Skip to contentSkip to conversation

Question on xfstest 274

DISCUSSION Eric Sandeen 2 participants 3 messages Jan 27, 2012 – Jan 30, 2012

lore
Share & tools
lore
#0 lore raw

Hi, I was wondering if you could describe your intent for 274 just a little more.

The test says:

# preallocation test

but that is the only comment other than copyright. ;)

I see that it does:

# Make a 1G fs
# Create a single 4k file
# Allocate 1M past the EOF on that file
# Completely fill remaining space, using 2 other files # Write 8k past EOF on the original file which has blocks past EOF

but I am not certain what you are testing. I presume that you are testing the fact that the 1M past EOF should be truncated, freeing up space, and allowing the 8k write to succeed? Is that right?

However, at least on ext4 I noticed that the "fill the fs" stage does not succeed; perhaps that should be tested as well.

I can do some similar work on this like I did for 275, but I need to be sure I understand your original intent for the test, first.

Thanks,
-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#1 lore raw
Quoted 7 lines · Eric Sandeen

Hi, I was wondering if you could describe your intent for 274 just a little more.

The test says:

# preallocation test

but that is the only comment other than copyright. ;)

Maybe the comment is too simple :)

Quoted 9 lines

I see that it does:

# Make a 1G fs
# Create a single 4k file
# Allocate 1M past the EOF on that file
# Completely fill remaining space, using 2 other files # Write 8k past EOF on the original file which has blocks past EOF

Quoted 4 lines

but I am not certain what you are testing. I presume that you are testing the fact that the 1M past EOF should be truncated, freeing up space, and allowing the 8k write to succeed? Is that right?

Actually I write 8K at the end of 4K(seek=1), and what I want to test is the 8K-write should be succeed because the preallocation 1M. At the mean while, the file should not be truncated.

Thanks,
wubo

Quoted 10 lines

However, at least on ext4 I noticed that the "fill the fs" stage does not succeed; perhaps that should be tested as well.

I can do some similar work on this like I did for 275, but I need to be sure I understand your original intent for the test, first.

Thanks,
-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#2 lore raw
Quoted 20 lines · WuBo
Quoted 7 lines · Eric Sandeen

Hi, I was wondering if you could describe your intent for 274 just a little more.

The test says:

# preallocation test

but that is the only comment other than copyright. ;)

Maybe the comment is too simple :)

Quoted 9 lines

I see that it does:

# Make a 1G fs
# Create a single 4k file
# Allocate 1M past the EOF on that file
# Completely fill remaining space, using 2 other files # Write 8k past EOF on the original file which has blocks past EOF

Quoted 8 lines
Quoted 4 lines

but I am not certain what you are testing. I presume that you are testing the fact that the 1M past EOF should be truncated, freeing up space, and allowing the 8k write to succeed? Is that right?

Actually I write 8K at the end of 4K(seek=1), and what I want to test is the 8K-write should be succeed because the preallocation 1M. At the mean while, the file should not be truncated.

Ah, right - I missed the conv=notrunc sorry. Ok, that all makes sense. I think I will try to clean up the test just a bit.

Thanks,
-Eric

Quoted 10 lines

Thanks,
wubo

Quoted 7 lines

However, at least on ext4 I noticed that the "fill the fs" stage does not succeed; perhaps that should be tested as well.

I can do some similar work on this like I did for 275, but I need to be sure I understand your original intent for the test, first.

Quoted 9 lines
Quoted 3 lines

Thanks,
-Eric

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs