> -----Original Message----- > From: Weiny, Ira [mailto:ira.weiny@xxxxxxxxx] > Sent: Tuesday, October 24, 2017 11:20 PM > To: Oded Nissan <odedni@xxxxxxxxxxxx>; 'Hal Rosenstock > (hal@xxxxxxxxxxxxxxxxxx)' <hal@xxxxxxxxxxxxxxxxxx> > Cc: linux-rdma@xxxxxxxxxxxxxxx > Subject: RE: ibqueryerrors.c: Add support for additional counters in > PortCountersExtended > > > > > Hi Ira, > > > > Users don't set thresholds for old/new counters, they just set > > thresholds and if new counters are supported then thresholds are > > checked against new counters and if only old counters are supported > > then thresholds are checked against old counters. > > But you are not checking the same set of counters. > > IB_PC_EXT_ERR_SYM_F, > IB_PC_EXT_LINK_RECOVERS_F, > IB_PC_EXT_LINK_DOWNED_F, > IB_PC_EXT_ERR_RCV_F, > IB_PC_EXT_ERR_PHYSRCV_F, > IB_PC_EXT_ERR_SWITCH_REL_F, > IB_PC_EXT_XMT_DISCARDS_F, > IB_PC_EXT_ERR_XMTCONSTR_F, > IB_PC_EXT_ERR_RCVCONSTR_F, > IB_PC_EXT_ERR_LOCALINTEG_F, > IB_PC_EXT_ERR_EXCESS_OVR_F, > IB_PC_EXT_VL15_DROPPED_F, > IB_PC_EXT_XMT_WAIT_F, > IB_PC_EXT_QP1_DROP_F, > IB_PC_EXT_ERR_LAST_F, > > > VS. > > /* > * GS Performance > */ > IB_PC_FIRST_F, > IB_PC_PORT_SELECT_F = IB_PC_FIRST_F, > IB_PC_COUNTER_SELECT_F, > IB_PC_ERR_SYM_F, > IB_PC_LINK_RECOVERS_F, > IB_PC_LINK_DOWNED_F, > IB_PC_ERR_RCV_F, > IB_PC_ERR_PHYSRCV_F, > IB_PC_ERR_SWITCH_REL_F, > IB_PC_XMT_DISCARDS_F, > IB_PC_ERR_XMTCONSTR_F, > IB_PC_ERR_RCVCONSTR_F, > IB_PC_COUNTER_SELECT2_F, > IB_PC_ERR_LOCALINTEG_F, > IB_PC_ERR_EXCESS_OVR_F, > IB_PC_VL15_DROPPED_F, > IB_PC_XMT_BYTES_F, > IB_PC_RCV_BYTES_F, > IB_PC_XMT_PKTS_F, > IB_PC_RCV_PKTS_F, > IB_PC_XMT_WAIT_F, > IB_PC_LAST_F, > > For example you no longer can specify IB_PC_XMT_BYTES_F. > > As I don't think anyone uses these thresholds for this counter I'm going to let > the patch go. But generally this probably should be cleaned up somehow. I > just don't have the time as I am sure you don't... ;-) > > Ira > > These counters: IB_PC_XMT_BYTES_F, IB_PC_RCV_BYTES_F, IB_PC_XMT_PKTS_F, IB_PC_RCV_PKTS_F, Are not error counters and also in the old version users couldn't set thresholds for them. Oded > > > > -----Original Message----- > > From: Weiny, Ira [mailto:ira.weiny@xxxxxxxxx] > > Sent: Saturday, September 30, 2017 4:42 AM > > To: 'Hal Rosenstock (hal@xxxxxxxxxxxxxxxxxx)' > > <hal@xxxxxxxxxxxxxxxxxx>; Oded Nissan <odedni@xxxxxxxxxxxx> > > Cc: linux-rdma@xxxxxxxxxxxxxxx > > Subject: RE: ibqueryerrors.c: Add support for additional counters in > > PortCountersExtended > > > > Hal, Oded, > > > > I have lost the original email from this submission but got it off of > patchwork. > > > > Question on the hunk below. > > > > diff --git a/src/ibqueryerrors.c b/src/ibqueryerrors.c index > > 304dc154c4b4..4ea762a1f344 100644 > > --- a/src/ibqueryerrors.c > > +++ b/src/ibqueryerrors.c > > @@ -105,15 +105,15 @@ static unsigned valid_gid(ib_gid_t * gid) > > return memcmp(&zero_gid, gid, sizeof(*gid)); } > > > > -static void set_thres(char *name, uint32_t val) > > +static void set_thres(char *name, uint64_t val) > > { > > int f; > > int n; > > char tmp[256]; > > - for (f = IB_PC_FIRST_F; f <= IB_PC_LAST_F; f++) { > > + for (f = IB_PC_EXT_ERR_SYM_F; f <= IB_PC_EXT_XMT_WAIT_F; f++) > > + { > > > > How does this change allow for users to set the old Port Counters > thresholds? > > > > Ira > > > > > > if (strcmp(name, mad_field_name(f)) == 0) { > > mad_encode_field(thresholds, f, &val); > > - snprintf(tmp, 255, "[%s = %u]", name, val); > > + snprintf(tmp, 255, "[%s = %lu]", name, val); > > threshold_str = realloc(threshold_str, > > strlen(threshold_str)+strlen(tmp)+1); > > if (!threshold_str) { -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
