Re: erofs: use wrapper i_blocksize() in erofs_file_read_iter()

[Gao Xiang]

On 2023/3/9 15:42, Yangtao Li wrote:
> > FYI it seems that GENMASK macro is widely used in driver and arch code base, while it's rarely used in fs, except for f2fs.
>
> I think the following usage can be changed to bitmap api, just like in f2fs?
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=c4ca1f7164734a1baf40d4ff1552172a07d4fc4d
>
> fs/erofs/fscache.c:135:         unsigned long flags = 1 << NETFS_SREQ_ONDEMAND;
> fs/erofs/internal.h:250:#define SECTORS_PER_BLOCK       (1 << SECTORS_PER_BLOCK)
> fs/erofs/internal.h:252:#define EROFS_BLKSIZ            (1 << LOG_BLOCK_SIZE)
> fs/erofs/internal.h:354:        return (value >> bit) & ((1 << bits) - 1);
> fs/erofs/zmap.c:66:             ((1 << Z_EROFS_VLE_DI_CLUSTER_TYPE_BITS) - 1);
> fs/erofs/zmap.c:69:             m->clusterofs = 1 << vi->z_logical_clusterbits;
> fs/erofs/zmap.c:114:    const unsigned int lomask = (1 << lclusterbits) - 1;
> fs/erofs/zmap.c:141:    const unsigned int lomask = (1 << lclusterbits) - 1;
> fs/erofs/zmap.c:147:    if (1 << amortizedshift == 4)
> fs/erofs/zmap.c:149:    else if (1 << amortizedshift == 2 && lclusterbits == 12)
> fs/erofs/zmap.c:169:            m->clusterofs = 1 << lclusterbits;
> fs/erofs/zmap.c:291:    pos += lcn * (1 << amortizedshift);
> fs/erofs/zmap.c:409:            m->compressedblks = 1 << (lclusterbits - LOG_BLOCK_SIZE);
> fs/erofs/zmap.c:457:                              m->clusterofs != 1 << lclusterbits);
> fs/erofs/zmap.c:497:    endoff = ofs & ((1 << lclusterbits) - 1);
> fs/erofs/erofs_fs.h:120:        ((1 << (EROFS_I_DATALAYOUT_BIT + EROFS_I_DATALAYOUT_BITS)) - 1)
> fs/erofs/erofs_fs.h:279:#define Z_EROFS_ALL_COMPR_ALGS          ((1 << Z_EROFS_COMPRESSION_MAX) - 1)
> fs/erofs/erofs_fs.h:377:#define Z_EROFS_VLE_DI_PARTIAL_REF              (1 << 15)
> fs/erofs/erofs_fs.h:384:#define Z_EROFS_VLE_DI_D0_CBLKCNT               (1 << 11)
> fs/erofs/erofs_fs.h:427:                .h_clusterbits = 1 << Z_EROFS_FRAGMENT_INODE_BIT
> fs/erofs/data.c:379:                    blksize_mask = (1 << inode->i_blkbits) - 1;
> fs/erofs/zdata.c:133:#define Z_EROFS_PAGE_EIO                   (1 << 30)

Is there some benefit to use these? BIT(1) vs 1 << 1? also almost all
filesystems rarely use such APIs honestly.

Thanks,
Gao Xiang

> Thx,
> Yangtao