| Message ID | 20240916125707.127118-3-luca.stefani.ge1@gmail.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | btrfs: Don't block system suspend during fstrim | expand |
On Mon, Sep 16, 2024 at 02:56:15PM +0200, Luca Stefani wrote: > Sometimes the system isn't able to suspend because the task > responsible for trimming the device isn't able to finish in > time, especially since we have a free extent discarding phase, > which can trim a lot of unallocated space, and there is no > limits on the trim size (unlike the block group part). > > Since discard isn't a critical call it can be interrupted > at any time, in such cases we stop the trim, report the amount > of discarded bytes and return failure. > > Link: https://bugzilla.kernel.org/show_bug.cgi?id=219180 > Link: https://bugzilla.suse.com/show_bug.cgi?id=1229737 > Signed-off-by: Luca Stefani <luca.stefani.ge1@gmail.com> I went through the cancellation points, some of them don't seem to be necessary, eg. in a big loop when some function is called to do trim (extents, bitmaps) and then again does the signal and freezing check. Next, some of the functions are called from async discard and errors are not checked: btrfs_trim_block_group_bitmaps() called from btrfs_discard_workfn(). Ther's also check for signals pending in trim_bitmaps() in free-space-cache.c. Given that the space cache code is on the way out we don't necesssarily need to fix it but if the patch gets backported to older kernels it still makes sense. > --- > fs/btrfs/extent-tree.c | 23 ++++++++++++++++++++++- > 1 file changed, 22 insertions(+), 1 deletion(-) > > diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c > index 79b9243c9cd6..cef368a30731 100644 > --- a/fs/btrfs/extent-tree.c > +++ b/fs/btrfs/extent-tree.c > @@ -16,6 +16,7 @@ > #include <linux/percpu_counter.h> > #include <linux/lockdep.h> > #include <linux/crc32c.h> > +#include <linux/freezer.h> > #include "ctree.h" > #include "extent-tree.h" > #include "transaction.h" > @@ -1235,6 +1236,11 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, > return ret; > } > > +static bool btrfs_trim_interrupted(void) > +{ > + return fatal_signal_pending(current) || freezing(current); > +} > + > static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, > u64 *discarded_bytes) > { > @@ -1316,6 +1322,11 @@ static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, > start += bytes_to_discard; > bytes_left -= bytes_to_discard; > *discarded_bytes += bytes_to_discard; > + > + if (btrfs_trim_interrupted()) { > + ret = -ERESTARTSYS; > + break; > + } > } > > return ret; > @@ -6470,7 +6481,7 @@ static int btrfs_trim_free_extents(struct btrfs_device *device, u64 *trimmed) > start += len; > *trimmed += bytes; > > - if (fatal_signal_pending(current)) { > + if (btrfs_trim_interrupted()) { > ret = -ERESTARTSYS; > break; > } > @@ -6519,6 +6530,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) > > cache = btrfs_lookup_first_block_group(fs_info, range->start); > for (; cache; cache = btrfs_next_block_group(cache)) { > + if (btrfs_trim_interrupted()) { > + bg_ret = -ERESTARTSYS; > + break; > + } > + > if (cache->start >= range_end) { > btrfs_put_block_group(cache); > break; > @@ -6558,6 +6574,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) > > mutex_lock(&fs_devices->device_list_mutex); > list_for_each_entry(device, &fs_devices->devices, dev_list) { > + if (btrfs_trim_interrupted()) { > + dev_ret = -ERESTARTSYS; This one seems redundant. > + break; > + } > + > if (test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state)) > continue; > > -- > 2.46.0 >
On 17/09/24 18:24, David Sterba wrote: > On Mon, Sep 16, 2024 at 02:56:15PM +0200, Luca Stefani wrote: >> Sometimes the system isn't able to suspend because the task >> responsible for trimming the device isn't able to finish in >> time, especially since we have a free extent discarding phase, >> which can trim a lot of unallocated space, and there is no >> limits on the trim size (unlike the block group part). >> >> Since discard isn't a critical call it can be interrupted >> at any time, in such cases we stop the trim, report the amount >> of discarded bytes and return failure. >> >> Link: https://bugzilla.kernel.org/show_bug.cgi?id=219180 >> Link: https://bugzilla.suse.com/show_bug.cgi?id=1229737 >> Signed-off-by: Luca Stefani <luca.stefani.ge1@gmail.com> > > I went through the cancellation points, some of them don't seem to be > necessary, eg. in a big loop when some function is called to do trim > (extents, bitmaps) and then again does the signal and freezing check. > > Next, some of the functions are called from async discard and errors are > not checked: btrfs_trim_block_group_bitmaps() called from > btrfs_discard_workfn(). Both btrfs_trim_block_group_bitmaps and btrfs_trim_block_group_extents ret codes are never checked indeed in btrfs_discard_workfn. I'll fix that up in another CL. > > Ther's also check for signals pending in trim_bitmaps() in > free-space-cache.c. Given that the space cache code is on the way out we > don't necesssarily need to fix it but if the patch gets backported to > older kernels it still makes sense. Ah I missed this one, will fix it. There's a few more instances of fatal_signal_pending but I don't know if they should be translated or not, will focus on the one you mentioned and trim_no_bitmap which seems to do similar checks for fatal signals. > >> --- >> fs/btrfs/extent-tree.c | 23 ++++++++++++++++++++++- >> 1 file changed, 22 insertions(+), 1 deletion(-) >> >> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c >> index 79b9243c9cd6..cef368a30731 100644 >> --- a/fs/btrfs/extent-tree.c >> +++ b/fs/btrfs/extent-tree.c >> @@ -16,6 +16,7 @@ >> #include <linux/percpu_counter.h> >> #include <linux/lockdep.h> >> #include <linux/crc32c.h> >> +#include <linux/freezer.h> >> #include "ctree.h" >> #include "extent-tree.h" >> #include "transaction.h" >> @@ -1235,6 +1236,11 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, >> return ret; >> } >> >> +static bool btrfs_trim_interrupted(void) >> +{ >> + return fatal_signal_pending(current) || freezing(current); >> +} >> + >> static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, >> u64 *discarded_bytes) >> { >> @@ -1316,6 +1322,11 @@ static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, >> start += bytes_to_discard; >> bytes_left -= bytes_to_discard; >> *discarded_bytes += bytes_to_discard; >> + >> + if (btrfs_trim_interrupted()) { >> + ret = -ERESTARTSYS; >> + break; >> + } >> } >> >> return ret; >> @@ -6470,7 +6481,7 @@ static int btrfs_trim_free_extents(struct btrfs_device *device, u64 *trimmed) >> start += len; >> *trimmed += bytes; >> >> - if (fatal_signal_pending(current)) { >> + if (btrfs_trim_interrupted()) { >> ret = -ERESTARTSYS; >> break; >> } >> @@ -6519,6 +6530,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) >> >> cache = btrfs_lookup_first_block_group(fs_info, range->start); >> for (; cache; cache = btrfs_next_block_group(cache)) { >> + if (btrfs_trim_interrupted()) { >> + bg_ret = -ERESTARTSYS; >> + break; >> + } >> + >> if (cache->start >= range_end) { >> btrfs_put_block_group(cache); >> break; >> @@ -6558,6 +6574,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) >> >> mutex_lock(&fs_devices->device_list_mutex); >> list_for_each_entry(device, &fs_devices->devices, dev_list) { >> + if (btrfs_trim_interrupted()) { >> + dev_ret = -ERESTARTSYS; > > This one seems redundant. > >> + break; >> + } >> + >> if (test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state)) >> continue; >> >> -- >> 2.46.0 >>
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 79b9243c9cd6..cef368a30731 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -16,6 +16,7 @@ #include <linux/percpu_counter.h> #include <linux/lockdep.h> #include <linux/crc32c.h> +#include <linux/freezer.h> #include "ctree.h" #include "extent-tree.h" #include "transaction.h" @@ -1235,6 +1236,11 @@ static int remove_extent_backref(struct btrfs_trans_handle *trans, return ret; } +static bool btrfs_trim_interrupted(void) +{ + return fatal_signal_pending(current) || freezing(current); +} + static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, u64 *discarded_bytes) { @@ -1316,6 +1322,11 @@ static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len, start += bytes_to_discard; bytes_left -= bytes_to_discard; *discarded_bytes += bytes_to_discard; + + if (btrfs_trim_interrupted()) { + ret = -ERESTARTSYS; + break; + } } return ret; @@ -6470,7 +6481,7 @@ static int btrfs_trim_free_extents(struct btrfs_device *device, u64 *trimmed) start += len; *trimmed += bytes; - if (fatal_signal_pending(current)) { + if (btrfs_trim_interrupted()) { ret = -ERESTARTSYS; break; } @@ -6519,6 +6530,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) cache = btrfs_lookup_first_block_group(fs_info, range->start); for (; cache; cache = btrfs_next_block_group(cache)) { + if (btrfs_trim_interrupted()) { + bg_ret = -ERESTARTSYS; + break; + } + if (cache->start >= range_end) { btrfs_put_block_group(cache); break; @@ -6558,6 +6574,11 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range) mutex_lock(&fs_devices->device_list_mutex); list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (btrfs_trim_interrupted()) { + dev_ret = -ERESTARTSYS; + break; + } + if (test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state)) continue;
Sometimes the system isn't able to suspend because the task responsible for trimming the device isn't able to finish in time, especially since we have a free extent discarding phase, which can trim a lot of unallocated space, and there is no limits on the trim size (unlike the block group part). Since discard isn't a critical call it can be interrupted at any time, in such cases we stop the trim, report the amount of discarded bytes and return failure. Link: https://bugzilla.kernel.org/show_bug.cgi?id=219180 Link: https://bugzilla.suse.com/show_bug.cgi?id=1229737 Signed-off-by: Luca Stefani <luca.stefani.ge1@gmail.com> --- fs/btrfs/extent-tree.c | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-)