I am unable to reproduce the problem:

# rpm -q powertop
powertop-2.3-5.el7.x86_64

# powertop --csv=filename
Loaded 0 prior measurements
Preparing to take measurements
Taking 1 measurement(s) for a duration of 20 second(s) each.
PowerTOP outputing using base filename filename.csv

# ls
filename.csv

What I got was:
# powertop --time=1 --csv=powertop.output
Loaded 0 prior measurements
Preparing to take measurements
Taking 1 measurement(s) for a duration of 1 second(s) each.
PowerTOP outputing using base filename powertop.csv

it seems powertop would parse file.* into file.csv, so this contradicted with my understanding from the manual, I thought the output file should be file.* instead of file.csv.

(In reply to Jimmy Pan from comment #3)
> What I got was:
> # powertop --time=1 --csv=powertop.output
> Loaded 0 prior measurements
> Preparing to take measurements
> Taking 1 measurement(s) for a duration of 1 second(s) each.
> PowerTOP outputing using base filename powertop.csv
> 
> it seems powertop would parse file.* into file.csv, so this contradicted
> with my understanding from the manual, I thought the output file should be
> file.* instead of file.csv.

BTW, if you use --html=file.*, the output file would be file.* from what I have tried.

I am probably going to update the docs. The current state is not perfect, but it works good for general use cases. The current algorithm tries to get the basename and then adds the default suffix. If running in iteration mode it outputs multiple reports, each in format: basename-timestamp.suffix. In such case it wouldn't be good for e.g. HTML to have basename-timestamp without the suffix. Adding another switch or template string to support the suffix specification seems to me as over-engineering this feature. Well, this could all be designed better way, but I think such discussion should go upstream.

This request was resolved in Red Hat Enterprise Linux 7.0.

Contact your manager or support representative in case you have further questions about the request.

The comment above is incorrect. The correct version is bellow.
I'm sorry for any inconvenience.
---------------------------------------------------------------

This request was NOT resolved in Red Hat Enterprise Linux 7.0.

Contact your manager or support representative in case you need
to escalate this bug.

Created attachment 1024712 [details]
Proposed fix

I rewrote the functionality and propose the patch upstream. With the patch applied I think it behaves more logically:

powertop --html
generates 'powertop.html' file

powertop --html=myfile.suffix
generates 'myfile.suffix' file

powertop -i 2 --html
generates 'powertop-TIMESTAMPS.html' files

powertop -i 2 --html=myfile.suffix
generates 'myfile-TIMESTAMPS.suffix' files

powertop -i 2 --html=myfile
generates 'myfile-TIMESTAMPS' files

Similarly for CSV.

Testing procedure:

echo "# version"
rpm -q powertop
echo

for output in --html --html=test --html=test.ext --csv --csv=test --csv=test.ext; do
    for iteration in '' '--iteration=2'; do
        mkdir -p /tmp/powertop${output}${iteration}; cd /tmp/powertop${output}${iteration}
        echo "# powertop $output $iteration --time=1"
        powertop powertop $output $iteration --time=1 >/dev/null 2>/dev/null
        ls
        echo
    done
done

Broken:

# version
powertop-2.3-5.el7.x86_64

# powertop --html  --time=1
powertop.html

# powertop --html --iteration=2 --time=1
powertop-20150916-120554.html
powertop-20150916-120556.html

# powertop --html=test  --time=1
test.html

# powertop --html=test --iteration=2 --time=1
test-20150916-120601.html
test-20150916-120602.html

# powertop --html=test.ext  --time=1
test.html

# powertop --html=test.ext --iteration=2 --time=1
test-20150916-120607.html
test-20150916-120609.html

# powertop --csv  --time=1
powertop.csv

# powertop --csv --iteration=2 --time=1
powertop-20150916-120614.csv
powertop-20150916-120615.csv

# powertop --csv=test  --time=1
test.csv

# powertop --csv=test --iteration=2 --time=1
test-20150916-120620.csv
test-20150916-120622.csv

# powertop --csv=test.ext  --time=1
test.csv

# powertop --csv=test.ext --iteration=2 --time=1
test-20150916-120627.csv
test-20150916-120628.csv

Fixed:

# version
powertop-2.3-9.el7.x86_64

# powertop --html  --time=1
powertop.html

# powertop --html --iteration=2 --time=1
powertop-20150916-120332.html
powertop-20150916-120333.html

# powertop --html=test  --time=1
test

# powertop --html=test --iteration=2 --time=1
test-20150916-120338
test-20150916-120340

# powertop --html=test.ext  --time=1
test.ext

# powertop --html=test.ext --iteration=2 --time=1
test-20150916-120345.ext
test-20150916-120346.ext

# powertop --csv  --time=1
powertop.csv

# powertop --csv --iteration=2 --time=1
powertop-20150916-120351.csv
powertop-20150916-120353.csv

# powertop --csv=test  --time=1
test

# powertop --csv=test --iteration=2 --time=1
test-20150916-120358
test-20150916-120359

# powertop --csv=test.ext  --time=1
test.ext

# powertop --csv=test.ext --iteration=2 --time=1
test-20150916-120404.ext
test-20150916-120406.ext

The --csv and --html options now behave as described in Comment 10.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2015-2225.html