On Fri, Oct 30, 2020 at 01:46:18PM +0530, Chandan Babu R wrote:
> On Friday 30 October 2020 3:51:30 AM IST Darrick J. Wong wrote:
> > On Thu, Oct 29, 2020 at 03:43:46PM +0530, Chandan Babu R wrote:
> > > This commit moves over the code which computes stripe alignment and
> > > extent size hint alignment into a separate function. Apart from
> > > xfs_bmap_btalloc(), the new function will be used by another function
> > > introduced in a future commit.
> > >
> > > Reviewed-by: Allison Henderson <allison.henderson@xxxxxxxxxx>
> > > Signed-off-by: Chandan Babu R <chandanrlinux@xxxxxxxxx>
> > > ---
> > > fs/xfs/libxfs/xfs_bmap.c | 88 +++++++++++++++++++++++-----------------
> > > 1 file changed, 51 insertions(+), 37 deletions(-)
> > >
> > > diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
> > > index 64c4d0e384a5..935f2d506748 100644
> > > --- a/fs/xfs/libxfs/xfs_bmap.c
> > > +++ b/fs/xfs/libxfs/xfs_bmap.c
> > > @@ -3463,13 +3463,58 @@ xfs_bmap_btalloc_accounting(
> > > args->len);
> > > }
> > >
> > > +static void
> >
> > Why not return stripe_align instead of passing pointers?
>
> xfs_bmap_exact_minlen_extent_alloc() introduced in the last patch would invoke
> this function passing NULL value as the third argument i.e. it does not need
> "stripe alignment" to be computed. Hence xfs_bmap_exact_minlen_extent_alloc()
> would ignore the return value of xfs_bmap_compute_alignments(). This was the
> reason for deciding on passing a pointer to the stripe_align variable as an
> argument.
I would have thought that an out parameter that isn't used by all
callers would be the perfect use of a return, especially since passing
by pointer means that the compiler has to spill stripe_align to a memory
location both here and in the callers and cannot keep the value in
registers.
--D
> >
> > > +xfs_bmap_compute_alignments(
> > > + struct xfs_bmalloca *ap,
> > > + struct xfs_alloc_arg *args,
> > > + int *stripe_align)
> > > +{
> > > + struct xfs_mount *mp = args->mp;
> > > + xfs_extlen_t align = 0; /* minimum allocation alignment */
> > > + int error;
> > > +
> > > + /* stripe alignment for allocation is determined by mount parameters */
> > > + *stripe_align = 0;
> > > + if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC))
> > > + *stripe_align = mp->m_swidth;
> > > + else if (mp->m_dalign)
> > > + *stripe_align = mp->m_dalign;
> > > +
> > > + if (ap->flags & XFS_BMAPI_COWFORK)
> > > + align = xfs_get_cowextsz_hint(ap->ip);
> > > + else if (ap->datatype & XFS_ALLOC_USERDATA)
> > > + align = xfs_get_extsz_hint(ap->ip);
> > > + if (align) {
> > > + error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
> > > + align, 0, ap->eof, 0, ap->conv,
> > > + &ap->offset, &ap->length);
> > > + ASSERT(!error);
> > > + ASSERT(ap->length);
> > > + }
> > > +
> > > + /* apply extent size hints if obtained earlier */
> > > + if (align) {
> > > + args->prod = align;
> > > + div_u64_rem(ap->offset, args->prod, &args->mod);
> > > + if (args->mod)
> > > + args->mod = args->prod - args->mod;
> > > + } else if (mp->m_sb.sb_blocksize >= PAGE_SIZE) {
> > > + args->prod = 1;
> > > + args->mod = 0;
> > > + } else {
> > > + args->prod = PAGE_SIZE >> mp->m_sb.sb_blocklog;
> > > + div_u64_rem(ap->offset, args->prod, &args->mod);
> > > + if (args->mod)
> > > + args->mod = args->prod - args->mod;
> > > + }
> > > +}
> > > +
> > > STATIC int
> > > xfs_bmap_btalloc(
> > > struct xfs_bmalloca *ap) /* bmap alloc argument struct */
> > > {
> > > xfs_mount_t *mp; /* mount point structure */
> > > xfs_alloctype_t atype = 0; /* type for allocation routines */
> > > - xfs_extlen_t align = 0; /* minimum allocation alignment */
> > > xfs_agnumber_t fb_agno; /* ag number of ap->firstblock */
> > > xfs_agnumber_t ag;
> > > xfs_alloc_arg_t args;
> > > @@ -3489,25 +3534,11 @@ xfs_bmap_btalloc(
> > >
> > > mp = ap->ip->i_mount;
> > >
> > > - /* stripe alignment for allocation is determined by mount parameters */
> > > - stripe_align = 0;
> > > - if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC))
> > > - stripe_align = mp->m_swidth;
> > > - else if (mp->m_dalign)
> > > - stripe_align = mp->m_dalign;
> > > -
> > > - if (ap->flags & XFS_BMAPI_COWFORK)
> > > - align = xfs_get_cowextsz_hint(ap->ip);
> > > - else if (ap->datatype & XFS_ALLOC_USERDATA)
> > > - align = xfs_get_extsz_hint(ap->ip);
> > > - if (align) {
> > > - error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev,
> > > - align, 0, ap->eof, 0, ap->conv,
> > > - &ap->offset, &ap->length);
> > > - ASSERT(!error);
> > > - ASSERT(ap->length);
> > > - }
> > > + memset(&args, 0, sizeof(args));
> > > + args.tp = ap->tp;
> > > + args.mp = mp;
> >
> > FWIW you might as well clean up the variable declarations while you're
> > moving this stuff around:
> >
> > STATIC int
> > xfs_bmap_btalloc(
> > struct xfs_bmalloca *ap)
> > {
> > struct xfs_mount *mp = ap->ip->i_mount;
> > struct xfs_alloc_arg args = { .tp = ap->tp, .mp = mp };
> >
> > And then you can get rid of the memset call.
>
> Sure, I will make the changes that have been suggested.
>
> >
> > AFAICT there aren't any data dependencies between the parts where we
> > initialize args.fsbno and where we set args.prod and args.mod, so I
> > guess this is a reasonable hoist.
> >
> > Other than those cleanups, this looks ok to me.
> >
> > --D
> >
> > >
> > > + xfs_bmap_compute_alignments(ap, &args, &stripe_align);
> > >
> > > nullfb = ap->tp->t_firstblock == NULLFSBLOCK;
> > > fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp,
> > > @@ -3538,9 +3569,6 @@ xfs_bmap_btalloc(
> > > * Normal allocation, done through xfs_alloc_vextent.
> > > */
> > > tryagain = isaligned = 0;
> > > - memset(&args, 0, sizeof(args));
> > > - args.tp = ap->tp;
> > > - args.mp = mp;
> > > args.fsbno = ap->blkno;
> > > args.oinfo = XFS_RMAP_OINFO_SKIP_UPDATE;
> > >
> > > @@ -3571,21 +3599,7 @@ xfs_bmap_btalloc(
> > > args.total = ap->total;
> > > args.minlen = ap->minlen;
> > > }
> > > - /* apply extent size hints if obtained earlier */
> > > - if (align) {
> > > - args.prod = align;
> > > - div_u64_rem(ap->offset, args.prod, &args.mod);
> > > - if (args.mod)
> > > - args.mod = args.prod - args.mod;
> > > - } else if (mp->m_sb.sb_blocksize >= PAGE_SIZE) {
> > > - args.prod = 1;
> > > - args.mod = 0;
> > > - } else {
> > > - args.prod = PAGE_SIZE >> mp->m_sb.sb_blocklog;
> > > - div_u64_rem(ap->offset, args.prod, &args.mod);
> > > - if (args.mod)
> > > - args.mod = args.prod - args.mod;
> > > - }
> > > +
> > > /*
> > > * If we are not low on available data blocks, and the underlying
> > > * logical volume manager is a stripe, and the file offset is zero then
> >
>
>
> --
> chandan
>
>
>
