[PATCH V2 1/1] mdadm/super1: Add MD_FEATURE_RAID0_LAYOUT for all raid0 after kernel v5.4

[PATCH V2 1/1] mdadm/super1: Add MD_FEATURE_RAID0_LAYOUT for all raid0 after kernel v5.4

[Xiao Ni]

After and include kernel v5.4, it adds one feature bit MD_FEATURE_RAID0_LAYOUT.
It must need to specify a layout for raid0 with more than one zone. But for
raid0 with one zone, in fact it also has a defalut layout.

Now for raid0 with one zone, *unknown* layout can be seen when running mdadm -D
command. It's the reason that mdadm doesn't set MD_FEATURE_RAID0_LAYOUT for
raid0 with one zone. Then in kernel space, super_1_validate sets mddev->layout
to -1 because of no MD_FEATURE_RAID0_LAYOUT. In fact, in raid0 io path, it
uses the default layout. So in fact after/include kernel v5.4, all raid0 device
have layout.

Fixes: 329dfc28debb ('Create: add support for RAID0 layouts.')
Signed-off-by: Xiao Ni <xni@redhat.com>
---
 super1.c | 21 ++-------------------
 1 file changed, 2 insertions(+), 19 deletions(-)

diff --git a/super1.c b/super1.c
index 856b02082662..653a2ea6c0e4 100644
--- a/super1.c
+++ b/super1.c
@@ -1978,26 +1978,10 @@ static int write_init_super1(struct supertype *st)
 	unsigned long long sb_offset;
 	unsigned long long data_offset;
 	long bm_offset;
-	int raid0_need_layout = 0;
 
-	for (di = st->info; di; di = di->next) {
+	for (di = st->info; di; di = di->next)
 		if (di->disk.state & (1 << MD_DISK_JOURNAL))
 			sb->feature_map |= __cpu_to_le32(MD_FEATURE_JOURNAL);
-		if (sb->level == 0 && sb->layout != 0) {
-			struct devinfo *di2 = st->info;
-			unsigned long long s1, s2;
-			s1 = di->dev_size;
-			if (di->data_offset != INVALID_SECTORS)
-				s1 -= di->data_offset;
-			s1 /= __le32_to_cpu(sb->chunksize);
-			s2 = di2->dev_size;
-			if (di2->data_offset != INVALID_SECTORS)
-				s2 -= di2->data_offset;
-			s2 /= __le32_to_cpu(sb->chunksize);
-			if (s1 != s2)
-				raid0_need_layout = 1;
-		}
-	}
 
 	for (di = st->info; di; di = di->next) {
 		if (di->disk.state & (1 << MD_DISK_FAULTY))
@@ -2139,8 +2123,7 @@ static int write_init_super1(struct supertype *st)
 			sb->bblog_offset = 0;
 		}
 
-		/* RAID0 needs a layout if devices aren't all the same size */
-		if (raid0_need_layout)
+		if (sb->level == 0 && sb->layout && get_linux_version() >= 5004000)
 			sb->feature_map |= __cpu_to_le32(MD_FEATURE_RAID0_LAYOUT);
 
 		sb->sb_csum = calc_sb_1_csum(sb);
-- 
2.32.0 (Apple Git-132)

Re: [PATCH V2 1/1] mdadm/super1: Add MD_FEATURE_RAID0_LAYOUT for all raid0 after kernel v5.4

[Mariusz Tkaczyk]

On Tue, 17 Oct 2023 11:51:42 +0800
Xiao Ni <xni@redhat.com> wrote:

> After and include kernel v5.4, it adds one feature bit
> MD_FEATURE_RAID0_LAYOUT. It must need to specify a layout for raid0 with more
> than one zone. But for raid0 with one zone, in fact it also has a defalut
> layout.
> 
> Now for raid0 with one zone, *unknown* layout can be seen when running mdadm
> -D command. It's the reason that mdadm doesn't set MD_FEATURE_RAID0_LAYOUT for
> raid0 with one zone. Then in kernel space, super_1_validate sets mddev->layout
> to -1 because of no MD_FEATURE_RAID0_LAYOUT. In fact, in raid0 io path, it
> uses the default layout. So in fact after/include kernel v5.4, all raid0
> device have layout.
> 
> Fixes: 329dfc28debb ('Create: add support for RAID0 layouts.')
> Signed-off-by: Xiao Ni <xni@redhat.com>
> ---
>  super1.c | 21 ++-------------------
>  1 file changed, 2 insertions(+), 19 deletions(-)
> 
> diff --git a/super1.c b/super1.c
> index 856b02082662..653a2ea6c0e4 100644
> --- a/super1.c
> +++ b/super1.c
> @@ -1978,26 +1978,10 @@ static int write_init_super1(struct supertype *st)
>  	unsigned long long sb_offset;
>  	unsigned long long data_offset;
>  	long bm_offset;
> -	int raid0_need_layout = 0;
>  
> -	for (di = st->info; di; di = di->next) {
> +	for (di = st->info; di; di = di->next)
>  		if (di->disk.state & (1 << MD_DISK_JOURNAL))
>  			sb->feature_map |= __cpu_to_le32(MD_FEATURE_JOURNAL);
> -		if (sb->level == 0 && sb->layout != 0) {
> -			struct devinfo *di2 = st->info;
> -			unsigned long long s1, s2;
> -			s1 = di->dev_size;
> -			if (di->data_offset != INVALID_SECTORS)
> -				s1 -= di->data_offset;
> -			s1 /= __le32_to_cpu(sb->chunksize);
> -			s2 = di2->dev_size;
> -			if (di2->data_offset != INVALID_SECTORS)
> -				s2 -= di2->data_offset;
> -			s2 /= __le32_to_cpu(sb->chunksize);
> -			if (s1 != s2)
> -				raid0_need_layout = 1;
> -		}
> -	}

We need to keep this code. Neil made MD_FEATURE_RAID0_LAYOUT always added for
device with various sizes. You are extending it not replacing.

I understand that now it sets MD_FEATURE_RAID0_LAYOUT if it detects
member devices with various sizes. Kernel version is irrelevant so I suspect
that if someone creates zoned raid0 array, it fails to start array if
MD_FEATURE_RAID0_LAYOUT is not supported by the MD driver. User must
acknowledge that by layout=dangerous (it means no layout I think).

We don't want remove this. It prevents users from data corruption.

Your change is to start always setting MD_FEATURE_RAID0_LAYOUT if it seems to be
safe i.e. kernel is >=5.4 but it does not invalidate the raid0_need_layout
routine from the reason raised above.

Please correct me if I missed something or if I'm wrong. I did not tested it.
I trust that you made necessary testing and can provide real-life input here.

Thanks,
Mariusz

Re: [PATCH V2 1/1] mdadm/super1: Add MD_FEATURE_RAID0_LAYOUT for all raid0 after kernel v5.4

[Xiao Ni]

On Tue, Oct 17, 2023 at 3:36 PM Mariusz Tkaczyk
<mariusz.tkaczyk@linux.intel.com> wrote:
>
> On Tue, 17 Oct 2023 11:51:42 +0800
> Xiao Ni <xni@redhat.com> wrote:
>
> > After and include kernel v5.4, it adds one feature bit
> > MD_FEATURE_RAID0_LAYOUT. It must need to specify a layout for raid0 with more
> > than one zone. But for raid0 with one zone, in fact it also has a defalut
> > layout.
> >
> > Now for raid0 with one zone, *unknown* layout can be seen when running mdadm
> > -D command. It's the reason that mdadm doesn't set MD_FEATURE_RAID0_LAYOUT for
> > raid0 with one zone. Then in kernel space, super_1_validate sets mddev->layout
> > to -1 because of no MD_FEATURE_RAID0_LAYOUT. In fact, in raid0 io path, it
> > uses the default layout. So in fact after/include kernel v5.4, all raid0
> > device have layout.
> >
> > Fixes: 329dfc28debb ('Create: add support for RAID0 layouts.')
> > Signed-off-by: Xiao Ni <xni@redhat.com>
> > ---
> >  super1.c | 21 ++-------------------
> >  1 file changed, 2 insertions(+), 19 deletions(-)
> >
> > diff --git a/super1.c b/super1.c
> > index 856b02082662..653a2ea6c0e4 100644
> > --- a/super1.c
> > +++ b/super1.c
> > @@ -1978,26 +1978,10 @@ static int write_init_super1(struct supertype *st)
> >       unsigned long long sb_offset;
> >       unsigned long long data_offset;
> >       long bm_offset;
> > -     int raid0_need_layout = 0;
> >
> > -     for (di = st->info; di; di = di->next) {
> > +     for (di = st->info; di; di = di->next)
> >               if (di->disk.state & (1 << MD_DISK_JOURNAL))
> >                       sb->feature_map |= __cpu_to_le32(MD_FEATURE_JOURNAL);
> > -             if (sb->level == 0 && sb->layout != 0) {
> > -                     struct devinfo *di2 = st->info;
> > -                     unsigned long long s1, s2;
> > -                     s1 = di->dev_size;
> > -                     if (di->data_offset != INVALID_SECTORS)
> > -                             s1 -= di->data_offset;
> > -                     s1 /= __le32_to_cpu(sb->chunksize);
> > -                     s2 = di2->dev_size;
> > -                     if (di2->data_offset != INVALID_SECTORS)
> > -                             s2 -= di2->data_offset;
> > -                     s2 /= __le32_to_cpu(sb->chunksize);
> > -                     if (s1 != s2)
> > -                             raid0_need_layout = 1;
> > -             }
> > -     }
>
> We need to keep this code. Neil made MD_FEATURE_RAID0_LAYOUT always added for
> device with various sizes. You are extending it not replacing.
>
> I understand that now it sets MD_FEATURE_RAID0_LAYOUT if it detects
> member devices with various sizes. Kernel version is irrelevant so I suspect
> that if someone creates zoned raid0 array, it fails to start array if
> MD_FEATURE_RAID0_LAYOUT is not supported by the MD driver. User must
> acknowledge that by layout=dangerous (it means no layout I think).
>
> We don't want remove this. It prevents users from data corruption.
>
> Your change is to start always setting MD_FEATURE_RAID0_LAYOUT if it seems to be
> safe i.e. kernel is >=5.4 but it does not invalidate the raid0_need_layout
> routine from the reason raised above.
>
> Please correct me if I missed something or if I'm wrong. I did not tested it.
> I trust that you made necessary testing and can provide real-life input here.
>
> Thanks,
> Mariusz
>

Hi Mariusz

Thanks. I'll add the check <5.4 situation

Best Regards
Xiao