7eabffafa
Denys Vlasenko
2009-10-14
Subject
dd: use correct multiplication factor and simplify code
Binary
No binary (build failed)
Architecture & Compiler Flags
—
Build Log
(log yok)
Commit Diff GitHub →
coreutils/dd.c+7 -19
| @@ -49,9 +49,9 @@ struct globals { | |
| static void dd_output_status(int UNUSED_PARAM cur_signal) | |
| { | |
| #if ENABLE_FEATURE_DD_THIRD_STATUS_LINE | |
| - | unsigned long long total; |
| - | unsigned long long diff_scaled; |
| - | unsigned long long diff_us = monotonic_us(); /* before fprintf */ |
| + | double seconds; |
| + | unsigned long long bytes_sec; |
| + | unsigned long long now_us = monotonic_us(); /* before fprintf */ |
| #endif | |
| /* Deliberately using %u, not %d */ | |
| @@ -72,24 +72,12 @@ static void dd_output_status(int UNUSED_PARAM cur_signal) | |
| * (echo DONE) | ./busybox dd >/dev/null | |
| * (sleep 1; echo DONE) | ./busybox dd >/dev/null | |
| */ | |
| - | diff_us -= G.begin_time_us; |
| - | /* We need to calculate "(total * 1000000) / usec" without overflow. |
| - | * this would work too, but is bigger than integer code below. |
| - | * total = G.total_bytes * (double)1000000 / (diff_us ? diff_us : 1); |
| - | */ |
| - | diff_scaled = diff_us; |
| - | total = G.total_bytes; |
| - | while (total > MAXINT(unsigned long long) / (1024 * 1024)) { |
| - | total >>= 1; |
| - | diff_scaled >>= 1; |
| - | } |
| - | total *= (1024 * 1024); /* should be 1000000, but it's +45 bytes */ |
| - | if (diff_scaled > 1) |
| - | total /= diff_scaled; |
| + | seconds = (now_us - G.begin_time_us) / 1000000.0; |
| + | bytes_sec = G.total_bytes / seconds; |
| fprintf(stderr, "%f seconds, %sB/s\n", | |
| - | diff_us / 1000000.0, |
| + | seconds, |
| /* show fractional digit, use suffixes */ | |
| - | make_human_readable_str(total, 1, 0) |
| + | make_human_readable_str(bytes_sec, 1, 0) |
| ); | |
| #endif | |
| } | |