Skip to contentSkip to conversation

RDMA/rw: Remove force_mr module parameter

PATCH In review Leon Romanovsky 2 participants 3 messages Jul 2, 2017 2 messages not in the archive InfiniBand

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 RDMA/rw: Remove force_mr module parameter

From: Leon Romanovsky <leonro-VPRAkNaXOzVWk0Htik3J/[email protected]>

The force_mr module parameter was introduced in the commit a060b5629ab0 ("IB/core: generic RDMA READ/WRITE API").

Quoted 1 line

From the beginning, the fourth parameter of the module_param_named

macro was set to 0. Passing 0 there means that module parameter is not created and that adding "options ib_core force_mr=1" to a modprobe.conf file has no effect.

Since the force_mr parameter is intended for debug and unlikely will be used by anyone except developers and exposure of module parameter can limit ourselves in debug ability. The preferable way will be to remove this module parameter entirely and don't fix the file permission parameter (the fourth parameter).

Fixes a060b5629ab0 ("IB/core: generic RDMA READ/WRITE API")

---
Chuck,

I added your SOB to this commit, because this commit message was highly influenced by yours initial patch.

Please let me know if it is fine with you.

Thanks

Changelog:
v0->v1:
1. Rewrote commit message and comment in the code
2. Removed extra blank line

---
 drivers/infiniband/core/rw.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
drivers/infiniband/core/rw.c +4 -3
index dbfd854c32c9..447099574ede 100644
@@ -22,9 +22,10 @@ enum {
 	RDMA_RW_SIG_MR,
 };
 
+/*
+ * Debug option to force MR registration logic.
+ */
 static bool rdma_rw_force_mr;
-module_param_named(force_mr, rdma_rw_force_mr, bool, 0);
-MODULE_PARM_DESC(force_mr, "Force usage of MRs for RDMA READ/WRITE operations");
 
 /*
  * Check if the device might use memory registration.  This is currently only
@@ -503,7 +504,7 @@ struct ib_send_wr *rdma_rw_ctx_wrs(struct rdma_rw_ctx *ctx, struct ib_qp *qp,
 		rdma_rw_update_lkey(&ctx->sig->data, true);
 		if (ctx->sig->prot.mr)
 			rdma_rw_update_lkey(&ctx->sig->prot, true);
-
+
 		ctx->sig->sig_mr->need_inval = true;
 		ib_update_fast_reg_key(ctx->sig->sig_mr,
 			ib_inc_rkey(ctx->sig->sig_mr->lkey));

--
2.13.2

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

A message that is not in the archive (sent off-list or not yet received): 20170702062425.6780-1-leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org

A message that is not in the archive (sent off-list or not yet received): [email protected]

#4 lore raw
Quoted 2 lines · Christoph Hellwig

NAK. It's a very useful debug helper and should instead also be exposed at runtime.

So dramatic,
Last time, you wrote that you (AUTHOR) didn't use this module parameter, so what did happen from that post?

----
https://www.spinics.net/lists/linux-rdma/msg49891.html
I think I only added it on requests from someone (you?), and when writing the code debugged it by manually forcing it to on.
----

No one is removing the logic and the code in the IB/core, the change is here is in removing visibility to the user.

Thanks