On Fri, Apr 28, 2017 at 04:27:29PM +0200, Wolfram Sang wrote:
> On Fri, Apr 28, 2017 at 04:41:33PM +0300, Dan Carpenter wrote:
>
> Adding Tobin to CC, he is way more into the driver than I am these
> days...
I've mocked up a patch set to address this issue (along with some
further refactoring of the containing function). I'm going to sit on
it for a couple days before submitting. I was not able to decipher the
reason for the line that caused the checker warning. FYI the patch removes
the if statement.
thanks,
Tobin.
> > Hello Wolfram Sang,
> >
> > The patch 13a9930d15b4: "staging: ks7010: add driver from Nanonote
> > extra-repository" from May 31, 2016, leads to the following static
> > checker warning:
> >
> > drivers/staging/ks7010/ks_wlan_net.c:1141 ks_wlan_get_range()
> > warn: dead code because of 'i > 2' and 'i < 13'
> >
> > drivers/staging/ks7010/ks_wlan_net.c
> > 1062 static int ks_wlan_get_range(struct net_device *dev,
> > 1063 struct iw_request_info *info,
> > 1064 struct iw_point *dwrq, char *extra)
> > 1065 {
> > 1066 struct ks_wlan_private *priv =
> > 1067 (struct ks_wlan_private *)netdev_priv(dev);
> > 1068 struct iw_range *range = (struct iw_range *)extra;
> > 1069 int i, k;
> > 1070
> > 1071 DPRINTK(2, "\n");
> > 1072
> > 1073 if (priv->sleep_mode == SLP_SLEEP)
> > 1074 return -EPERM;
> > 1075
> > 1076 /* for SLEEP MODE */
> > 1077 dwrq->length = sizeof(struct iw_range);
> > 1078 memset(range, 0, sizeof(*range));
> > 1079 range->min_nwid = 0x0000;
> > 1080 range->max_nwid = 0x0000;
> > 1081 range->num_channels = 14;
> > 1082 /* Should be based on cap_rid.country to give only
> > 1083 * what the current card support
> > 1084 */
> > 1085 k = 0;
> > 1086 for (i = 0; i < 13; i++) { /* channel 1 -- 13 */
> > ^^^^^^
> > 1087 range->freq[k].i = i + 1; /* List index */
> > 1088 range->freq[k].m = frequency_list[i] * 100000;
> > 1089 range->freq[k++].e = 1; /* Values in table in MHz -> * 10^5 * 10 */
> > 1090 }
> >
> > i is always 13 after the loop.
> >
> > 1091 range->num_frequency = k;
> > 1092 if (priv->reg.phy_type == D_11B_ONLY_MODE || priv->reg.phy_type == D_11BG_COMPATIBLE_MODE) { /* channel 14 */
> > 1093 range->freq[13].i = 14; /* List index */
> > 1094 range->freq[13].m = frequency_list[13] * 100000;
> > 1095 range->freq[13].e = 1; /* Values in table in MHz -> * 10^5 * 10 */
> > 1096 range->num_frequency = 14;
> > 1097 }
> > 1098
> > 1099 /* Hum... Should put the right values there */
> > 1100 range->max_qual.qual = 100;
> > 1101 range->max_qual.level = 256 - 128; /* 0 dBm? */
> > 1102 range->max_qual.noise = 256 - 128;
> > 1103 range->sensitivity = 1;
> > 1104
> > 1105 if (priv->reg.phy_type == D_11B_ONLY_MODE) {
> > 1106 range->bitrate[0] = 1e6;
> > 1107 range->bitrate[1] = 2e6;
> > 1108 range->bitrate[2] = 5.5e6;
> > 1109 range->bitrate[3] = 11e6;
> > 1110 range->num_bitrates = 4;
> > 1111 } else { /* D_11G_ONLY_MODE or D_11BG_COMPATIBLE_MODE */
> > 1112 range->bitrate[0] = 1e6;
> > 1113 range->bitrate[1] = 2e6;
> > 1114 range->bitrate[2] = 5.5e6;
> > 1115 range->bitrate[3] = 11e6;
> > 1116
> > 1117 range->bitrate[4] = 6e6;
> > 1118 range->bitrate[5] = 9e6;
> > 1119 range->bitrate[6] = 12e6;
> > 1120 if (IW_MAX_BITRATES < 9) {
> > 1121 range->bitrate[7] = 54e6;
> > 1122 range->num_bitrates = 8;
> > 1123 } else {
> > 1124 range->bitrate[7] = 18e6;
> > 1125 range->bitrate[8] = 24e6;
> > 1126 range->bitrate[9] = 36e6;
> > 1127 range->bitrate[10] = 48e6;
> > 1128 range->bitrate[11] = 54e6;
> > 1129
> > 1130 range->num_bitrates = 12;
> > 1131 }
> > 1132 }
> > 1133
> > 1134 /* Set an indication of the max TCP throughput
> > 1135 * in bit/s that we can expect using this interface.
> > 1136 * May be use for QoS stuff... Jean II
> > 1137 */
> > 1138 if (i > 2)
> > ^^^^^
> > So this condition is always true. I'm not sure what was intended,
> > though.
> >
> > 1139 range->throughput = 5000 * 1000;
> > 1140 else
> > 1141 range->throughput = 1500 * 1000;
> > 1142
> >
> > regards,
> > dan carpenter
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
