On 2024/8/6 18:49, Simon Horman wrote:
On Mon, Aug 05, 2024 at 05:05:50PM +0800, Wen Gu wrote:
Currently we have the statistics on sndbuf/RMB sizes of all connections
that have ever been on the link group, namely smc_stats_memsize. However
these statistics are incremental and since the ringbufs of link group
are allowed to be reused, we cannot know the actual allocated buffers
through these. So here introduces the statistic on actual allocated
ringbufs of the link group, it will be incremented when a new ringbuf is
added into buf_list and decremented when it is deleted from buf_list.
Signed-off-by: Wen Gu <guwen@xxxxxxxxxxxxxxxxx>
---
include/uapi/linux/smc.h | 4 ++++
net/smc/smc_core.c | 52 ++++++++++++++++++++++++++++++++++++----
net/smc/smc_core.h | 2 ++
3 files changed, 54 insertions(+), 4 deletions(-)
diff --git a/include/uapi/linux/smc.h b/include/uapi/linux/smc.h
index b531e3ef011a..d27b8dc50f90 100644
--- a/include/uapi/linux/smc.h
+++ b/include/uapi/linux/smc.h
@@ -127,6 +127,8 @@ enum {
SMC_NLA_LGR_R_NET_COOKIE, /* u64 */
SMC_NLA_LGR_R_PAD, /* flag */
SMC_NLA_LGR_R_BUF_TYPE, /* u8 */
+ SMC_NLA_LGR_R_SNDBUF_ALLOC, /* u64 */
+ SMC_NLA_LGR_R_RMB_ALLOC, /* u64 */
__SMC_NLA_LGR_R_MAX,
SMC_NLA_LGR_R_MAX = __SMC_NLA_LGR_R_MAX - 1
};
@@ -162,6 +164,8 @@ enum {
SMC_NLA_LGR_D_V2_COMMON, /* nest */
SMC_NLA_LGR_D_EXT_GID, /* u64 */
SMC_NLA_LGR_D_PEER_EXT_GID, /* u64 */
+ SMC_NLA_LGR_D_SNDBUF_ALLOC, /* u64 */
+ SMC_NLA_LGR_D_DMB_ALLOC, /* u64 */
__SMC_NLA_LGR_D_MAX,
SMC_NLA_LGR_D_MAX = __SMC_NLA_LGR_D_MAX - 1
};
diff --git a/net/smc/smc_core.c b/net/smc/smc_core.c
index 71fb334d8234..73c7999fc74f 100644
--- a/net/smc/smc_core.c
+++ b/net/smc/smc_core.c
@@ -221,6 +221,37 @@ static void smc_lgr_unregister_conn(struct smc_connection *conn)
write_unlock_bh(&lgr->conns_lock);
}
+/* must be called under lgr->{sndbufs|rmbs} lock */
+static inline void smc_lgr_buf_list_add(struct smc_link_group *lgr,
+ bool is_rmb,
+ struct list_head *buf_list,
+ struct smc_buf_desc *buf_desc)
Please do not use the inline keyword in .c files unless there is a
demonstrable reason to do so, e.g. performance. Rather, please allow
the compiler to inline functions as it sees fit.
The inline keyword in .h files is, of course, fine.
Yes.. I forgot to remove 'inline' when I moved these two helpers
from .h file to .c file. I will fix this in next version.
Thank you!
+{
+ list_add(&buf_desc->list, buf_list);
+ if (is_rmb) {
+ lgr->alloc_rmbs += buf_desc->len;
+ lgr->alloc_rmbs +=
+ lgr->is_smcd ? sizeof(struct smcd_cdc_msg) : 0;
+ } else {
+ lgr->alloc_sndbufs += buf_desc->len;
+ }
+}
+
+/* must be called under lgr->{sndbufs|rmbs} lock */
+static inline void smc_lgr_buf_list_del(struct smc_link_group *lgr,
+ bool is_rmb,
+ struct smc_buf_desc *buf_desc)
Ditto.
+{
+ list_del(&buf_desc->list);
+ if (is_rmb) {
+ lgr->alloc_rmbs -= buf_desc->len;
+ lgr->alloc_rmbs -=
+ lgr->is_smcd ? sizeof(struct smcd_cdc_msg) : 0;
+ } else {
+ lgr->alloc_sndbufs -= buf_desc->len;
+ }
+}
+
...
