Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux

Merge branch 'smc-fix'

D. Wythe says:

====================
fix unexpected SMC_CLC_DECL_ERR_REGRMB error

We can easily trigger the SMC_CLC_DECL_ERR_REGRMB exception within
following script:

server: smc_run nginx
client: smc_run ./wrk -c 2000 -t 8 -d 20 http://smc-server

And we can clearly see that this error is also divided into two types:

1. 0x09990003
2. 0x05000000/0x09990003

Which has the same root causes, but the immediate causes vary.

The root cause of this issues is that remove connections from link group
is not synchronous with add/delete rtoken entry, which means that even
the number of connections is less that SMC_RMBS_PER_LGR_MAX, it does not
mean that the connection can register rtoken successfully later. In
other words, the rtoken entry may released, This will cause an
unexpected SMC_CLC_DECL_ERR_REGRMB to be reported, and then this SMC
connections have to fallback to TCP.

This patch set handles two types of SMC_CLC_DECL_ERR_REGRMB exceptions
from different perspectives.

Patch 1: fix the 0x05000000/0x09990003 error.
Patch 2: fix the 0x09990003 error.

After those patches, there is no SMC_CLC_DECL_ERR_REGRMB exceptions in
my
test case any more.

v1 -> v2:
- add bugfix patch for SMC_CLC_DECL_ERR_REGRMB cause by server side
v2 -> v3:
- fix incorrect mail thread
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

+3 -2
+3 -2
net/smc/smc_core.c
··· 1161 1161 cancel_work_sync(&conn->abort_work); 1162 1162 } 1163 1163 if (!list_empty(&lgr->list)) { 1164 - smc_lgr_unregister_conn(conn); 1165 1164 smc_buf_unuse(conn, lgr); /* allow buffer reuse */ 1165 + smc_lgr_unregister_conn(conn); 1166 1166 } 1167 1167 1168 1168 if (!lgr->conns_num) ··· 1864 1864 (ini->smcd_version == SMC_V2 || 1865 1865 lgr->vlan_id == ini->vlan_id) && 1866 1866 (role == SMC_CLNT || ini->is_smcd || 1867 - lgr->conns_num < SMC_RMBS_PER_LGR_MAX)) { 1867 + (lgr->conns_num < SMC_RMBS_PER_LGR_MAX && 1868 + !bitmap_full(lgr->rtokens_used_mask, SMC_RMBS_PER_LGR_MAX)))) { 1868 1869 /* link group found */ 1869 1870 ini->first_contact_local = 0; 1870 1871 conn->lgr = lgr;