Add an SD card reader test.

Add an SD card reader test.

BUG=none
TEST=Automatic: Plug in an SD card and run this test.

An inserted SD card is required to get the USB device powered up.
We look for the Realtek card reader USB device, then for a
removable block device.

[petkov, 2010-08-13 19:48:17]

What's the purpose of this test?

Isn't checking for the specific Realtek device ID too specific? This should be
covered by the hardware_Components test already.

In general, this test and the USB tests -- when do you expect to run them? Are
they orphaned right now?

[bfreed, 2010-08-18 23:26:09]

On 2010/08/13 19:48:17, petkov wrote:
> What's the purpose of this test?
> 
> Isn't checking for the specific Realtek device ID too specific? This should be
> covered by the hardware_Components test already.

Yes, indeed it is.  I have now relaxed the USB check for one that looks for any
kind of "CARD READER" device.

The check for a non-sda partition is still needed to confirm the driver is
basically working.

> In general, this test and the USB tests -- when do you expect to run them? Are
> they orphaned right now?

This will be added to the HWQual suite.

[petkov, 2010-08-18 23:52:44]

http://codereview.chromium.org/3129013/diff/4001/5001
File client/site_tests/hardware_CardReader/control (right):

http://codereview.chromium.org/3129013/diff/4001/5001#newcode7
client/site_tests/hardware_CardReader/control:7: PURPOSE = 'Verify the Realtek
card reader basically works.'
Is this still a RealTek specific test? If so, you may as well leave the USB
vendor/device IDs in...

http://codereview.chromium.org/3129013/diff/4001/5002
File client/site_tests/hardware_CardReader/hardware_CardReader.py (right):

http://codereview.chromium.org/3129013/diff/4001/5002#newcode15
client/site_tests/hardware_CardReader/hardware_CardReader.py:15: # Verbose lsusb
output shows iProduct "USB2.0-CRW" and
Are these comments RealTek device specific?

http://codereview.chromium.org/3129013/diff/4001/5002#newcode26
client/site_tests/hardware_CardReader/hardware_CardReader.py:26: for line in
partition_output:
can't you just do:

for line in open("/proc/partitions"):

?

http://codereview.chromium.org/3129013/diff/4001/5002#newcode27
client/site_tests/hardware_CardReader/hardware_CardReader.py:27: part =
re.search(r' (sd[b-z])$', line)
This will break on ARM. Maybe that's expected anyway? Maybe mark somewhere that
this test works only on x86.

http://codereview.chromium.org/3129013/diff/4001/5002#newcode29
client/site_tests/hardware_CardReader/hardware_CardReader.py:29:
logging.info("Found partition %s" % part.group(1))
Use , instead of % for formatting logging arguments.

http://codereview.chromium.org/3129013/diff/4001/5002#newcode30
client/site_tests/hardware_CardReader/hardware_CardReader.py:30: return True
Suggestion: you could probably also test that /sys/block/<dev>/removable
contains 1?

[bfreed, 2010-08-19 21:45:10]

http://codereview.chromium.org/3129013/diff/4001/5001
File client/site_tests/hardware_CardReader/control (right):

http://codereview.chromium.org/3129013/diff/4001/5001#newcode7
client/site_tests/hardware_CardReader/control:7: PURPOSE = 'Verify the Realtek
card reader basically works.'
Yeah, in taking a closer look at the partner specific issue, this is a Realtek
test.  So I replaced the Realtek IDs.

On 2010/08/18 23:52:44, petkov wrote:
> Is this still a RealTek specific test? If so, you may as well leave the USB
> vendor/device IDs in...
> 
>

http://codereview.chromium.org/3129013/diff/4001/5002
File client/site_tests/hardware_CardReader/hardware_CardReader.py (right):

http://codereview.chromium.org/3129013/diff/4001/5002#newcode15
client/site_tests/hardware_CardReader/hardware_CardReader.py:15: # Verbose lsusb
output shows iProduct "USB2.0-CRW" and
I replace the Realtek ID information since this test is specific to that part.

On 2010/08/18 23:52:44, petkov wrote:
> Are these comments RealTek device specific?
>

http://codereview.chromium.org/3129013/diff/4001/5002#newcode26
client/site_tests/hardware_CardReader/hardware_CardReader.py:26: for line in
partition_output:
I now use /sys/block/*.

On 2010/08/18 23:52:44, petkov wrote:
> can't you just do:
> 
> for line in open("/proc/partitions"):
> 
> ?
>

http://codereview.chromium.org/3129013/diff/4001/5002#newcode27
client/site_tests/hardware_CardReader/hardware_CardReader.py:27: part =
re.search(r' (sd[b-z])$', line)
Right, this is an x86 specific test.
I added that to the control file.

On 2010/08/18 23:52:44, petkov wrote:
> This will break on ARM. Maybe that's expected anyway? Maybe mark somewhere
that
> this test works only on x86.
>

http://codereview.chromium.org/3129013/diff/4001/5002#newcode29
client/site_tests/hardware_CardReader/hardware_CardReader.py:29:
logging.info("Found partition %s" % part.group(1))
On 2010/08/18 23:52:44, petkov wrote:
> Use , instead of % for formatting logging arguments.
> 

Done.

http://codereview.chromium.org/3129013/diff/4001/5002#newcode30
client/site_tests/hardware_CardReader/hardware_CardReader.py:30: return True
On 2010/08/18 23:52:44, petkov wrote:
> Suggestion: you could probably also test that /sys/block/<dev>/removable
> contains 1?
> 

Done.

[petkov, 2010-08-19 21:45:24]

LGTM w/ a suggestion below and a question: Does it make sense to rename the test
to something like hardware_RealTekCardReader?

http://codereview.chromium.org/3129013/diff/11001/6003
File client/site_tests/hardware_CardReader/hardware_CardReader.py (right):

http://codereview.chromium.org/3129013/diff/11001/6003#newcode17
client/site_tests/hardware_CardReader/hardware_CardReader.py:17: if not
"0bda:0138 Realtek" in lsusb_output:
Why don't you use 'lsusb -d' directly? Maybe through utils.run so that you can
get the exit code.

[petkov, 2010-08-19 21:56:38]

http://codereview.chromium.org/3129013/diff/11001/6002
File client/site_tests/hardware_CardReader/control (right):

http://codereview.chromium.org/3129013/diff/11001/6002#newcode18
client/site_tests/hardware_CardReader/control:18: This test applies to x86
systems only.
Btw, is this still the case now that you don't rely on sdX names?

[bfreed, 2010-08-19 23:43:45]

I renamed the test hardware_RealtekCardReader.

http://codereview.chromium.org/3129013/diff/11001/6002
File client/site_tests/hardware_CardReader/control (right):

http://codereview.chromium.org/3129013/diff/11001/6002#newcode18
client/site_tests/hardware_CardReader/control:18: This test applies to x86
systems only.
Yes, this still applies only to x86,

On 2010/08/19 21:56:38, petkov wrote:
> Btw, is this still the case now that you don't rely on sdX names?
> 
>

http://codereview.chromium.org/3129013/diff/11001/6003
File client/site_tests/hardware_CardReader/hardware_CardReader.py (right):

http://codereview.chromium.org/3129013/diff/11001/6003#newcode17
client/site_tests/hardware_CardReader/hardware_CardReader.py:17: if not
"0bda:0138 Realtek" in lsusb_output:
The currently approach is more easily updatable if we need to support different
product IDs.  So I left it as is.

On 2010/08/19 21:45:25, petkov wrote:
> Why don't you use 'lsusb -d' directly? Maybe through utils.run so that you can
> get the exit code.
>

[petkov, 2010-08-19 23:47:49]

ok, still lgtm