two raid consistency bugs
Hello btrfs devs,
I found two raid-related bugs in btrfs. Two test cases are provided.
BUG 1. fstests: btrfs/348: test ambiguous generation handling on raid1 profile
This test simulates a ambiguous generation which can be caused by, for example, two successive power failures. Please note this is not related to degraded mounts or nodatacow.
This bug may affect several raid levels, take raid1 (say disk A and B) as an example:
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
At second power failure during a different transaction N, nothing is written to disk A, but metadata trees and super is committed to disk B.
This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
I'd like to discuss possible solutions:
1) Turn btrfs metadata trees into merkle trees
This is the most CoW flavor solution. With a strong checksum algorithm, merkle tree can gurantee ambiguous is detected and fixed. However it is difficult to implement because metadata checksumming is done at bio time (not at tree manipulation time), also on-disk format is changed.
2) A write-intent bitmap
This is the traditional solution to raid consistency problem. This can also helps resync nodatacow data. However it seems there is an anandoned series of write-intent patches in btrfs mailing list.
3) Generation redzone
Introduce a generation redzone value to superblock, which is updated at mount time, for example:
1st mount: generation=N redzone=N -> next generation id is N+1
2nd mount: generation=N redzone=N+1 -> next generation id is N+2 3rd mount: generation=N redzone=N+2 -> next generation id is N+3
However it seems there can be infinte TRANS_STATE_UNBLOCKED transactions, so it's hard to decide how may delta should we add to redzone value to get next generation id.
Also, this redzone value is not applicable to tree-log, so BUG 2 (see below) can't be solved.
4) Dirty workaround
Pin metadata to a dedicated device (which can be LVM raid1) and ask user to run single metadata profile. This is too dirty and should not used.
BUG 2. fstests: btrfs/349: test if latest tree-log is choosen at mount time on raid1 profile
This test simulate a scenario that tree-log only exists in secondary device, and test if latest tree-log is choosen at mount time. Currently, the tree-log in the device with lowest devid is choosen. So fsync'ed data may loss if tree-log only exists in secondary device.
A proposed draft fix is available at:
https://lore.kernel.org/linux-btrfs/[email protected]/
However, this draft fix also suffers from the above ambiguous generation problem.
Zhang Boyang
PATCH 1/2 fstests: btrfs/348: test ambiguous generation handling on raid1 profile
This test simulates a ambiguous generation which can be caused by, for example, two successive power failures.
Please note this is not related to degraded mounts or nodatacow.
Unfortunately, there is currently no fix for this problem.
--- tests/btrfs/348 | 79 +++++++++++++++++++++++++ tests/btrfs/348.out | 141 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 220 insertions(+) create mode 100755 tests/btrfs/348 create mode 100644 tests/btrfs/348.out
tests/btrfs/348 new file +79 -0
@@ -0,0 +1,79 @@ +#! /bin/bash +# SPDX-License-Identifier: GPL-2.0 +# Copyright (c) 2026 Zhang Boyang. All Rights Reserved. +# +# FS QA Test 348 +# +# Test btrfs ambiguous generation handling on raid1 profile. +# This can be caused by, for example, two successive power failures. +# +. ./common/preamble +_begin_fstest auto quick read_repair raid + +. ./common/filter +. ./common/fail_make_request +. ./common/dmdust + +_require_fail_make_request +_require_scratch_dev_pool 2 +_require_dm_target dust + +_scratch_dev_pool_get 2 +_check_minimal_fs_size $(( 1024 * 1024 * 1024 )) +# a strong checksum like sha256 can't protect you +_scratch_pool_mkfs "-m raid1 -d raid1 -b 1G -n 4K --csum sha256" >> $seqres.full 2>&1 +dev2=`echo $SCRATCH_DEV_POOL | $AWK_PROG '{print $2}'` +_allow_fail_make_request + +echo "Step 1: prepare filesystem" +_scratch_mount -o notreelog +# create a 2-level FS_TREE, this is not strictly necessary +# however this helps improve test result reproducibility (see below) +for ((i = 0; i < 1000; i++)); do + $XFS_IO_PROG -f -c "pwrite -S 0xff 0 2K" $SCRATCH_MNT/foobar$i >> $seqres.full 2>&1 +done +# the test file should reside in a leaf which is not read at mount time +# so we can control which mirror to use by using _btrfs_buffered_read_on_mirror +testfile="$SCRATCH_MNT/foobar42" +_scratch_unmount + +echo "Step 2: simulate failure #1" +_init_dust +_mount_dust -o notreelog +# simulate dev1 disappears just before writing superblock +for ((i = 0; i < 4096 / 512; i++)); do + $DMSETUP_PROG message $DUST_NAME 0 addbadblock $((0x10000 / 512 + $i)) 255 +done +$DMSETUP_PROG message $DUST_NAME 0 enable +# simulate dev2 disappears before writing anything +_bdev_fail_make_request $dev2 1 +# this write should fail +$XFS_IO_PROG -f -c "pwrite -W -S 0xaa 0 2K" "$testfile" | _filter_xfs_io +_cleanup_dust +_bdev_fail_make_request $dev2 0 + +echo "Step 3: simulate failure #2" +_scratch_mount -o notreelog +# simulate dev1 disappears before writing anything +_bdev_fail_make_request $SCRATCH_DEV 1 +# this write should success +$XFS_IO_PROG -f -c "pwrite -W -S 0xbb 0 2K" "$testfile" | _filter_xfs_io +# simulate dev2 disappears after writing superblock +_bdev_fail_make_request $dev2 1 +_scratch_unmount +_bdev_fail_make_request $SCRATCH_DEV 0 +_bdev_fail_make_request $dev2 0 + +echo "Step 4: simulate recovery" +_scratch_mount -o notreelog +# this read should trigger resync +_btrfs_buffered_read_on_mirror 0 2 "$testfile" 0 2K +echo "Step 5: verify written data" +# read data should be 0xbb +$XFS_IO_PROG -c "pread -v 0 2K" "$testfile" | _filter_xfs_io_offset +_scratch_unmount + +_disallow_fail_make_request +_scratch_dev_pool_put +# success, all done +_exit 0
tests/btrfs/348.out new file +141 -0
141 lines not on this page · raw message
Signature · 1 line
2.53.0
PATCH 2/2 fstests: btrfs/349: test if latest tree-log is choosen at mount time on raid1 profile
In reply to Zhang BoyangThis test simulate a scenario that tree-log only exists in secondary device, and test if latest tree-log is choosen at mount time.
Currently, the tree-log in the device with lowest devid is choosen. So fsync'ed data may loss if tree-log only exists in secondary device.
--- tests/btrfs/349 | 53 +++++++++++++++++++++++++++++++++++++++++++++ tests/btrfs/349.out | 14 ++++++++++++ 2 files changed, 67 insertions(+) create mode 100755 tests/btrfs/349 create mode 100644 tests/btrfs/349.out
tests/btrfs/349 new file +53 -0
@@ -0,0 +1,53 @@ +#! /bin/bash +# SPDX-License-Identifier: GPL-2.0 +# Copyright (c) 2026 Zhang Boyang. All Rights Reserved. +# +# FS QA Test 349 +# +# Test if latest tree-log is choosen at mount time on raid1 profile. +# +. ./common/preamble +_begin_fstest auto quick read_repair raid + +. ./common/filter +. ./common/fail_make_request + +_require_fail_make_request +_require_scratch_dev_pool 2 + +_scratch_dev_pool_get 2 +_check_minimal_fs_size $(( 1024 * 1024 * 1024 )) +# a strong checksum like sha256 can't protect you +_scratch_pool_mkfs "-m raid1 -d raid1 -b 1G --csum sha256" >> $seqres.full 2>&1 +dev2=`echo $SCRATCH_DEV_POOL | $AWK_PROG '{print $2}'` +_allow_fail_make_request + +echo "Step 1: prepare filesystem" +_scratch_mount -o treelog +$XFS_IO_PROG -f -c "pwrite -W -S 0xaa 0 8" $SCRATCH_MNT/foobar | _filter_xfs_io +_scratch_unmount + +echo "Step 2: simulate failure" +_scratch_mount -o treelog +# simulate dev1 disappears before writing anything +_bdev_fail_make_request $SCRATCH_DEV 1 +# this write should success +$XFS_IO_PROG -f -c "pwrite -W -S 0xbb 0 8" $SCRATCH_MNT/foobar | _filter_xfs_io +# simulate dev2 disappears after writing TREE_LOG +_bdev_fail_make_request $dev2 1 +_scratch_unmount +_bdev_fail_make_request $SCRATCH_DEV 0 +_bdev_fail_make_request $dev2 0 + +echo "Step 3: simulate recovery" +# this mount should start tree-log replay +_scratch_mount -o treelog +echo "Step 4: verify written data" +# read data should be 0xbb +$XFS_IO_PROG -c "pread -v 0 8" $SCRATCH_MNT/foobar | _filter_xfs_io_offset +_scratch_unmount + +_disallow_fail_make_request +_scratch_dev_pool_put +# success, all done +_exit 0
tests/btrfs/349.out new file +14 -0
@@ -0,0 +1,14 @@ +QA output created by 349 +Allow global fail_make_request feature +Step 1: prepare filesystem +wrote 8/8 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +Step 2: simulate failure +wrote 8/8 bytes at offset 0 +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +Step 3: simulate recovery +Step 4: verify written data +XXXXXXXX: bb bb bb bb bb bb bb bb ........ +read 8/8 bytes +XXX Bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec) +Disallow global fail_make_request feature
Signature · 1 line
2.53.0
Quoted 11 lines · Zhang Boyang
Hello btrfs devs,
I found two raid-related bugs in btrfs. Two test cases are provided.
BUG 1. fstests: btrfs/348: test ambiguous generation handling on raid1 profile
This test simulates a ambiguous generation which can be caused by, for example, two successive power failures. Please note this is not related to degraded mounts or nodatacow.
Quoted 8 lines
This bug may affect several raid levels, take raid1 (say disk A and B) as an example:
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
At second power failure during a different transaction N,
If it's a different transaction, why it will still have the same transid N?
Quoted 5 lines
nothing is
written to disk A, but metadata trees and super is committed to disk B.This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
Hi,
Quoted 8 lines · Qu Wenruo
Quoted 5 lines
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
At second power failure during a different transaction N,
If it's a different transaction, why it will still have the same transid N?
Because the power failure occurred just before writing superblock (transid N). The superblock on disk still has transid N-1.
After reboot, looking at superblock which transid is N-1, btrfs has no idea of transid N existed previously, so it uses transid N for new transcation.
Quoted 5 lines
Quoted 5 lines
nothing is
written to disk A, but metadata trees and super is committed to disk B.This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
By the way, I came up another solution:
If generation mismatch between devices (or log-tree mismatch) is detected at mount time, set a dirty flag in superblock on device which is behind.
If dirty flag is set for a device, disable read load balancing for that device. So (meta)data only read from latest device(s).
If dirty flag is detected, ask user to run a scrub. The dirty flag is cleared after a successful scrub.
Zhang Boyang
Quoted 7 lines · Zhang Boyang
Hi,
Quoted 4 lines · Qu Wenruo
Quoted 4 lines
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
At second power failure during a different transaction N,
If it's a different transaction, why it will still have the same transid N?
Because the power failure occurred just before writing superblock (transid N). The superblock on disk still has transid N-1.
After reboot, looking at superblock which transid is N-1, btrfs has no idea of transid N existed previously, so it uses transid N for new transcation.
Then there should be no problem at least at the next mount after the power loss.
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
nothing is
written to disk A, but metadata trees and super is committed to disk B.This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
At the next mount, btrfs should detect device B has the latest super block, and use that as the super block to mount.
Since metadata are all written to device B, even disk A may have some stale tree blocks with transid N, stale tree blocks still need to meet other conditions like root owner, level, first key checks.
I won't say that's impossible, and won't say we shouldn't do anything to address it, but this is a variant of the split brain problems mentioned in the past.
I strongly recommend to find out that thread and check if any of the ideas are explored before and if they have their limits.
Quoted 6 lines
By the way, I came up another solution:
If generation mismatch between devices (or log-tree mismatch) is detected at mount time, set a dirty flag in superblock on device which is behind.
If dirty flag is set for a device, disable read load balancing for that device. So (meta)data only read from latest device(s).
What if some metadata only arrives at that stale device, but not completely reached the good device?
That kills the only chance to get the good metadata mirror.
And this won't solve the split brain situation either.
Quoted 5 lines
If dirty flag is detected, ask user to run a scrub. The dirty flag is cleared after a successful scrub.
Zhang Boyang
Hi,
Quoted 25 lines · Qu Wenruo
Quoted 7 lines · Zhang Boyang
Hi,
Quoted 4 lines · Qu Wenruo
Quoted 4 lines
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
At second power failure during a different transaction N,
If it's a different transaction, why it will still have the same transid N?
Because the power failure occurred just before writing superblock (transid N). The superblock on disk still has transid N-1.
After reboot, looking at superblock which transid is N-1, btrfs has no idea of transid N existed previously, so it uses transid N for new transcation.
Then there should be no problem at least at the next mount after the power loss.
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
nothing is
written to disk A, but metadata trees and super is committed to disk B.
Quoted 9 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
At the next mount, btrfs should detect device B has the latest super block, and use that as the super block to mount.
Since metadata are all written to device B, even disk A may have some stale tree blocks with transid N, stale tree blocks still need to meet other conditions like root owner, level, first key checks.
I won't say that's impossible, and won't say we shouldn't do anything to address it, but this is a variant of the split brain problems mentioned in the past.
I think this is a very special variant of split brain problem. This bug can occur even if no degraded mounts are involved. All devices are presented to btrfs at every mounts. Personally I don't like to call this bug as a split brain problem because I think split brain should only related to degraded mounts.
Quoted 2 lines
I strongly recommend to find out that thread and check if any of the ideas are explored before and if they have their limits.
I did read some of these threads. If split brain problems are solved, this bug can be solved as well. But I think fixing this particular bug is also beneficial.
Quoted 9 lines
Quoted 6 lines
By the way, I came up another solution:
If generation mismatch between devices (or log-tree mismatch) is detected at mount time, set a dirty flag in superblock on device which is behind.
If dirty flag is set for a device, disable read load balancing for that device. So (meta)data only read from latest device(s).
What if some metadata only arrives at that stale device, but not completely reached the good device?
Let me add another rule: always write superblocks of dirty devices after superblocks of (one of) latest devices are fully written.
Therefore, as long as there is no died device, there is at least one device not dirty so we can get golden data from it.
Of course performance may hit but we are safe.
Quoted 2 lines
That kills the only chance to get the good metadata mirror.
A recovery option can be introduced to disable the above policy.
Quoted 2 lines
And this won't solve the split brain situation either.
I'm not aiming to solve the split brain problems completely.
But waiting split brain problems to be solved is also a solution to this particular bug though.
Quoted 6 lines
Quoted 5 lines
If dirty flag is detected, ask user to run a scrub. The dirty flag is cleared after a successful scrub.
Zhang Boyang
Zhang Boyang
Quoted 38 lines · Zhang Boyang
Hi,
Quoted 25 lines · Qu Wenruo
Quoted 7 lines · Zhang Boyang
Hi,
Quoted 4 lines · Qu Wenruo
Quoted 4 lines
At first power failure during transaction N, metadata trees of generation N are written to disk A, but super is not committed. Nothing is written to disk B.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
At second power failure during a different transaction N,
If it's a different transaction, why it will still have the same transid N?
Because the power failure occurred just before writing superblock (transid N). The superblock on disk still has transid N-1.
After reboot, looking at superblock which transid is N-1, btrfs has no idea of transid N existed previously, so it uses transid N for new transcation.
Then there should be no problem at least at the next mount after the power loss.
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
nothing is
written to disk A, but metadata trees and super is committed to disk B.Quoted 9 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
This creates a ambiguous generation N in two disks. Currently btrfs can't detect this, and can lead to severe damages.
At the next mount, btrfs should detect device B has the latest super block, and use that as the super block to mount.
Since metadata are all written to device B, even disk A may have some stale tree blocks with transid N, stale tree blocks still need to meet other conditions like root owner, level, first key checks.
I won't say that's impossible, and won't say we shouldn't do anything to address it, but this is a variant of the split brain problems mentioned in the past.
Quoted 6 lines
I think this is a very special variant of split brain problem. This bug can occur even if no degraded mounts are involved. All devices are presented to btrfs at every mounts. Personally I don't like to call this bug as a split brain problem because I think split brain should only related to degraded mounts.
Quoted 2 lines
I strongly recommend to find out that thread and check if any of the ideas are explored before and if they have their limits.
I did read some of these threads. If split brain problems are solved, this bug can be solved as well. But I think fixing this particular bug is also beneficial.
I do not agree.
You're introducing more and more code just to handle some very specific corner cases.
E.g. for your particular case, you will need a very specific write situation. You're introducing a feature that is very hard to hit under most situations, but we will always bear the burden.
I am not even sure if you'll still contribute in the next 5 years, thus I won't bet my 5 cents on that this feature will be properly maintained.
To me, if you really bother this particular situation for whatever reason, just introduce a special harden mount option, that any barrier/super block write failure will mark the fs error.
That will be a much safer bet than any of your proposal.
Hi,
Quoted 4 lines · Qu Wenruo
You're introducing more and more code just to handle some very specific corner cases.
E.g. for your particular case, you will need a very specific write situation. You're introducing a feature that is very hard to hit under most situations, but we will always bear the burden.
Technically the runtime burden is small because generation mismatch is rare. If no generation mismatch is detected, no device will be marked dirty and no runtime burden is involved. On the other side, if generation mismatch is detected, you always want to run a scrub to resync your array.
Quoted 2 lines
I am not even sure if you'll still contribute in the next 5 years, thus I won't bet my 5 cents on that this feature will be properly maintained.
For maintenance burden, yes, there is maintenance burden. So I'm here to discuss which way to fix this bug is proper.
Quoted 3 lines
To me, if you really bother this particular situation for whatever reason, just introduce a special harden mount option, that any barrier/ super block write failure will mark the fs error.
That will be a much safer bet than any of your proposal.
This can be a solution, at the cost of one failing drive will render entire raid array unavailable.
Personally I like the idea of turning metadata trees into merkle trees (so I wrote it first). But since you are the maintainer, I will let you decide which way is appropriate.
Zhang Boyang
Quoted 15 lines · Zhang Boyang
Hi,
Quoted 4 lines · Qu Wenruo
You're introducing more and more code just to handle some very specific corner cases.
E.g. for your particular case, you will need a very specific write situation. You're introducing a feature that is very hard to hit under most situations, but we will always bear the burden.
Technically the runtime burden is small because generation mismatch is rare. If no generation mismatch is detected, no device will be marked dirty and no runtime burden is involved. On the other side, if generation mismatch is detected, you always want to run a scrub to resync your array.
Quoted 2 lines
I am not even sure if you'll still contribute in the next 5 years, thus I won't bet my 5 cents on that this feature will be properly maintained.
For maintenance burden, yes, there is maintenance burden. So I'm here to discuss which way to fix this bug is proper.
Quoted 7 lines
Quoted 3 lines
To me, if you really bother this particular situation for whatever reason, just introduce a special harden mount option, that any barrier/ super block write failure will mark the fs error.
That will be a much safer bet than any of your proposal.
This can be a solution, at the cost of one failing drive will render entire raid array unavailable.
Personally I like the idea of turning metadata trees into merkle trees (so I wrote it first). But since you are the maintainer, I will let you decide which way is appropriate.
I'm not the maintainer, I just default to not trust ideas/features from unproven contributors, especially when the idea/feature is not simple.
Especially considering the increased patches generated by LLMs.
So my suggestion is, if you really want to do long-term contributions, stop coming up with fancy ideas, start from basics like doing proper fstests runs with various combinations, report real bugs and fix simple ones by yourself.
Earn your name first then consider fancy ideas.
Quoted 4 lines
Zhang Boyang
Hi,
Quoted 6 lines · Qu Wenruo
Quoted 1 line
Personally I like the idea of turning metadata trees into merkle trees (so I wrote it first). But since you are the maintainer, I will let you decide which way is appropriate.
I'm not the maintainer, I just default to not trust ideas/features from unproven contributors, especially when the idea/feature is not simple.
Especially considering the increased patches generated by LLMs.
I understand your concern.
Quoted 5 lines
So my suggestion is, if you really want to do long-term contributions, stop coming up with fancy ideas, start from basics like doing proper fstests runs with various combinations, report real bugs and fix simple ones by yourself.
Earn your name first then consider fancy ideas.
That's exactly what I'm doing. I'm not requesting devs to fix this bug immediately or implement my fancy idea. I wrote two test cases and hope it will be helpful to btrfs development. It would be better if these two test cases can be merged to test suite.
I even wrote a draft patch to fix the second bug. However, the first bug is not easily solvable. So I had to discuss possible solutions. If possible, I would have written a draft patch for the first bug, too.
Zhang Boyang
Quoted 13 lines · Zhang Boyang
Hi,
Quoted 6 lines · Qu Wenruo
Quoted 1 line
Personally I like the idea of turning metadata trees into merkle trees (so I wrote it first). But since you are the maintainer, I will let you decide which way is appropriate.
I'm not the maintainer, I just default to not trust ideas/features from unproven contributors, especially when the idea/feature is not simple.
Especially considering the increased patches generated by LLMs.
I understand your concern.
Quoted 1 line
Quoted 6 lines
Quoted 4 lines
So my suggestion is, if you really want to do long-term contributions, stop coming up with fancy ideas, start from basics like doing proper fstests runs with various combinations, report real bugs and fix simple ones by yourself.
Earn your name first then consider fancy ideas.
That's exactly what I'm doing. I'm not requesting devs to fix this bug immediately or implement my fancy idea. I wrote two test cases and hope it will be helpful to btrfs development. It would be better if these two test cases can be merged to test suite.
No, we will not push for test cases without a fix.
There is already a known failure, generic/747, and at least I'm not the only one regretting pushing that.
Quoted 2 lines
I even wrote a draft patch to fix the second bug. However, the first bug is not easily solvable. So I had to discuss possible solutions. If possible, I would have written a draft patch for the first bug, too.
This is already what I call fancy ideas.
Anyway, this will mostly just get yourself ignored if you really choose to start from fancy projects. Those long existing problems are still there for a reason.
If you believe you're smarter than everyone else, then go ahead, no one is going to make you change your mind.
Quoted 3 lines
Zhang Boyang
Hi,
Quoted 6 lines · Qu Wenruo
Quoted 1 line
That's exactly what I'm doing. I'm not requesting devs to fix this bug immediately or implement my fancy idea. I wrote two test cases and hope it will be helpful to btrfs development. It would be better if these two test cases can be merged to test suite.
No, we will not push for test cases without a fix.
There is already a known failure, generic/747, and at least I'm not the only one regretting pushing that.
That's a strange policy. Is this a btrfs policy only or it applies to whole xfstests?
Quoted 8 lines
Quoted 2 lines
I even wrote a draft patch to fix the second bug. However, the first bug is not easily solvable. So I had to discuss possible solutions. If possible, I would have written a draft patch for the first bug, too.
This is already what I call fancy ideas.
Anyway, this will mostly just get yourself ignored if you really choose to start from fancy projects. Those long existing problems are still there for a reason.
Ok, the first bug is a bit fancy, but what about the second one? I don't think the second bug is hard to solve (if you ignore the first bug).
Quoted 2 lines
If you believe you're smarter than everyone else, then go ahead, no one is going to make you change your mind.
This is a strange accusation and I don't like your attitude. I never thought I am smarter than everyone else.
Zhang Boyang
Quoted 16 lines · Qu Wenruo
Quoted 12 lines · Zhang Boyang
Hi,
Quoted 6 lines · Qu Wenruo
Quoted 1 line
Personally I like the idea of turning metadata trees into merkle trees (so I wrote it first). But since you are the maintainer, I will let you decide which way is appropriate.
I'm not the maintainer, I just default to not trust ideas/features from unproven contributors, especially when the idea/feature is not simple.
Especially considering the increased patches generated by LLMs.
I understand your concern.
Quoted 3 lines
So my suggestion is, if you really want to do long-term contributions, stop coming up with fancy ideas, start from basics like doing proper fstests runs with various combinations, report real bugs and fix simple ones by yourself.
Quoted 7 lines
Quoted 3 lines
Quoted 2 lines
Earn your name first then consider fancy ideas.
That's exactly what I'm doing. I'm not requesting devs to fix this bug immediately or implement my fancy idea. I wrote two test cases and hope it will be helpful to btrfs development. It would be better if these two test cases can be merged to test suite.
No, we will not push for test cases without a fix.
There is already a known failure, generic/747, and at least I'm not the only one regretting pushing that.
Then why not delete it ?
Quoted 11 lines
Quoted 2 lines
I even wrote a draft patch to fix the second bug. However, the first bug is not easily solvable. So I had to discuss possible solutions. If possible, I would have written a draft patch for the first bug, too.
This is already what I call fancy ideas.
Anyway, this will mostly just get yourself ignored if you really choose to start from fancy projects. Those long existing problems are still there for a reason.
If you believe you're smarter than everyone else, then go ahead, no one is going to make you change your mind.
Quoted 1 line
Zhang Boyang