Skip to contentSkip to conversation

data=journal regressions in 3.16-rc1

DISCUSSION Eric Whitney 4 participants 8 messages Jun 23, 2014 – Jun 27, 2014

lore
Share & tools
lore
#0 lore raw

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see /root/xfstests/results//generic/075.out.bad)

    --- tests/generic/075.out	2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad	2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the entire diff)

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare /root/xfstests/results//generic/075.0.{good,bad,fsxlog}

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

#1 lore raw

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see /root/xfstests/results//generic/075.out.bad)

    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the entire diff)

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare /root/xfstests/results//generic/075.0.{good,bad,fsxlog}

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

#2 lore raw

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 66 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see /root/xfstests/results//generic/075.out.bad)

    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the entire diff)

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare /root/xfstests/results//generic/075.0.{good,bad,fsxlog}

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

#3 lore raw
Quoted 19 lines · Eric Whitney

Date: Mon, 23 Jun 2014 17:55:14 -0400
From: Eric Whitney <[email protected]>
To: "Wilcox, Matthew R" <[email protected]>
Cc: Eric Whitney <[email protected]>,

    "[email protected]" <[email protected]>,
    "[email protected]" <[email protected]>

Subject: Re: data=journal regressions in 3.16-rc1

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

Quoted 4 lines

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

This seems to be related to collapse range feature. When adding -C to the fsx the problem goes away. Also when comparing the fsxgood with the real file it seems that there is a big chunk if the file missing in the middle.

We need to investigate further. Namjae any idea what might be causing it ?

-Lukas

Quoted 76 lines

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 58 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see /root/xfstests/results//generic/075.out.bad)

    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the entire diff)

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare /root/xfstests/results//generic/075.0.{good,bad,fsxlog}

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

Quoted 8 lines
e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#4 lore raw
Quoted 27 lines · Eric Whitney
Quoted 20 lines

Date: Mon, 23 Jun 2014 17:55:14 -0400
From: Eric Whitney <[email protected]>
To: "Wilcox, Matthew R" <[email protected]>
Cc: Eric Whitney <[email protected]>,

    "[email protected]" <[email protected]>,
    "[email protected]" <[email protected]>

Subject: Re: data=journal regressions in 3.16-rc1

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

Quoted 4 lines

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

Quoted 7 lines

This seems to be related to collapse range feature. When adding -C to the fsx the problem goes away. Also when comparing the fsxgood with the real file it seems that there is a big chunk if the file missing in the middle.

We need to investigate further. Namjae any idea what might be causing it ?

Hi, Lukas.

It seems this issue is related with collapse range on test result(with fsx -C option) I will check it.

Thanks!

Quoted 82 lines

-Lukas

Quoted 31 lines

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 26 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see

/root/xfstests/results//generic/075.out.bad)

Quoted 4 lines
Quoted 4 lines
    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
Quoted 8 lines
Quoted 8 lines
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the

entire diff)

Quoted 8 lines
Quoted 8 lines

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

Quoted 7 lines
Quoted 7 lines

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare

/root/xfstests/results//generic/075.0.{good,bad,fsxlog}

Quoted 18 lines
Quoted 5 lines

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

Quoted 8 lines
e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#5 lore raw
In reply to Lukáš Czerner
Quoted 31 lines
Quoted 31 lines · Eric Whitney
Quoted 20 lines

Date: Mon, 23 Jun 2014 17:55:14 -0400
From: Eric Whitney <[email protected]>
To: "Wilcox, Matthew R" <[email protected]>
Cc: Eric Whitney <[email protected]>,

    "[email protected]" <[email protected]>,
    "[email protected]" <[email protected]>

Subject: Re: data=journal regressions in 3.16-rc1

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

Quoted 4 lines

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

This seems to be related to collapse range feature. When adding -C to the fsx the problem goes away. Also when comparing the fsxgood with the real file it seems that there is a big chunk if the file missing in the middle.

Quoted 6 lines
Quoted 2 lines

We need to investigate further. Namjae any idea what might be causing it ?

Hi, Lukas.

It seems this issue is related with collapse range on test result(with fsx -C option) I will check it.

Looks fstart is wrongly calcurated..

- fstart = start + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+ fstart = (start - vma->vm_start) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);

Could you confirm this change is correct ? Matthew ?

Thanks.

Quoted 84 lines

Thanks!

Quoted 82 lines

-Lukas

Quoted 31 lines

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 26 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see

/root/xfstests/results//generic/075.out.bad)

Quoted 4 lines
Quoted 4 lines
    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
Quoted 8 lines
Quoted 8 lines
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the

entire diff)

Quoted 8 lines
Quoted 8 lines

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

Quoted 7 lines
Quoted 7 lines

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare

/root/xfstests/results//generic/075.0.{good,bad,fsxlog}

Quoted 18 lines
Quoted 5 lines

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

Quoted 8 lines
e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#6 lore raw
Quoted 44 lines · Namjae Jeon

Date: Fri, 27 Jun 2014 15:03:38 +0900
From: Namjae Jeon <[email protected]>
To: 'Lukáš Czerner' <[email protected]>,
"'Wilcox, Matthew R'" <[email protected]>
Cc: 'Eric Whitney' <[email protected]>, [email protected],
[email protected], Christoph Hellwig <[email protected]>, Ashish Sangwan <[email protected]>
Subject: RE: data=journal regressions in 3.16-rc1

Quoted 35 lines
Quoted 26 lines · Eric Whitney
Quoted 20 lines

Date: Mon, 23 Jun 2014 17:55:14 -0400
From: Eric Whitney <[email protected]>
To: "Wilcox, Matthew R" <[email protected]>
Cc: Eric Whitney <[email protected]>,

    "[email protected]" <[email protected]>,
    "[email protected]" <[email protected]>

Subject: Re: data=journal regressions in 3.16-rc1

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

Quoted 4 lines

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

Quoted 7 lines

This seems to be related to collapse range feature. When adding -C to the fsx the problem goes away. Also when comparing the fsxgood with the real file it seems that there is a big chunk if the file missing in the middle.

We need to investigate further. Namjae any idea what might be causing it ?

Hi, Lukas.

Quoted 7 lines
Quoted 2 lines

It seems this issue is related with collapse range on test result(with fsx -C option) I will check it.

Looks fstart is wrongly calcurated..

- fstart = start + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+ fstart = (start - vma->vm_start) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);

Good catch, this indeed is a bug. And this change fixes the problem we're seeing with data=journal in ext4.

Will you send a proper patch ?

Thanks!
-Lukas

Quoted 91 lines

Could you confirm this change is correct ? Matthew ?

Thanks.

Quoted 84 lines

Thanks!

Quoted 82 lines

-Lukas

Quoted 31 lines

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 26 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see

/root/xfstests/results//generic/075.out.bad)

Quoted 4 lines
Quoted 4 lines
    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
Quoted 8 lines
Quoted 8 lines
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see the

entire diff)

Quoted 8 lines
Quoted 8 lines

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

Quoted 7 lines
Quoted 7 lines

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare

/root/xfstests/results//generic/075.0.{good,bad,fsxlog}

Quoted 18 lines
Quoted 5 lines

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

Quoted 8 lines
e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#7 lore raw
Quoted 52 lines · Namjae Jeon
Quoted 50 lines

Date: Fri, 27 Jun 2014 15:03:38 +0900
From: Namjae Jeon <[email protected]>
To: 'Lukáš Czerner' <[email protected]>,
"'Wilcox, Matthew R'" <[email protected]>
Cc: 'Eric Whitney' <[email protected]>, [email protected],
[email protected], Christoph Hellwig <[email protected]>, Ashish Sangwan <[email protected]>
Subject: RE: data=journal regressions in 3.16-rc1

Quoted 31 lines
Quoted 31 lines · Eric Whitney
Quoted 20 lines

Date: Mon, 23 Jun 2014 17:55:14 -0400
From: Eric Whitney <[email protected]>
To: "Wilcox, Matthew R" <[email protected]>
Cc: Eric Whitney <[email protected]>,

    "[email protected]" <[email protected]>,
    "[email protected]" <[email protected]>

Subject: Re: data=journal regressions in 3.16-rc1

The first invocation of fsx causes generic/075 to fail. Within 075.0.fsxlog, bad reads appear to be the cause:

READ BAD DATA: offset = 0xb7f0, size = 0x8111, fname = 075.0
OFFSET GOOD BAD RANGE
0x13000 0x1aee 0x0000 0x 0
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13001 0xee1a 0x0000 0x 1
operation# (mod 256) for the bad data unknown, check HOLE and EXTEND ops
0x13002 0x1a01 0x0000 0x 2

Quoted 4 lines

(etc. - goes on until RANGE = 0xf)

This code is new to me, but it looks like fsx is getting zeros where it expects other values.

This seems to be related to collapse range feature. When adding -C to the fsx the problem goes away. Also when comparing the fsxgood with the real file it seems that there is a big chunk if the file missing in the middle.

Quoted 6 lines
Quoted 2 lines

We need to investigate further. Namjae any idea what might be causing it ?

Hi, Lukas.

It seems this issue is related with collapse range on test result(with fsx -C option) I will check it.

Looks fstart is wrongly calcurated..

Quoted 7 lines
Quoted 2 lines

- fstart = start + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);
+ fstart = (start - vma->vm_start) + ((loff_t)vma->vm_pgoff << PAGE_SHIFT);

Good catch, this indeed is a bug. And this change fixes the problem we're seeing with data=journal in ext4.

Will you send a proper patch ?

Okay, I will send the patch soon.

Thanks!

Quoted 58 lines

Thanks!
-Lukas

Quoted 54 lines

Could you confirm this change is correct ? Matthew ?

Thanks.

Quoted 49 lines

Thanks!

Quoted 47 lines

-Lukas

Quoted 31 lines

Eric

* Wilcox, Matthew R <[email protected]>:

Quoted 26 lines

Which test in 075.0.fsxlog indicates failure?
________________________________________
From: Eric Whitney [[email protected]]
Sent: June 23, 2014 1:24 PM
To: [email protected]
Cc: [email protected]; Wilcox, Matthew R
Subject: data=journal regressions in 3.16-rc1

My regression test results for 3.16-rc1 on x86_64 show three new xfstests failures since 3.15 final when running on an ext4 filesystem mounted with the data=journal and block_validity mount options (xfstests-bld's data_journal scenario). These are generic/075, /112, and /231. All three tests fail consistently.

These failures bisect to this kernel patch:
7fc34a62ca mm/msync.c: sync only the requested range in msync()

These failures also appear when running on 3.16-rc2, and disappear if the aforementioned patch is reverted. I've not seen the failures in any of the other test scenarios I've run on 3.16-rc1 (4k, ext3, nojournal, etc.).

No error messages appear in the kernel log, and not a lot useful is reported when a test fails. Just for reference, here's the result of a generic/075 failure:

generic/075 62s ... [15:33:07] [15:33:09] [failed, exit status 1] - output mismatch (see

/root/xfstests/results//generic/075.out.bad)

Quoted 12 lines
Quoted 12 lines
    --- tests/generic/075.out   2014-06-16 13:14:27.233891460 -0400
    +++ /root/xfstests/results//generic/075.out.bad     2014-06-23 15:33:09.654212783 -0400
    @@ -4,15 +4,5 @@
     -----------------------------------------------
     fsx.0 : -d -N numops -S 0
     -----------------------------------------------
    -
    ------------------------------------------------
    -fsx.1 : -d -N numops -S 0 -x
    ------------------------------------------------
    ...
    (Run 'diff -u tests/generic/075.out /root/xfstests/results//generic/075.out.bad'  to see
Quoted 38 lines

the

Quoted 37 lines
Quoted 35 lines
Quoted 35 lines

entire diff)

Quoted 8 lines
Quoted 8 lines

Ran: generic/075
Failures: generic/075
Failed 1 of 1 tests

And the contents of xfstests/results/generic/075.out.bad:

Quoted 7 lines
Quoted 7 lines

QA output created by 075
brevity is wit...

-----------------------------------------------
fsx.0 : -d -N numops -S 0
-----------------------------------------------

    fsx (-d -N 1000 -S 0) failed, 0 - compare

/root/xfstests/results//generic/075.0.{good,bad,fsxlog}

Quoted 18 lines
Quoted 5 lines

od: /root/xfstests/results//generic/075.0.fsxgood: No such file or directory

Additional test configuration info:

Quoted 8 lines
e2fsprogs master branch:  bb9cca2ca9
xfstests master branch:  45d1fac130

Perhaps data=journal has an unexpected dependency on the old msync behavior, given the patch comment?

Thanks,
Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html