a 100644 >> --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c >> +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c >> @@ -670,7 +670,7 @@ static int mlx5e_macsec_del_txsa(struct macsec_context *ctx) >> >> mlx5e_macsec_cleanup_sa(macsec, tx_sa, true); >> mlx5_destroy_encryption_key(macsec->mdev, tx_sa->enc_key_id); >> - kfree_rcu(tx_sa); >> + kfree_rcu_mightsleep(tx_sa); >> macsec_device->tx_sa[assoc_num] = NULL; >> >> out: >> @@ -849,7 +849,7 @@ static void macsec_del_rxsc_ctx(struct mlx5e_macsec *macsec, struct mlx5e_macsec >> xa_erase(&macsec->sc_xarray, rx_sc->sc_xarray_element->fs_id); >> metadata_dst_free(rx_sc->md_dst); >> kfree(rx_sc->sc_xarray_element); >> - kfree_rcu(rx_sc); >> + kfree_rcu_mightsleep(rx_sc); >> } >> >> static int mlx5e_macsec_del_rxsc(struct macsec_context *ctx) >> -- >> 2.40.0.rc1.284.g88254d51c5-goog >> ^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() [not found] <20230315181902.4177819-1-joel@joelfernandes.org> 2023-03-15 18:18 ` [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() Joel Fernandes (Google) @ 2023-03-15 18:19 ` Joel Fernandes (Google) 2023-03-15 18:38 ` Bob Pearson 1 sibling, 1 reply; 8+ messages in thread From: Joel Fernandes (Google) @ 2023-03-15 18:19 UTC (permalink / raw) To: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, Devesh Sharma, Bob Pearson Cc: Joel Fernandes (Google), Paul E . McKenney, Zhu Yanjun, linux-rdma, linux-kernel The k[v]free_rcu() macro's single-argument form is deprecated. Therefore switch to the new k[v]free_rcu_mightsleep() variant. The goal is to avoid accidental use of the single-argument forms, which can introduce functionality bugs in atomic contexts and latency bugs in non-atomic contexts. There is no functionality change with this patch. Link: https://lore.kernel.org/rcu/20230201150815.409582-1-urezki@gmail.com Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> Fixes: 72a03627443d ("RDMA/rxe: Remove rxe_alloc()") Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> --- drivers/infiniband/sw/rxe/rxe_mr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index b10aa1580a64..ae3a100e18fb 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -731,7 +731,7 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) return -EINVAL; rxe_cleanup(mr); - kfree_rcu(mr); + kfree_rcu_mightsleep(mr); return 0; } -- 2.40.0.rc1.284.g88254d51c5-goog ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() 2023-03-15 18:19 ` [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() Joel Fernandes (Google) @ 2023-03-15 18:38 ` Bob Pearson 2023-03-15 18:41 ` Joel Fernandes 0 siblings, 1 reply; 8+ messages in thread From: Bob Pearson @ 2023-03-15 18:38 UTC (permalink / raw) To: Joel Fernandes (Google), Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, Devesh Sharma Cc: Paul E . McKenney, Zhu Yanjun, linux-rdma, linux-kernel On 3/15/23 13:19, Joel Fernandes (Google) wrote: > The k[v]free_rcu() macro's single-argument form is deprecated. > Therefore switch to the new k[v]free_rcu_mightsleep() variant. The goal > is to avoid accidental use of the single-argument forms, which can > introduce functionality bugs in atomic contexts and latency bugs in > non-atomic contexts. > > There is no functionality change with this patch. > > Link: https://lore.kernel.org/rcu/20230201150815.409582-1-urezki@gmail.com > Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> > Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> > Reviewed-by: Paul E. McKenney <paulmck@kernel.org> > Fixes: 72a03627443d ("RDMA/rxe: Remove rxe_alloc()") > Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> > --- > drivers/infiniband/sw/rxe/rxe_mr.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c > index b10aa1580a64..ae3a100e18fb 100644 > --- a/drivers/infiniband/sw/rxe/rxe_mr.c > +++ b/drivers/infiniband/sw/rxe/rxe_mr.c > @@ -731,7 +731,7 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) > return -EINVAL; > > rxe_cleanup(mr); > - kfree_rcu(mr); > + kfree_rcu_mightsleep(mr); > return 0; > } > Please replace kfree_rcu_mightsleep() by kfree(). There is no need to delay the kfree(mr). Bob ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() 2023-03-15 18:38 ` Bob Pearson @ 2023-03-15 18:41 ` Joel Fernandes 2023-03-15 19:05 ` Bob Pearson 0 siblings, 1 reply; 8+ messages in thread From: Joel Fernandes @ 2023-03-15 18:41 UTC (permalink / raw) To: Bob Pearson Cc: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, Devesh Sharma, Paul E . McKenney, Zhu Yanjun, linux-rdma, linux-kernel On Wed, Mar 15, 2023 at 2:38 PM Bob Pearson <rpearsonhpe@gmail.com> wrote: > > On 3/15/23 13:19, Joel Fernandes (Google) wrote: > > The k[v]free_rcu() macro's single-argument form is deprecated. > > Therefore switch to the new k[v]free_rcu_mightsleep() variant. The goal > > is to avoid accidental use of the single-argument forms, which can > > introduce functionality bugs in atomic contexts and latency bugs in > > non-atomic contexts. > > > > There is no functionality change with this patch. > > > > Link: https://lore.kernel.org/rcu/20230201150815.409582-1-urezki@gmail.com > > Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> > > Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> > > Reviewed-by: Paul E. McKenney <paulmck@kernel.org> > > Fixes: 72a03627443d ("RDMA/rxe: Remove rxe_alloc()") > > Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> > > --- > > drivers/infiniband/sw/rxe/rxe_mr.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c > > index b10aa1580a64..ae3a100e18fb 100644 > > --- a/drivers/infiniband/sw/rxe/rxe_mr.c > > +++ b/drivers/infiniband/sw/rxe/rxe_mr.c > > @@ -731,7 +731,7 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) > > return -EINVAL; > > > > rxe_cleanup(mr); > > - kfree_rcu(mr); > > + kfree_rcu_mightsleep(mr); > > return 0; > > } > > > Please replace kfree_rcu_mightsleep() by kfree(). There is no need to delay the kfree(mr). I thought you said either was Ok, but yeah that's fine with me. I was trying to play it safe ;-). An extra GP may not hurt, but not having one when it is needed might. I will update it to just be kfree. <quote> > rxe_cleanup(mr); > - kfree_rcu(mr); > + kfree_rcu_mightsleep(mr); > return 0; > } > I would prefer just - kfree_rcu(mr); + kfree(mr); but either one will work. Bob </quote> - Joel ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() 2023-03-15 18:41 ` Joel Fernandes @ 2023-03-15 19:05 ` Bob Pearson 2023-03-15 19:59 ` Joel Fernandes 0 siblings, 1 reply; 8+ messages in thread From: Bob Pearson @ 2023-03-15 19:05 UTC (permalink / raw) To: Joel Fernandes Cc: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, Devesh Sharma, Paul E . McKenney, Zhu Yanjun, linux-rdma, linux-kernel On 3/15/23 13:41, Joel Fernandes wrote: > On Wed, Mar 15, 2023 at 2:38 PM Bob Pearson <rpearsonhpe@gmail.com> wrote: >> >> On 3/15/23 13:19, Joel Fernandes (Google) wrote: >>> The k[v]free_rcu() macro's single-argument form is deprecated. >>> Therefore switch to the new k[v]free_rcu_mightsleep() variant. The goal >>> is to avoid accidental use of the single-argument forms, which can >>> introduce functionality bugs in atomic contexts and latency bugs in >>> non-atomic contexts. >>> >>> There is no functionality change with this patch. >>> >>> Link: https://lore.kernel.org/rcu/20230201150815.409582-1-urezki@gmail.com >>> Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> >>> Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> >>> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> >>> Fixes: 72a03627443d ("RDMA/rxe: Remove rxe_alloc()") >>> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> >>> --- >>> drivers/infiniband/sw/rxe/rxe_mr.c | 2 +- >>> 1 file changed, 1 insertion(+), 1 deletion(-) >>> >>> diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c >>> index b10aa1580a64..ae3a100e18fb 100644 >>> --- a/drivers/infiniband/sw/rxe/rxe_mr.c >>> +++ b/drivers/infiniband/sw/rxe/rxe_mr.c >>> @@ -731,7 +731,7 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) >>> return -EINVAL; >>> >>> rxe_cleanup(mr); >>> - kfree_rcu(mr); >>> + kfree_rcu_mightsleep(mr); >>> return 0; >>> } >>> >> Please replace kfree_rcu_mightsleep() by kfree(). There is no need to delay the kfree(mr). > > I thought you said either was Ok, but yeah that's fine with me. I was > trying to play it safe ;-). An extra GP may not hurt, but not having > one when it is needed might. I will update it to just be kfree. Thanks. The reason to not add the pause is that it really isn't required and will just make people think that it is. With the current state of the driver the mr cleanup code will disable looking up the mr from its rkey or lkey and then wait until all the references to the mr are dropped before calling kfree. This will always work (unless a new bug is introduced) so no reason to add the rcu delay. Bob > > <quote> >> rxe_cleanup(mr); >> - kfree_rcu(mr); >> + kfree_rcu_mightsleep(mr); >> return 0; >> } >> > > I would prefer just > - kfree_rcu(mr); > + kfree(mr); > > but either one will work. > Bob > </quote> > > - Joel ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() 2023-03-15 19:05 ` Bob Pearson @ 2023-03-15 19:59 ` Joel Fernandes 0 siblings, 0 replies; 8+ messages in thread From: Joel Fernandes @ 2023-03-15 19:59 UTC (permalink / raw) To: Bob Pearson Cc: Zhu Yanjun, Jason Gunthorpe, Leon Romanovsky, Devesh Sharma, Paul E . McKenney, Zhu Yanjun, linux-rdma, linux-kernel On Wed, Mar 15, 2023 at 3:05 PM Bob Pearson <rpearsonhpe@gmail.com> wrote: > > On 3/15/23 13:41, Joel Fernandes wrote: > > On Wed, Mar 15, 2023 at 2:38 PM Bob Pearson <rpearsonhpe@gmail.com> wrote: > >> > >> On 3/15/23 13:19, Joel Fernandes (Google) wrote: > >>> The k[v]free_rcu() macro's single-argument form is deprecated. > >>> Therefore switch to the new k[v]free_rcu_mightsleep() variant. The goal > >>> is to avoid accidental use of the single-argument forms, which can > >>> introduce functionality bugs in atomic contexts and latency bugs in > >>> non-atomic contexts. > >>> > >>> There is no functionality change with this patch. > >>> > >>> Link: https://lore.kernel.org/rcu/20230201150815.409582-1-urezki@gmail.com > >>> Acked-by: Zhu Yanjun <zyjzyj2000@gmail.com> > >>> Reviewed-by: Bob Pearson <rpearsonhpe@gmail.com> > >>> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> > >>> Fixes: 72a03627443d ("RDMA/rxe: Remove rxe_alloc()") > >>> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> > >>> --- > >>> drivers/infiniband/sw/rxe/rxe_mr.c | 2 +- > >>> 1 file changed, 1 insertion(+), 1 deletion(-) > >>> > >>> diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c > >>> index b10aa1580a64..ae3a100e18fb 100644 > >>> --- a/drivers/infiniband/sw/rxe/rxe_mr.c > >>> +++ b/drivers/infiniband/sw/rxe/rxe_mr.c > >>> @@ -731,7 +731,7 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) > >>> return -EINVAL; > >>> > >>> rxe_cleanup(mr); > >>> - kfree_rcu(mr); > >>> + kfree_rcu_mightsleep(mr); > >>> return 0; > >>> } > >>> > >> Please replace kfree_rcu_mightsleep() by kfree(). There is no need to delay the kfree(mr). > > > > I thought you said either was Ok, but yeah that's fine with me. I was > > trying to play it safe ;-). An extra GP may not hurt, but not having > > one when it is needed might. I will update it to just be kfree. > > Thanks. The reason to not add the pause is that it really isn't required and will just make > people think that it is. With the current state of the driver the mr cleanup code will disable > looking up the mr from its rkey or lkey and then wait until all the references to the mr are dropped > before calling kfree. This will always work (unless a new bug is introduced) so no reason to > add the rcu delay. > Sounds good! I updated it in my queue, and I will repost it in the future pending test results and other comments. thanks, - Joel ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2023-03-27 15:09 UTC | newest] Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- [not found] <20230315181902.4177819-1-joel@joelfernandes.org> 2023-03-15 18:18 ` [PATCH v2 06/14] net/mlx5: Rename kfree_rcu() to kfree_rcu_mightsleep() Joel Fernandes (Google) 2023-03-26 12:34 ` Joel Fernandes 2023-03-27 15:09 ` Saeed Mahameed 2023-03-15 18:19 ` [PATCH v2 13/14] RDMA/rxe: Rename kfree_rcu() to kvfree_rcu_mightsleep() Joel Fernandes (Google) 2023-03-15 18:38 ` Bob Pearson 2023-03-15 18:41 ` Joel Fernandes 2023-03-15 19:05 ` Bob Pearson 2023-03-15 19:59 ` Joel Fernandes
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).