From: "Krzysztof Błaszkowski" <kb@sysmikro.com.pl>
To: Carlos Maiolino <cmaiolino@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Subject: Re: freevxfs
Date: Thu, 26 May 2016 16:44:06 +0200 [thread overview]
Message-ID: <1464273846.17980.14.camel@linux-q3cb.site> (raw)
In-Reply-To: <20160526115004.GA734@redhat.com>
Hi Carlos,
On Thu, 2016-05-26 at 13:50 +0200, Carlos Maiolino wrote:
> On Wed, May 25, 2016 at 11:27:27PM +0200, Krzysztof Błaszkowski wrote:
> > Hi,
> >
> > Please find included patchset which addresses all major features we have
> > discussed (unless I forgot about something).
> >
>
> Hi Krzysztof,
>
> please send the patches one-by one, inlined in the e-mail message, not attached
> to the e-mail, mainly all of them attached to a single message.
>
> You can use git send-email to send the patches.
Thanks for the tip. Will send the email shortly soon.
I will not use git send-mail because can't recall password for my mail account now.
I hope only that these patches can be applied still after receiving them.
(patches won't be malformed).
Thanks
>
> See: Documentation/SubmittingPatches
>
> Cheers
>
> > I verified correctness of operation with my regression tests and HP-UX's
> > vxfs image. No difference spotted. However I can't do this with SCO
> > image because I do not have such. I reckon that it is highly possible
> > that SCO image will work as usual because the only difference is data
> > endianess and offset of super block.
> >
> > I included for reference these scripts just in case one is curious.
> > (chksum8-sorted.log comes from hp-ux's cksum)
> >
> > Regards,
> > --
> > Krzysztof Blaszkowski
>
> > From 3d3b4e1ed5df014ae191e0566ff86a17d7d9ac05 Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 21:50:11 +0200
> > Subject: [PATCH 1/7] kconfig note
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/Kconfig | 10 +++++++++-
> > 1 files changed, 9 insertions(+), 1 deletions(-)
> >
> > diff --git a/fs/freevxfs/Kconfig b/fs/freevxfs/Kconfig
> > index 8dc1cd5..a4c9075 100644
> > --- a/fs/freevxfs/Kconfig
> > +++ b/fs/freevxfs/Kconfig
> > @@ -5,12 +5,20 @@ config VXFS_FS
> > FreeVxFS is a file system driver that support the VERITAS VxFS(TM)
> > file system format. VERITAS VxFS(TM) is the standard file system
> > of SCO UnixWare (and possibly others) and optionally available
> > - for Sunsoft Solaris, HP-UX and many other operating systems.
> > + for Sunsoft Solaris, HP-UX and many other operating systems. However
> > + these particular OS implementations of vxfs may differ in on-disk
> > + data endianess and/or superblock offset. The vxfs module has been
> > + tested with SCO UnixWare and HP-UX B.10.20 (pa-risc 1.1 arch.)
> > Currently only readonly access is supported.
> >
> > NOTE: the file system type as used by mount(1), mount(2) and
> > fstab(5) is 'vxfs' as it describes the file system format, not
> > the actual driver.
> >
> > + There is a userspace utility for HP-UX logical volumes which makes
> > + creating HP-UX logical volumes easy from HP-UX disk block device file
> > + or regular file with image of the disk. See:
> > + https://sourceforge.net/projects/linux-vxfs/
> > +
> > To compile this as a module, choose M here: the module will be
> > called freevxfs. If unsure, say N.
> > --
> > 1.7.3.4
> >
>
> > From c617f6bceedc2f68c62e7432f12b59124bab34f7 Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:12:05 +0200
> > Subject: [PATCH 2/7] cpu endian vs file system endian
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs.h | 25 +++++++++++-
> > fs/freevxfs/vxfs_bmap.c | 24 ++++++-----
> > fs/freevxfs/vxfs_fshead.c | 25 +++++++++++-
> > fs/freevxfs/vxfs_inode.c | 61 +++++++++++++++++++++++++++-
> > fs/freevxfs/vxfs_inode.h | 8 ++--
> > fs/freevxfs/vxfs_lookup.c | 15 ++++++-
> > fs/freevxfs/vxfs_olt.c | 15 ++++---
> > fs/freevxfs/vxfs_super.c | 97 ++++++++++++++++++++++++++++++++-------------
> > 8 files changed, 215 insertions(+), 55 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs.h b/fs/freevxfs/vxfs.h
> > index c8a9265..5dc8949 100644
> > --- a/fs/freevxfs/vxfs.h
> > +++ b/fs/freevxfs/vxfs.h
> > @@ -152,7 +152,7 @@ struct vxfs_sb {
> > /*
> > * Actually much more...
> > */
> > -};
> > +} __packed;
> >
> >
> > /*
> > @@ -168,9 +168,32 @@ struct vxfs_sb_info {
> > ino_t vsi_fshino; /* fileset header inode */
> > daddr_t vsi_oltext; /* OLT extent */
> > daddr_t vsi_oltsize; /* OLT size */
> > + int byte_order;
> > + int silent;
> > +};
> > +
> > +enum {
> > + BO_LE = 1,
> > + BO_BE
> > };
> >
> >
> > +static inline u32 fs32_to_cpu(int bo, u32 a)
> > +{
> > + return (bo == BO_BE) ? be32_to_cpu(a) : le32_to_cpu(a);
> > +}
> > +
> > +static inline u16 fs16_to_cpu(int bo, u16 a)
> > +{
> > + return (bo == BO_BE) ? be16_to_cpu(a) : le16_to_cpu(a);
> > +}
> > +
> > +static inline u64 fs64_to_cpu(int bo, u64 a)
> > +{
> > + return (bo == BO_BE) ? be64_to_cpu(a) : le64_to_cpu(a);
> > +}
> > +
> > +
> > /*
> > * File modes. File types above 0xf000 are vxfs internal only, they should
> > * not be passed back to higher levels of the system. vxfs file types must
> > diff --git a/fs/freevxfs/vxfs_bmap.c b/fs/freevxfs/vxfs_bmap.c
> > index f86fd3c..95afd98 100644
> > --- a/fs/freevxfs/vxfs_bmap.c
> > +++ b/fs/freevxfs/vxfs_bmap.c
> > @@ -92,7 +92,8 @@ vxfs_bmap_ext4(struct inode *ip, long bn)
> > goto fail_buf;
> >
> > indir = (u32 *)buf->b_data;
> > - bno = indir[(bn/indsize) % (indsize*bn)] + (bn%indsize);
> > + bno = fs32_to_cpu(VXFS_SBI(sb)->byte_order,
> > + indir[(bn/indsize) % (indsize*bn)]) + (bn % indsize);
> >
> > brelse(buf);
> > return bno;
> > @@ -130,6 +131,7 @@ vxfs_bmap_indir(struct inode *ip, long indir, int size, long block)
> > struct buffer_head *bp = NULL;
> > daddr_t pblock = 0;
> > int i;
> > + int bo = VXFS_SBI(ip->i_sb)->byte_order;
> >
> > for (i = 0; i < size * VXFS_TYPED_PER_BLOCK(ip->i_sb); i++) {
> > struct vxfs_typed *typ;
> > @@ -142,24 +144,24 @@ vxfs_bmap_indir(struct inode *ip, long indir, int size, long block)
> >
> > typ = ((struct vxfs_typed *)bp->b_data) +
> > (i % VXFS_TYPED_PER_BLOCK(ip->i_sb));
> > - off = (typ->vt_hdr & VXFS_TYPED_OFFSETMASK);
> > + off = fs64_to_cpu(bo, typ->vt_hdr) & VXFS_TYPED_OFFSETMASK;
> >
> > if (block < off) {
> > brelse(bp);
> > continue;
> > }
> >
> > - switch ((u_int32_t)(typ->vt_hdr >> VXFS_TYPED_TYPESHIFT)) {
> > + switch ((u_int32_t)(fs64_to_cpu(bo, typ->vt_hdr) >> VXFS_TYPED_TYPESHIFT)) {
> > case VXFS_TYPED_INDIRECT:
> > - pblock = vxfs_bmap_indir(ip, typ->vt_block,
> > - typ->vt_size, block - off);
> > + pblock = vxfs_bmap_indir(ip, fs32_to_cpu(bo, typ->vt_block),
> > + fs32_to_cpu(bo, typ->vt_size), block - off);
> > if (pblock == -2)
> > break;
> > goto out;
> > case VXFS_TYPED_DATA:
> > - if ((block - off) >= typ->vt_size)
> > + if ((block - off) >= fs32_to_cpu(bo, typ->vt_size))
> > break;
> > - pblock = (typ->vt_block + block - off);
> > + pblock = fs32_to_cpu(bo, typ->vt_block) + block - off;
> > goto out;
> > case VXFS_TYPED_INDIRECT_DEV4:
> > case VXFS_TYPED_DATA_DEV4: {
> > @@ -168,12 +170,14 @@ vxfs_bmap_indir(struct inode *ip, long indir, int size, long block)
> >
> > printk(KERN_INFO "\n\nTYPED_DEV4 detected!\n");
> > printk(KERN_INFO "block: %Lu\tsize: %Ld\tdev: %d\n",
> > - (unsigned long long) typ4->vd4_block,
> > - (unsigned long long) typ4->vd4_size,
> > - typ4->vd4_dev);
> > + (unsigned long long) fs64_to_cpu(bo, typ4->vd4_block),
> > + (unsigned long long) fs64_to_cpu(bo, typ4->vd4_size),
> > + fs32_to_cpu(bo, typ4->vd4_dev));
> > goto fail;
> > }
> > default:
> > + printk(KERN_ERR "%s:%d vt_hdr %llu\n", __func__, __LINE__,
> > + fs64_to_cpu(bo, typ->vt_hdr));
> > BUG();
> > }
> > brelse(bp);
> > diff --git a/fs/freevxfs/vxfs_fshead.c b/fs/freevxfs/vxfs_fshead.c
> > index c9a6a94..6cbdde7 100644
> > --- a/fs/freevxfs/vxfs_fshead.c
> > +++ b/fs/freevxfs/vxfs_fshead.c
> > @@ -60,6 +60,29 @@ vxfs_dumpfsh(struct vxfs_fsh *fhp)
> > }
> > #endif
> >
> > +#define VXFS_FS32(field1, field2) fhp->field1 = fs32_to_cpu(bo, dbh->field2)
> > +static inline void dbh2fhp(struct vxfs_fsh *fhp, void *_dbh, int bo)
> > +{
> > + struct vxfs_fsh *dbh = (struct vxfs_fsh *)_dbh;
> > +
> > + VXFS_FS32(fsh_version, fsh_version);
> > + VXFS_FS32(fsh_fsindex, fsh_fsindex);
> > + VXFS_FS32(fsh_time, fsh_time);
> > + VXFS_FS32(fsh_utime, fsh_utime);
> > + VXFS_FS32(fsh_extop, fsh_extop);
> > + VXFS_FS32(fsh_ninodes, fsh_ninodes);
> > + VXFS_FS32(fsh_nau, fsh_nau);
> > + VXFS_FS32(fsh_old_ilesize, fsh_old_ilesize);
> > + VXFS_FS32(fsh_dflags, fsh_dflags);
> > + VXFS_FS32(fsh_quota, fsh_quota);
> > + VXFS_FS32(fsh_maxinode, fsh_maxinode);
> > + VXFS_FS32(fsh_iauino, fsh_iauino);
> > + VXFS_FS32(fsh_ilistino[0], fsh_ilistino[0]);
> > + VXFS_FS32(fsh_ilistino[1], fsh_ilistino[1]);
> > + VXFS_FS32(fsh_lctino, fsh_lctino);
> > +}
> > +
> > +
> > /**
> > * vxfs_getfsh - read fileset header into memory
> > * @ip: the (fake) fileset header inode
> > @@ -83,7 +106,7 @@ vxfs_getfsh(struct inode *ip, int which)
> >
> > if (!(fhp = kmalloc(sizeof(*fhp), GFP_KERNEL)))
> > goto out;
> > - memcpy(fhp, bp->b_data, sizeof(*fhp));
> > + dbh2fhp(fhp, bp->b_data, VXFS_SBI(ip->i_sb)->byte_order);
> >
> > put_bh(bp);
> > return (fhp);
> > diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
> > index 363e3ae..53b8757 100644
> > --- a/fs/freevxfs/vxfs_inode.c
> > +++ b/fs/freevxfs/vxfs_inode.c
> > @@ -67,6 +67,63 @@ vxfs_dumpi(struct vxfs_inode_info *vip, ino_t ino)
> > }
> > #endif
> >
> > +#define VXFS_FS32(field1, field2) vip->field1 = fs32_to_cpu(bo, dip->field2)
> > +#define VXFS_FS64(field1, field2) vip->field1 = fs64_to_cpu(bo, dip->field2)
> > +#define VXFS_FS16(field1, field2) vip->field1 = fs16_to_cpu(bo, dip->field2)
> > +
> > +static inline void dip2vip_cpy(struct vxfs_inode_info *vip, struct vxfs_dinode *dip, int bo)
> > +{
> > + int j;
> > +
> > + VXFS_FS32(vdi_mode, vdi_mode);
> > + VXFS_FS32(vdi_nlink, vdi_nlink);
> > + VXFS_FS32(vdi_uid, vdi_uid);
> > + VXFS_FS32(vdi_gid, vdi_gid);
> > + VXFS_FS64(vdi_size, vdi_size);
> > + VXFS_FS32(vdi_atime, vdi_atime);
> > + VXFS_FS32(vdi_autime, vdi_autime);
> > + VXFS_FS32(vdi_mtime, vdi_mtime);
> > + VXFS_FS32(vdi_mutime, vdi_mutime);
> > + VXFS_FS32(vdi_ctime, vdi_ctime);
> > + VXFS_FS32(vdi_cutime, vdi_cutime);
> > + vip->vdi_aflags = dip->vdi_aflags;
> > + vip->vdi_orgtype = dip->vdi_orgtype;
> > + VXFS_FS16(vdi_eopflags, vdi_eopflags);
> > + VXFS_FS32(vdi_eopdata, vdi_eopdata);
> > +
> > + VXFS_FS32(vdi_ftarea.i_regular.reserved, vdi_ftarea.i_regular.reserved);
> > + VXFS_FS32(vdi_ftarea.i_regular.fixextsize, vdi_ftarea.i_regular.fixextsize);
> > + VXFS_FS32(vdi_blocks, vdi_blocks);
> > + VXFS_FS32(vdi_gen, vdi_gen);
> > + VXFS_FS64(vdi_version, vdi_version);
> > +
> > + switch (dip->vdi_orgtype) {
> > + case VXFS_ORG_EXT4:
> > + VXFS_FS32(vdi_org.ext4.ve4_spare, vdi_org.ext4.ve4_spare);
> > + VXFS_FS32(vdi_org.ext4.ve4_indsize, vdi_org.ext4.ve4_indsize);
> > + for (j = 0; j < VXFS_NIADDR; j++) {
> > + VXFS_FS32(vdi_org.ext4.ve4_indir[j], vdi_org.ext4.ve4_indir[j]);
> > + }
> > + for (j = 0; j < VXFS_NDADDR; j++) {
> > + VXFS_FS32(vdi_org.ext4.ve4_direct[j].extent, vdi_org.ext4.ve4_direct[j].extent);
> > + VXFS_FS32(vdi_org.ext4.ve4_direct[j].size, vdi_org.ext4.ve4_direct[j].size);
> > + }
> > + break;
> > + case VXFS_ORG_IMMED:
> > + memcpy(&vip->vdi_org.immed, &dip->vdi_org.immed, sizeof(vip->vdi_org.immed));
> > + break;
> > + case VXFS_ORG_TYPED:
> > + for (j = 0; j < VXFS_NTYPED; j++) {
> > + VXFS_FS64(vdi_org.typed[j].vt_hdr, vdi_org.typed[j].vt_hdr);
> > + VXFS_FS32(vdi_org.typed[j].vt_block, vdi_org.typed[j].vt_block);
> > + VXFS_FS32(vdi_org.typed[j].vt_size, vdi_org.typed[j].vt_size);
> > + }
> > + break;
> > + };
> > +
> > + VXFS_FS32(vdi_iattrino, vdi_iattrino);
> > +}
> > +
> >
> > /**
> > * vxfs_blkiget - find inode based on extent #
> > @@ -101,7 +158,7 @@ vxfs_blkiget(struct super_block *sbp, u_long extent, ino_t ino)
> > if (!(vip = kmem_cache_alloc(vxfs_inode_cachep, GFP_KERNEL)))
> > goto fail;
> > dip = (struct vxfs_dinode *)(bp->b_data + offset);
> > - memcpy(vip, dip, sizeof(*vip));
> > + dip2vip_cpy(vip, dip, VXFS_SBI(sbp)->byte_order);
> > #ifdef DIAGNOSTIC
> > vxfs_dumpi(vip, ino);
> > #endif
> > @@ -143,7 +200,7 @@ __vxfs_iget(ino_t ino, struct inode *ilistp)
> > if (!(vip = kmem_cache_alloc(vxfs_inode_cachep, GFP_KERNEL)))
> > goto fail;
> > dip = (struct vxfs_dinode *)(kaddr + offset);
> > - memcpy(vip, dip, sizeof(*vip));
> > + dip2vip_cpy(vip, dip, VXFS_SBI(ilistp->i_sb)->byte_order);
> > #ifdef DIAGNOSTIC
> > vxfs_dumpi(vip, ino);
> > #endif
> > diff --git a/fs/freevxfs/vxfs_inode.h b/fs/freevxfs/vxfs_inode.h
> > index 240aeb1..9a2c376 100644
> > --- a/fs/freevxfs/vxfs_inode.h
> > +++ b/fs/freevxfs/vxfs_inode.h
> > @@ -77,13 +77,13 @@ struct vxfs_ext4 {
> > vx_daddr_t extent; /* Extent number */
> > int32_t size; /* Size of extent */
> > } ve4_direct[VXFS_NDADDR];
> > -};
> > +} __packed;
> >
> > struct vxfs_typed {
> > u_int64_t vt_hdr; /* Header, 0xTTOOOOOOOOOOOOOO; T=type,O=offs */
> > vx_daddr_t vt_block; /* Extent block */
> > int32_t vt_size; /* Size in blocks */
> > -};
> > +} __packed;
> >
> > struct vxfs_typed_dev4 {
> > u_int64_t vd4_hdr; /* Header, 0xTTOOOOOOOOOOOOOO; T=type,O=offs */
> > @@ -91,7 +91,7 @@ struct vxfs_typed_dev4 {
> > u_int64_t vd4_size; /* Size in blocks */
> > int32_t vd4_dev; /* Device ID */
> > u_int32_t __pad1;
> > -};
> > +} __packed;
> >
> > /*
> > * The inode as contained on the physical device.
> > @@ -134,7 +134,7 @@ struct vxfs_dinode {
> > struct vxfs_typed typed[VXFS_NTYPED];
> > } vdi_org;
> > u_int32_t vdi_iattrino;
> > -};
> > +} __packed;
> >
> > #define vdi_rdev vdi_ftarea.rdev
> > #define vdi_dotdot vdi_ftarea.dotdot
> > diff --git a/fs/freevxfs/vxfs_lookup.c b/fs/freevxfs/vxfs_lookup.c
> > index 99c7f0a..cea158a 100644
> > --- a/fs/freevxfs/vxfs_lookup.c
> > +++ b/fs/freevxfs/vxfs_lookup.c
> > @@ -96,6 +96,15 @@ vxfs_next_entry(struct vxfs_direct *de)
> > return ((struct vxfs_direct *)((char*)de + de->d_reclen));
> > }
> >
> > +/*
> > + * VXFS_dirblk_ovh is the overhead of a specific dirblock.
> > + */
> > +static inline u_long VXFS_dirblk_ovh(struct vxfs_dirblk *dbp, int bo)
> > +{
> > + return (sizeof(short) * fs16_to_cpu(bo, dbp->d_nhash)) + 4;
> > +}
> > +
> > +
> > /**
> > * vxfs_find_entry - find a mathing directory entry for a dentry
> > * @ip: directory inode
> > @@ -242,6 +251,8 @@ vxfs_readdir(struct file *fp, struct dir_context *ctx)
> > u_long bsize = sbp->s_blocksize;
> > u_long page, npages, block, pblocks, nblocks, offset;
> > loff_t pos;
> > + int bo = VXFS_SBI(sbp)->byte_order;
> > +
> >
> > if (ctx->pos == 0) {
> > if (!dir_emit_dot(fp, ctx))
> > @@ -297,8 +308,8 @@ vxfs_readdir(struct file *fp, struct dir_context *ctx)
> >
> > offset = (char *)de - kaddr;
> > ctx->pos = ((page << PAGE_CACHE_SHIFT) | offset) + 2;
> > - if (!dir_emit(ctx, de->d_name, de->d_namelen,
> > - de->d_ino, DT_UNKNOWN)) {
> > + if (!dir_emit(ctx, de->d_name, fs16_to_cpu(bo, de->d_namelen),
> > + fs32_to_cpu(bo, de->d_ino), DT_UNKNOWN)) {
> > vxfs_put_page(pp);
> > return 0;
> > }
> > diff --git a/fs/freevxfs/vxfs_olt.c b/fs/freevxfs/vxfs_olt.c
> > index 0495008..6b50188 100644
> > --- a/fs/freevxfs/vxfs_olt.c
> > +++ b/fs/freevxfs/vxfs_olt.c
> > @@ -43,14 +43,14 @@ static inline void
> > vxfs_get_fshead(struct vxfs_oltfshead *fshp, struct vxfs_sb_info *infp)
> > {
> > BUG_ON(infp->vsi_fshino);
> > - infp->vsi_fshino = fshp->olt_fsino[0];
> > + infp->vsi_fshino = fs32_to_cpu(infp->byte_order, fshp->olt_fsino[0]);
> > }
> >
> > static inline void
> > vxfs_get_ilist(struct vxfs_oltilist *ilistp, struct vxfs_sb_info *infp)
> > {
> > BUG_ON(infp->vsi_iext);
> > - infp->vsi_iext = ilistp->olt_iext[0];
> > + infp->vsi_iext = fs32_to_cpu(infp->byte_order, ilistp->olt_iext[0]);
> > }
> >
> > static inline u_long
> > @@ -80,6 +80,7 @@ vxfs_read_olt(struct super_block *sbp, u_long bsize)
> > struct buffer_head *bp;
> > struct vxfs_olt *op;
> > char *oaddr, *eaddr;
> > + int bo = infp->byte_order;
> >
> >
> > bp = sb_bread(sbp, vxfs_oblock(sbp, infp->vsi_oltext, bsize));
> > @@ -87,7 +88,7 @@ vxfs_read_olt(struct super_block *sbp, u_long bsize)
> > goto fail;
> >
> > op = (struct vxfs_olt *)bp->b_data;
> > - if (op->olt_magic != VXFS_OLT_MAGIC) {
> > + if (fs32_to_cpu(bo, op->olt_magic) != VXFS_OLT_MAGIC) {
> > printk(KERN_NOTICE "vxfs: ivalid olt magic number\n");
> > goto fail;
> > }
> > @@ -102,14 +103,14 @@ vxfs_read_olt(struct super_block *sbp, u_long bsize)
> > goto fail;
> > }
> >
> > - oaddr = bp->b_data + op->olt_size;
> > + oaddr = bp->b_data + fs32_to_cpu(bo, op->olt_size);
> > eaddr = bp->b_data + (infp->vsi_oltsize * sbp->s_blocksize);
> >
> > while (oaddr < eaddr) {
> > struct vxfs_oltcommon *ocp =
> > (struct vxfs_oltcommon *)oaddr;
> >
> > - switch (ocp->olt_type) {
> > + switch (fs32_to_cpu(bo, ocp->olt_type)) {
> > case VXFS_OLT_FSHEAD:
> > vxfs_get_fshead((struct vxfs_oltfshead *)oaddr, infp);
> > break;
> > @@ -118,11 +119,11 @@ vxfs_read_olt(struct super_block *sbp, u_long bsize)
> > break;
> > }
> >
> > - oaddr += ocp->olt_size;
> > + oaddr += fs32_to_cpu(bo, ocp->olt_size);
> > }
> >
> > brelse(bp);
> > - return 0;
> > + return (infp->vsi_fshino && infp->vsi_iext) ? 0 : -EINVAL;
> >
> > fail:
> > brelse(bp);
> > diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
> > index 7ca8c75..6a19802 100644
> > --- a/fs/freevxfs/vxfs_super.c
> > +++ b/fs/freevxfs/vxfs_super.c
> > @@ -109,14 +109,15 @@ static int
> > vxfs_statfs(struct dentry *dentry, struct kstatfs *bufp)
> > {
> > struct vxfs_sb_info *infp = VXFS_SBI(dentry->d_sb);
> > + struct vxfs_sb *raw_sb = infp->vsi_raw;
> >
> > bufp->f_type = VXFS_SUPER_MAGIC;
> > bufp->f_bsize = dentry->d_sb->s_blocksize;
> > - bufp->f_blocks = infp->vsi_raw->vs_dsize;
> > - bufp->f_bfree = infp->vsi_raw->vs_free;
> > + bufp->f_blocks = fs32_to_cpu(infp->byte_order, raw_sb->vs_dsize);
> > + bufp->f_bfree = fs32_to_cpu(infp->byte_order, raw_sb->vs_free);
> > bufp->f_bavail = 0;
> > bufp->f_files = 0;
> > - bufp->f_ffree = infp->vsi_raw->vs_ifree;
> > + bufp->f_ffree = fs32_to_cpu(infp->byte_order, raw_sb->vs_ifree);
> > bufp->f_namelen = VXFS_NAMELEN;
> >
> > return 0;
> > @@ -129,6 +130,46 @@ static int vxfs_remount(struct super_block *sb, int *flags, char *data)
> > return 0;
> > }
> >
> > +
> > +static int vxfs_try_sb_magic(struct super_block *sbp, int blk, u32 magic)
> > +{
> > + struct buffer_head *bp;
> > + struct vxfs_sb *rsbp;
> > + struct vxfs_sb_info *infp = VXFS_SBI(sbp);
> > + int rc = -ENOMEM;
> > +
> > + bp = sb_bread(sbp, blk);
> > + do {
> > + if (!bp || !buffer_mapped(bp)) {
> > + if (!infp->silent) {
> > + printk(KERN_WARNING "vxfs: unable to read"
> > + " disk superblock at %d\n", blk);
> > + }
> > + break;
> > + }
> > +
> > + rc = -EINVAL;
> > + rsbp = (struct vxfs_sb *)bp->b_data;
> > + if (rsbp->vs_magic != magic) {
> > + if (!infp->silent)
> > + printk(KERN_NOTICE "vxfs: WRONG superblock magic\n");
> > + break;
> > + }
> > +
> > + rc = 0;
> > + infp->vsi_raw = rsbp;
> > + infp->vsi_bp = bp;
> > + } while (0);
> > +
> > + if (rc) {
> > + infp->vsi_raw = NULL;
> > + infp->vsi_bp = NULL;
> > + brelse(bp);
> > + }
> > +
> > + return rc;
> > +}
> > +
> > /**
> > * vxfs_read_super - read superblock into memory and initialize filesystem
> > * @sbp: VFS superblock (to fill)
> > @@ -149,10 +190,10 @@ static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
> > {
> > struct vxfs_sb_info *infp;
> > struct vxfs_sb *rsbp;
> > - struct buffer_head *bp = NULL;
> > u_long bsize;
> > struct inode *root;
> > int ret = -EINVAL;
> > + u32 j;
> >
> > sbp->s_flags |= MS_RDONLY;
> >
> > @@ -162,48 +203,47 @@ static int vxfs_fill_super(struct super_block *sbp, void *dp, int silent)
> > return -ENOMEM;
> > }
> >
> > + infp->silent = silent;
> > bsize = sb_min_blocksize(sbp, BLOCK_SIZE);
> > if (!bsize) {
> > printk(KERN_WARNING "vxfs: unable to set blocksize\n");
> > goto out;
> > }
> >
> > - bp = sb_bread(sbp, 1);
> > - if (!bp || !buffer_mapped(bp)) {
> > - if (!silent) {
> > - printk(KERN_WARNING
> > - "vxfs: unable to read disk superblock\n");
> > + sbp->s_fs_info = infp;
> > + do {
> > + if (!vxfs_try_sb_magic(sbp, 1, cpu_to_le32(VXFS_SUPER_MAGIC))) {
> > + infp->byte_order = BO_LE; /* SCO */
> > + break;
> > + }
> > +
> > + if (!vxfs_try_sb_magic(sbp, 8, cpu_to_be32(VXFS_SUPER_MAGIC))) {
> > + infp->byte_order = BO_BE; /* HP-UX pa-risc likely */
> > + break;
> > }
> > - goto out;
> > - }
> >
> > - rsbp = (struct vxfs_sb *)bp->b_data;
> > - if (rsbp->vs_magic != VXFS_SUPER_MAGIC) {
> > - if (!silent)
> > - printk(KERN_NOTICE "vxfs: WRONG superblock magic\n");
> > goto out;
> > - }
> > + } while (0);
> >
> > - if ((rsbp->vs_version < 2 || rsbp->vs_version > 4) && !silent) {
> > - printk(KERN_NOTICE "vxfs: unsupported VxFS version (%d)\n",
> > - rsbp->vs_version);
> > + rsbp = infp->vsi_raw;
> > + j = fs32_to_cpu(infp->byte_order, rsbp->vs_version);
> > + if ((j < 2 || j > 4) && !silent) {
> > + printk(KERN_NOTICE "vxfs: unsupported VxFS version (%d)\n", j);
> > goto out;
> > }
> >
> > #ifdef DIAGNOSTIC
> > - printk(KERN_DEBUG "vxfs: supported VxFS version (%d)\n", rsbp->vs_version);
> > + printk(KERN_DEBUG "vxfs: supported VxFS version (%d)\n", j);
> > printk(KERN_DEBUG "vxfs: blocksize: %d\n", rsbp->vs_bsize);
> > #endif
> >
> > - sbp->s_magic = rsbp->vs_magic;
> > - sbp->s_fs_info = infp;
> > + sbp->s_magic = fs32_to_cpu(infp->byte_order, rsbp->vs_magic);
> >
> > - infp->vsi_raw = rsbp;
> > - infp->vsi_bp = bp;
> > - infp->vsi_oltext = rsbp->vs_oltext[0];
> > - infp->vsi_oltsize = rsbp->vs_oltsize;
> > + infp->vsi_oltext = fs32_to_cpu(infp->byte_order, rsbp->vs_oltext[0]);
> > + infp->vsi_oltsize = fs32_to_cpu(infp->byte_order, rsbp->vs_oltsize);
> >
> > - if (!sb_set_blocksize(sbp, rsbp->vs_bsize)) {
> > + j = fs32_to_cpu(infp->byte_order, rsbp->vs_bsize);
> > + if (!sb_set_blocksize(sbp, j)) {
> > printk(KERN_WARNING "vxfs: unable to set final block size\n");
> > goto out;
> > }
> > @@ -237,7 +277,8 @@ out_free_ilist:
> > vxfs_put_fake_inode(infp->vsi_ilist);
> > vxfs_put_fake_inode(infp->vsi_stilist);
> > out:
> > - brelse(bp);
> > + if (infp->vsi_bp)
> > + brelse(infp->vsi_bp);
> > kfree(infp);
> > return ret;
> > }
> > --
> > 1.7.3.4
> >
>
> > From b95e9b749b85be347c953ce33d2b3c5dde4e56d6 Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:13:20 +0200
> > Subject: [PATCH 3/7] missing kfree and kfree on kmem_cache obj
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs_extern.h | 1 +
> > fs/freevxfs/vxfs_fshead.c | 8 +++++---
> > fs/freevxfs/vxfs_inode.c | 5 +++++
> > 3 files changed, 11 insertions(+), 3 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs_extern.h b/fs/freevxfs/vxfs_extern.h
> > index 881aa3d..0a3ff86 100644
> > --- a/fs/freevxfs/vxfs_extern.h
> > +++ b/fs/freevxfs/vxfs_extern.h
> > @@ -64,6 +64,7 @@ extern struct vxfs_inode_info * vxfs_blkiget(struct super_block *, u_long, ino_t
> > extern struct vxfs_inode_info * vxfs_stiget(struct super_block *, ino_t);
> > extern struct inode * vxfs_iget(struct super_block *, ino_t);
> > extern void vxfs_evict_inode(struct inode *);
> > +extern void vxfs_inode_info_free(struct vxfs_inode_info *vip);
> >
> > /* vxfs_lookup.c */
> > extern const struct inode_operations vxfs_dir_inode_ops;
> > diff --git a/fs/freevxfs/vxfs_fshead.c b/fs/freevxfs/vxfs_fshead.c
> > index 6cbdde7..67ca2f9 100644
> > --- a/fs/freevxfs/vxfs_fshead.c
> > +++ b/fs/freevxfs/vxfs_fshead.c
> > @@ -183,7 +183,7 @@ vxfs_read_fshead(struct super_block *sbp)
> > infp->vsi_stilist = vxfs_get_fake_inode(sbp, tip);
> > if (!infp->vsi_stilist) {
> > printk(KERN_ERR "vxfs: unable to get structural list inode\n");
> > - kfree(tip);
> > + vxfs_inode_info_free(tip);
> > goto out_free_pfp;
> > }
> > if (!VXFS_ISILT(VXFS_INO(infp->vsi_stilist))) {
> > @@ -198,7 +198,7 @@ vxfs_read_fshead(struct super_block *sbp)
> > infp->vsi_ilist = vxfs_get_fake_inode(sbp, tip);
> > if (!infp->vsi_ilist) {
> > printk(KERN_ERR "vxfs: unable to get inode list inode\n");
> > - kfree(tip);
> > + vxfs_inode_info_free(tip);
> > goto out_iput_stilist;
> > }
> > if (!VXFS_ISILT(VXFS_INO(infp->vsi_ilist))) {
> > @@ -206,6 +206,8 @@ vxfs_read_fshead(struct super_block *sbp)
> > VXFS_INO(infp->vsi_ilist)->vii_mode & VXFS_TYPE_MASK);
> > goto out_iput_ilist;
> > }
> > + kfree(pfp);
> > + kfree(sfp);
> >
> > return 0;
> >
> > @@ -221,6 +223,6 @@ vxfs_read_fshead(struct super_block *sbp)
> > iput(infp->vsi_fship);
> > return -EINVAL;
> > out_free_fship:
> > - kfree(vip);
> > + vxfs_inode_info_free(vip);
> > return -EINVAL;
> > }
> > diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
> > index 53b8757..16d8d27 100644
> > --- a/fs/freevxfs/vxfs_inode.c
> > +++ b/fs/freevxfs/vxfs_inode.c
> > @@ -415,3 +415,8 @@ vxfs_evict_inode(struct inode *ip)
> > clear_inode(ip);
> > call_rcu(&ip->i_rcu, vxfs_i_callback);
> > }
> > +
> > +void vxfs_inode_info_free(struct vxfs_inode_info *vip)
> > +{
> > + kmem_cache_free(vxfs_inode_cachep, vip);
> > +}
> > --
> > 1.7.3.4
> >
>
> > From e4979c6a29eb79d59d8bc2eca9acc0d2b416780f Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:15:20 +0200
> > Subject: [PATCH 4/7] super_operations.destroy_inode
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs_extern.h | 1 +
> > fs/freevxfs/vxfs_inode.c | 16 ++++++++++++----
> > fs/freevxfs/vxfs_super.c | 1 +
> > 3 files changed, 14 insertions(+), 4 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs_extern.h b/fs/freevxfs/vxfs_extern.h
> > index 0a3ff86..4d8298b 100644
> > --- a/fs/freevxfs/vxfs_extern.h
> > +++ b/fs/freevxfs/vxfs_extern.h
> > @@ -65,6 +65,7 @@ extern struct vxfs_inode_info * vxfs_stiget(struct super_block *, ino_t);
> > extern struct inode * vxfs_iget(struct super_block *, ino_t);
> > extern void vxfs_evict_inode(struct inode *);
> > extern void vxfs_inode_info_free(struct vxfs_inode_info *vip);
> > +extern void vxfs_destroy_inode(struct inode *ip);
> >
> > /* vxfs_lookup.c */
> > extern const struct inode_operations vxfs_dir_inode_ops;
> > diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
> > index 16d8d27..9b45ad7 100644
> > --- a/fs/freevxfs/vxfs_inode.c
> > +++ b/fs/freevxfs/vxfs_inode.c
> > @@ -397,7 +397,15 @@ vxfs_iget(struct super_block *sbp, ino_t ino)
> > static void vxfs_i_callback(struct rcu_head *head)
> > {
> > struct inode *inode = container_of(head, struct inode, i_rcu);
> > - kmem_cache_free(vxfs_inode_cachep, inode->i_private);
> > + void *priv = inode->i_private;
> > +
> > + inode->i_private = NULL;
> > + kmem_cache_free(vxfs_inode_cachep, priv);
> > +}
> > +
> > +void vxfs_destroy_inode(struct inode *ip)
> > +{
> > + call_rcu(&ip->i_rcu, vxfs_i_callback);
> > }
> >
> > /**
> > @@ -405,17 +413,17 @@ static void vxfs_i_callback(struct rcu_head *head)
> > * @ip: inode to discard.
> > *
> > * Description:
> > - * vxfs_evict_inode() is called on the final iput and frees the private
> > - * inode area.
> > + * vxfs_evict_inode() is called on the final iput
> > */
> > void
> > vxfs_evict_inode(struct inode *ip)
> > {
> > truncate_inode_pages_final(&ip->i_data);
> > + invalidate_inode_buffers(ip);
> > clear_inode(ip);
> > - call_rcu(&ip->i_rcu, vxfs_i_callback);
> > }
> >
> > +
> > void vxfs_inode_info_free(struct vxfs_inode_info *vip)
> > {
> > kmem_cache_free(vxfs_inode_cachep, vip);
> > diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
> > index 6a19802..11a535a 100644
> > --- a/fs/freevxfs/vxfs_super.c
> > +++ b/fs/freevxfs/vxfs_super.c
> > @@ -59,6 +59,7 @@ static int vxfs_statfs(struct dentry *, struct kstatfs *);
> > static int vxfs_remount(struct super_block *, int *, char *);
> >
> > static const struct super_operations vxfs_super_ops = {
> > + .destroy_inode = vxfs_destroy_inode,
> > .evict_inode = vxfs_evict_inode,
> > .put_super = vxfs_put_super,
> > .statfs = vxfs_statfs,
> > --
> > 1.7.3.4
> >
>
> > From 4a6ae0e351b6046367bbd2939547f292ed2b6e47 Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:28:37 +0200
> > Subject: [PATCH 5/7] refactoring of vxfs_readir() and find_entry()
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs_lookup.c | 267 +++++++++++++++++++++------------------------
> > 1 files changed, 124 insertions(+), 143 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs_lookup.c b/fs/freevxfs/vxfs_lookup.c
> > index cea158a..8eacb27 100644
> > --- a/fs/freevxfs/vxfs_lookup.c
> > +++ b/fs/freevxfs/vxfs_lookup.c
> > @@ -61,48 +61,6 @@ const struct file_operations vxfs_dir_operations = {
> > .iterate = vxfs_readdir,
> > };
> >
> > -
> > -static inline u_long
> > -dir_pages(struct inode *inode)
> > -{
> > - return (inode->i_size + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
> > -}
> > -
> > -static inline u_long
> > -dir_blocks(struct inode *ip)
> > -{
> > - u_long bsize = ip->i_sb->s_blocksize;
> > - return (ip->i_size + bsize - 1) & ~(bsize - 1);
> > -}
> > -
> > -/*
> > - * NOTE! unlike strncmp, vxfs_match returns 1 for success, 0 for failure.
> > - *
> > - * len <= VXFS_NAMELEN and de != NULL are guaranteed by caller.
> > - */
> > -static inline int
> > -vxfs_match(int len, const char * const name, struct vxfs_direct *de)
> > -{
> > - if (len != de->d_namelen)
> > - return 0;
> > - if (!de->d_ino)
> > - return 0;
> > - return !memcmp(name, de->d_name, len);
> > -}
> > -
> > -static inline struct vxfs_direct *
> > -vxfs_next_entry(struct vxfs_direct *de)
> > -{
> > - return ((struct vxfs_direct *)((char*)de + de->d_reclen));
> > -}
> > -
> > -/*
> > - * VXFS_dirblk_ovh is the overhead of a specific dirblock.
> > - */
> > -static inline u_long VXFS_dirblk_ovh(struct vxfs_dirblk *dbp, int bo)
> > -{
> > - return (sizeof(short) * fs16_to_cpu(bo, dbp->d_nhash)) + 4;
> > -}
> >
> >
> > /**
> > @@ -122,50 +80,65 @@ static inline u_long VXFS_dirblk_ovh(struct vxfs_dirblk *dbp, int bo)
> > static struct vxfs_direct *
> > vxfs_find_entry(struct inode *ip, struct dentry *dp, struct page **ppp)
> > {
> > - u_long npages, page, nblocks, pblocks, block;
> > - u_long bsize = ip->i_sb->s_blocksize;
> > - const char *name = dp->d_name.name;
> > - int namelen = dp->d_name.len;
> > -
> > - npages = dir_pages(ip);
> > - nblocks = dir_blocks(ip);
> > - pblocks = VXFS_BLOCK_PER_PAGE(ip->i_sb);
> > -
> > - for (page = 0; page < npages; page++) {
> > - caddr_t kaddr;
> > - struct page *pp;
> > -
> > - pp = vxfs_get_page(ip->i_mapping, page);
> > - if (IS_ERR(pp))
> > - continue;
> > - kaddr = (caddr_t)page_address(pp);
> > -
> > - for (block = 0; block <= nblocks && block <= pblocks; block++) {
> > - caddr_t baddr, limit;
> > - struct vxfs_dirblk *dbp;
> > - struct vxfs_direct *de;
> > -
> > - baddr = kaddr + (block * bsize);
> > - limit = baddr + bsize - VXFS_DIRLEN(1);
> > -
> > - dbp = (struct vxfs_dirblk *)baddr;
> > - de = (struct vxfs_direct *)(baddr + VXFS_DIRBLKOV(dbp));
> > -
> > - for (; (caddr_t)de <= limit; de = vxfs_next_entry(de)) {
> > - if (!de->d_reclen)
> > - break;
> > - if (!de->d_ino)
> > - continue;
> > - if (vxfs_match(namelen, name, de)) {
> > - *ppp = pp;
> > - return (de);
> > - }
> > + u_long bsize = ip->i_sb->s_blocksize;
> > + const char *name = dp->d_name.name;
> > + int namelen = dp->d_name.len;
> > + loff_t limit = VXFS_DIRROUND(ip->i_size);
> > + struct vxfs_direct *de_exit = NULL;
> > + loff_t pos = 0;
> > + int bo = VXFS_SBI(ip->i_sb)->byte_order;
> > +
> > + while (pos < limit) {
> > + struct page *pp;
> > + char *kaddr;
> > + int pg_ofs = pos & ~PAGE_CACHE_MASK;
> > +
> > + pp = vxfs_get_page(ip->i_mapping, pos >> PAGE_CACHE_SHIFT);
> > + if (IS_ERR(pp)) {
> > + return NULL;
> > + }
> > + kaddr = (char *)page_address(pp);
> > +
> > + while (pg_ofs < PAGE_SIZE && pos < limit) {
> > + struct vxfs_direct *de;
> > +
> > + if ((pos & (bsize - 1)) < 4) {
> > + struct vxfs_dirblk *dbp =
> > + (struct vxfs_dirblk *)(kaddr + (pos & ~PAGE_CACHE_MASK));
> > + int overhead = (sizeof(short) * fs16_to_cpu(bo, dbp->d_nhash)) + 4;
> > +
> > + pos += overhead;
> > + pg_ofs += overhead;
> > + }
> > + de = (struct vxfs_direct *)(kaddr + pg_ofs);
> > +
> > + if (!de->d_reclen) {
> > + pos += bsize - 1;
> > + pos &= ~(bsize - 1);
> > + break;
> > + }
> > +
> > + pg_ofs += fs16_to_cpu(bo, de->d_reclen);
> > + pos += fs16_to_cpu(bo, de->d_reclen);
> > + if (!de->d_ino) {
> > + continue;
> > + }
> > +
> > + if (namelen != fs16_to_cpu(bo, de->d_namelen))
> > + continue;
> > + if (!memcmp(name, de->d_name, namelen)) {
> > + *ppp = pp;
> > + de_exit = de;
> > + break;
> > }
> > }
> > - vxfs_put_page(pp);
> > + if (!de_exit)
> > + vxfs_put_page(pp);
> > + else
> > + break;
> > }
> >
> > - return NULL;
> > + return de_exit;
> > }
> >
> > /**
> > @@ -185,15 +158,17 @@ vxfs_inode_by_name(struct inode *dip, struct dentry *dp)
> > {
> > struct vxfs_direct *de;
> > struct page *pp;
> > - ino_t ino = 0;
> > + ino_t ino = 0;
> >
> > de = vxfs_find_entry(dip, dp, &pp);
> > if (de) {
> > - ino = de->d_ino;
> > + int bo = VXFS_SBI(dip->i_sb)->byte_order;
> > +
> > + ino = fs32_to_cpu(bo, de->d_ino);
> > kunmap(pp);
> > page_cache_release(pp);
> > }
> > -
> > +
> > return (ino);
> > }
> >
> > @@ -225,8 +200,8 @@ vxfs_lookup(struct inode *dip, struct dentry *dp, unsigned int flags)
> > ip = vxfs_iget(dip->i_sb, ino);
> > if (IS_ERR(ip))
> > return ERR_CAST(ip);
> > + d_add(dp, ip);
> > }
> > - d_add(dp, ip);
> > return NULL;
> > }
> >
> > @@ -249,76 +224,82 @@ vxfs_readdir(struct file *fp, struct dir_context *ctx)
> > struct inode *ip = file_inode(fp);
> > struct super_block *sbp = ip->i_sb;
> > u_long bsize = sbp->s_blocksize;
> > - u_long page, npages, block, pblocks, nblocks, offset;
> > - loff_t pos;
> > + loff_t pos, limit;
> > int bo = VXFS_SBI(sbp)->byte_order;
> >
> > -
> > if (ctx->pos == 0) {
> > if (!dir_emit_dot(fp, ctx))
> > - return 0;
> > - ctx->pos = 1;
> > + goto out;
> > + ctx->pos++;
> > }
> > if (ctx->pos == 1) {
> > if (!dir_emit(ctx, "..", 2, VXFS_INO(ip)->vii_dotdot, DT_DIR))
> > - return 0;
> > - ctx->pos = 2;
> > + goto out;
> > + ctx->pos++;
> > + }
> > +
> > + limit = VXFS_DIRROUND(ip->i_size);
> > + if (ctx->pos > limit) {
> > +#if 0
> > + ctx->pos = 0;
> > +#endif
> > + goto out;
> > }
> > - pos = ctx->pos - 2;
> > -
> > - if (pos > VXFS_DIRROUND(ip->i_size))
> > - return 0;
> > -
> > - npages = dir_pages(ip);
> > - nblocks = dir_blocks(ip);
> > - pblocks = VXFS_BLOCK_PER_PAGE(sbp);
> > -
> > - page = pos >> PAGE_CACHE_SHIFT;
> > - offset = pos & ~PAGE_CACHE_MASK;
> > - block = (u_long)(pos >> sbp->s_blocksize_bits) % pblocks;
> > -
> > - for (; page < npages; page++, block = 0) {
> > - char *kaddr;
> > - struct page *pp;
> > -
> > - pp = vxfs_get_page(ip->i_mapping, page);
> > - if (IS_ERR(pp))
> > - continue;
> > +
> > + pos = ctx->pos & ~3L;
> > +
> > + while (pos < limit) {
> > + struct page *pp;
> > + char *kaddr;
> > + int pg_ofs = pos & ~PAGE_CACHE_MASK;
> > + int rc = 0;
> > +
> > + pp = vxfs_get_page(ip->i_mapping, pos >> PAGE_CACHE_SHIFT);
> > + if (IS_ERR(pp)) {
> > + return -ENOMEM;
> > + }
> > kaddr = (char *)page_address(pp);
> >
> > - for (; block <= nblocks && block <= pblocks; block++) {
> > - char *baddr, *limit;
> > - struct vxfs_dirblk *dbp;
> > - struct vxfs_direct *de;
> > -
> > - baddr = kaddr + (block * bsize);
> > - limit = baddr + bsize - VXFS_DIRLEN(1);
> > -
> > - dbp = (struct vxfs_dirblk *)baddr;
> > - de = (struct vxfs_direct *)
> > - (offset ?
> > - (kaddr + offset) :
> > - (baddr + VXFS_DIRBLKOV(dbp)));
> > -
> > - for (; (char *)de <= limit; de = vxfs_next_entry(de)) {
> > - if (!de->d_reclen)
> > - break;
> > - if (!de->d_ino)
> > - continue;
> > -
> > - offset = (char *)de - kaddr;
> > - ctx->pos = ((page << PAGE_CACHE_SHIFT) | offset) + 2;
> > - if (!dir_emit(ctx, de->d_name, fs16_to_cpu(bo, de->d_namelen),
> > - fs32_to_cpu(bo, de->d_ino), DT_UNKNOWN)) {
> > - vxfs_put_page(pp);
> > - return 0;
> > - }
> > + while (pg_ofs < PAGE_SIZE && pos < limit) {
> > + struct vxfs_direct *de;
> > +
> > + if ((pos & (bsize - 1)) < 4) {
> > + struct vxfs_dirblk *dbp =
> > + (struct vxfs_dirblk *)(kaddr + (pos & ~PAGE_CACHE_MASK));
> > + int overhead = (sizeof(short) * fs16_to_cpu(bo, dbp->d_nhash)) + 4;
> > +
> > + pos += overhead;
> > + pg_ofs += overhead;
> > + }
> > + de = (struct vxfs_direct *)(kaddr + pg_ofs);
> > +
> > + if (!de->d_reclen) {
> > + pos += bsize - 1;
> > + pos &= ~(bsize - 1);
> > + break;
> > + }
> > +
> > + pg_ofs += fs16_to_cpu(bo, de->d_reclen);
> > + pos += fs16_to_cpu(bo, de->d_reclen);
> > + if (!de->d_ino) {
> > + continue;
> > + }
> > +
> > + rc = dir_emit(ctx, de->d_name, fs16_to_cpu(bo, de->d_namelen),
> > + fs32_to_cpu(bo, de->d_ino), DT_UNKNOWN);
> > + if (!rc) {
> > + /* the dir entry was not submitted, so fix pos. */
> > + pos -= fs16_to_cpu(bo, de->d_reclen);
> > + break;
> > }
> > - offset = 0;
> > }
> > vxfs_put_page(pp);
> > - offset = 0;
> > + if (!rc)
> > + break;
> > }
> > - ctx->pos = ((page << PAGE_CACHE_SHIFT) | offset) + 2;
> > +
> > + ctx->pos = pos | 2;
> > +
> > +out:
> > return 0;
> > }
> > --
> > 1.7.3.4
> >
>
> > From e7f68291aada1535016c767a4f5a5fcfb19a1de6 Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:41:22 +0200
> > Subject: [PATCH 6/7] static cachep
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs_extern.h | 4 +++-
> > fs/freevxfs/vxfs_inode.c | 24 +++++++++++++++++++++++-
> > fs/freevxfs/vxfs_super.c | 26 ++++++++++----------------
> > 3 files changed, 36 insertions(+), 18 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs_extern.h b/fs/freevxfs/vxfs_extern.h
> > index 4d8298b..cc43fd0 100644
> > --- a/fs/freevxfs/vxfs_extern.h
> > +++ b/fs/freevxfs/vxfs_extern.h
> > @@ -55,7 +55,6 @@ extern const struct inode_operations vxfs_immed_symlink_iops;
> >
> > /* vxfs_inode.c */
> > extern const struct address_space_operations vxfs_immed_aops;
> > -extern struct kmem_cache *vxfs_inode_cachep;
> > extern void vxfs_dumpi(struct vxfs_inode_info *, ino_t);
> > extern struct inode * vxfs_get_fake_inode(struct super_block *,
> > struct vxfs_inode_info *);
> > @@ -66,6 +65,9 @@ extern struct inode * vxfs_iget(struct super_block *, ino_t);
> > extern void vxfs_evict_inode(struct inode *);
> > extern void vxfs_inode_info_free(struct vxfs_inode_info *vip);
> > extern void vxfs_destroy_inode(struct inode *ip);
> > +extern int vxfs_ii_cache_init(void);
> > +extern void vxfs_ii_cache_destroy(void);
> > +
> >
> > /* vxfs_lookup.c */
> > extern const struct inode_operations vxfs_dir_inode_ops;
> > diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
> > index 9b45ad7..73ac417 100644
> > --- a/fs/freevxfs/vxfs_inode.c
> > +++ b/fs/freevxfs/vxfs_inode.c
> > @@ -41,7 +41,7 @@
> > #include "vxfs_extern.h"
> >
> >
> > -struct kmem_cache *vxfs_inode_cachep;
> > +static struct kmem_cache *vxfs_inode_cachep;
> >
> >
> > #ifdef DIAGNOSTIC
> > @@ -428,3 +428,25 @@ void vxfs_inode_info_free(struct vxfs_inode_info *vip)
> > {
> > kmem_cache_free(vxfs_inode_cachep, vip);
> > }
> > +
> > +
> > +int vxfs_ii_cache_init(void)
> > +{
> > + vxfs_inode_cachep = kmem_cache_create("vxfs_inode",
> > + sizeof(struct vxfs_inode_info), 0,
> > + SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL);
> > +
> > + return vxfs_inode_cachep ? 0 : -ENOMEM;
> > +}
> > +
> > +
> > +void vxfs_ii_cache_destroy(void)
> > +{
> > + /*
> > + * Make sure all delayed rcu free inodes are flushed before we
> > + * destroy cache.
> > + */
> > + rcu_barrier();
> > + kmem_cache_destroy(vxfs_inode_cachep);
> > +}
> > +
> > diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
> > index 11a535a..7579500 100644
> > --- a/fs/freevxfs/vxfs_super.c
> > +++ b/fs/freevxfs/vxfs_super.c
> > @@ -306,29 +306,23 @@ MODULE_ALIAS("vxfs");
> > static int __init
> > vxfs_init(void)
> > {
> > - int rv;
> > + int rc = vxfs_ii_cache_init();
> >
> > - vxfs_inode_cachep = kmem_cache_create("vxfs_inode",
> > - sizeof(struct vxfs_inode_info), 0,
> > - SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL);
> > - if (!vxfs_inode_cachep)
> > - return -ENOMEM;
> > - rv = register_filesystem(&vxfs_fs_type);
> > - if (rv < 0)
> > - kmem_cache_destroy(vxfs_inode_cachep);
> > - return rv;
> > + if (!rc) {
> > + rc = register_filesystem(&vxfs_fs_type);
> > + if (rc < 0)
> > + vxfs_ii_cache_destroy();
> > + }
> > + printk(KERN_DEBUG "%s: **** %s %s rc %d\n", __func__, __DATE__, __TIME__, rc);
> > +
> > + return rc;
> > }
> >
> > static void __exit
> > vxfs_cleanup(void)
> > {
> > unregister_filesystem(&vxfs_fs_type);
> > - /*
> > - * Make sure all delayed rcu free inodes are flushed before we
> > - * destroy cache.
> > - */
> > - rcu_barrier();
> > - kmem_cache_destroy(vxfs_inode_cachep);
> > + vxfs_ii_cache_destroy();
> > }
> >
> > module_init(vxfs_init);
> > --
> > 1.7.3.4
> >
>
> > From 2c0008e9f2f4e62bb3e10eecad54e2a0140e0c4c Mon Sep 17 00:00:00 2001
> > From: KB <kb@sysmikro.com.pl>
> > Date: Wed, 25 May 2016 22:58:08 +0200
> > Subject: [PATCH 7/7] the credits
> >
> >
> > Signed-off-by: KB <kb@sysmikro.com.pl>
> > ---
> > fs/freevxfs/vxfs.h | 3 +++
> > fs/freevxfs/vxfs_fshead.c | 4 ++++
> > fs/freevxfs/vxfs_inode.c | 4 ++++
> > fs/freevxfs/vxfs_lookup.c | 4 ++++
> > fs/freevxfs/vxfs_super.c | 6 +++++-
> > 5 files changed, 20 insertions(+), 1 deletions(-)
> >
> > diff --git a/fs/freevxfs/vxfs.h b/fs/freevxfs/vxfs.h
> > index 5dc8949..35f56b7 100644
> > --- a/fs/freevxfs/vxfs.h
> > +++ b/fs/freevxfs/vxfs.h
> > @@ -2,6 +2,9 @@
> > * Copyright (c) 2000-2001 Christoph Hellwig.
> > * All rights reserved.
> > *
> > + * (c) 2016 Krzysztof Blaszkowski
> > + * Many bug fixes, improvements & tests with HP-UX B.10.20 (pa-risc)
> > + *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > diff --git a/fs/freevxfs/vxfs_fshead.c b/fs/freevxfs/vxfs_fshead.c
> > index 67ca2f9..44b87d0 100644
> > --- a/fs/freevxfs/vxfs_fshead.c
> > +++ b/fs/freevxfs/vxfs_fshead.c
> > @@ -2,6 +2,10 @@
> > * Copyright (c) 2000-2001 Christoph Hellwig.
> > * All rights reserved.
> > *
> > + *
> > + * (c) 2016 Krzysztof Blaszkowski
> > + * Many bug fixes, improvements & tests with HP-UX B.10.20 (pa-risc)
> > + *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > diff --git a/fs/freevxfs/vxfs_inode.c b/fs/freevxfs/vxfs_inode.c
> > index 73ac417..4c8a625 100644
> > --- a/fs/freevxfs/vxfs_inode.c
> > +++ b/fs/freevxfs/vxfs_inode.c
> > @@ -2,6 +2,10 @@
> > * Copyright (c) 2000-2001 Christoph Hellwig.
> > * All rights reserved.
> > *
> > + *
> > + * (c) 2016 Krzysztof Blaszkowski
> > + * Many bug fixes, improvements & tests with HP-UX B.10.20 (pa-risc)
> > + *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > diff --git a/fs/freevxfs/vxfs_lookup.c b/fs/freevxfs/vxfs_lookup.c
> > index 8eacb27..173aeea 100644
> > --- a/fs/freevxfs/vxfs_lookup.c
> > +++ b/fs/freevxfs/vxfs_lookup.c
> > @@ -2,6 +2,10 @@
> > * Copyright (c) 2000-2001 Christoph Hellwig.
> > * All rights reserved.
> > *
> > + *
> > + * (c) 2016 Krzysztof Blaszkowski
> > + * Many bug fixes, improvements & tests with HP-UX B.10.20 (pa-risc)
> > + *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
> > index 7579500..bd121aa 100644
> > --- a/fs/freevxfs/vxfs_super.c
> > +++ b/fs/freevxfs/vxfs_super.c
> > @@ -2,6 +2,10 @@
> > * Copyright (c) 2000-2001 Christoph Hellwig.
> > * All rights reserved.
> > *
> > + *
> > + * (c) 2016 Krzysztof Blaszkowski
> > + * Many bug fixes, improvements & tests with HP-UX B.10.20 (pa-risc)
> > + *
> > * Redistribution and use in source and binary forms, with or without
> > * modification, are permitted provided that the following conditions
> > * are met:
> > @@ -48,7 +52,7 @@
> > #include "vxfs_inode.h"
> >
> >
> > -MODULE_AUTHOR("Christoph Hellwig");
> > +MODULE_AUTHOR("Christoph Hellwig, Krzysztof Blaszkowski");
> > MODULE_DESCRIPTION("Veritas Filesystem (VxFS) driver");
> > MODULE_LICENSE("Dual BSD/GPL");
> >
> > --
> > 1.7.3.4
> >
>
>
> > 3531594148 32829 /usr/sprockets/bin/bw_map
> > 3973234498 518896 /usr/sprockets/bin/cifslist
> > 3973234498 518896 /usr/sprockets/bin/cifslogout
> > 4171014129 518904 /usr/sprockets/bin/cifsmount
> > 4171014129 518904 /usr/sprockets/bin/cifsumount
> > 256627809 1171456 /usr/sprockets/bin/convert
> > 4193697053 242843 /usr/sprockets/bin/createLicensingFiles
> > 2780783712 51 /usr/sprockets/bin/createSprocFile
> > 3285775584 825 /usr/sprockets/bin/dcexec
> > 4073641898 94 /usr/sprockets/bin/defaultResources
> > 4039074287 6610 /usr/sprockets/bin/disableNIS
> > 313350104 20480 /usr/sprockets/bin/dos2ux
> > 1872597524 65536 /usr/sprockets/bin/doschmod
> > 1872597524 65536 /usr/sprockets/bin/doscp
> > 1872597524 65536 /usr/sprockets/bin/dosdf
> > 1872597524 65536 /usr/sprockets/bin/doslib
> > 1872597524 65536 /usr/sprockets/bin/dosll
> > 1872597524 65536 /usr/sprockets/bin/dosls
> > 1872597524 65536 /usr/sprockets/bin/dosmkdir
> > 1872597524 65536 /usr/sprockets/bin/dosrm
> > 1872597524 65536 /usr/sprockets/bin/dosrmdir
> > 1968087462 7112 /usr/sprockets/bin/enableNIS
> > 3282246790 360 /usr/sprockets/bin/endexec
> > 2151751185 553 /usr/sprockets/bin/endsession
> > 3927613487 1115 /usr/sprockets/bin/env_var_sub
> > 3738550324 157 /usr/sprockets/bin/format_dos
> > 3888232633 172649 /usr/sprockets/bin/fppv
> > 2279585942 466 /usr/sprockets/bin/getHostname
> > 2946778792 5648 /usr/sprockets/bin/getSharedConnections
> > 3306388627 20480 /usr/sprockets/bin/ibmsync
> > 3041502972 28781 /usr/sprockets/bin/ipconfig700
> > 3113659230 391815 /usr/sprockets/bin/kpp
> > 908094753 24630 /usr/sprockets/bin/kxst
> > 1329462070 36864 /usr/sprockets/bin/lanload
> > 1015322628 36947 /usr/sprockets/bin/listener
> > 486757159 339968 /usr/sprockets/bin/lmgrd
> > 2620320940 339968 /usr/sprockets/bin/lmutil
> > 951644065 243039 /usr/sprockets/bin/mattrib
> > 951644065 243039 /usr/sprockets/bin/mbadblocks
> > 951644065 243039 /usr/sprockets/bin/mcd
> > 3650177936 951 /usr/sprockets/bin/mcheck
> > 514586869 56 /usr/sprockets/bin/mcomp
> > 951644065 243039 /usr/sprockets/bin/mcopy
> > 951644065 243039 /usr/sprockets/bin/mdel
> > 951644065 243039 /usr/sprockets/bin/mdeltree
> > 951644065 243039 /usr/sprockets/bin/mdir
> > 951644065 243039 /usr/sprockets/bin/mformat
> > 951644065 243039 /usr/sprockets/bin/minfo
> > 951644065 243039 /usr/sprockets/bin/mlabel
> > 951644065 243039 /usr/sprockets/bin/mmd
> > 951644065 243039 /usr/sprockets/bin/mmount
> > 951644065 243039 /usr/sprockets/bin/mmove
> > 951644065 243039 /usr/sprockets/bin/mpartition
> > 662030829 1478557 /usr/sprockets/bin/mpv
> > 951644065 243039 /usr/sprockets/bin/mrd
> > 951644065 243039 /usr/sprockets/bin/mread
> > 951644065 243039 /usr/sprockets/bin/mren
> > 3443865238 32829 /usr/sprockets/bin/msgintcpt
> > 951644065 243039 /usr/sprockets/bin/mtools
> > 951644065 243039 /usr/sprockets/bin/mtoolstest
> > 951644065 243039 /usr/sprockets/bin/mtype
> > 951644065 243039 /usr/sprockets/bin/mwrite
> > 1915353244 82 /usr/sprockets/bin/mxtar
> > 951644065 243039 /usr/sprockets/bin/mzip
> > 2033389047 379158 /usr/sprockets/bin/nmbd
> > 1043940798 32843 /usr/sprockets/bin/nsbuild
> > 421327339 32843 /usr/sprockets/bin/nsdump
> > 2156955828 32843 /usr/sprockets/bin/nsquery
> > 1030470949 28737 /usr/sprockets/bin/pciinfo
> > 889541250 61560 /usr/sprockets/bin/pcipeek
> > 1431205578 28737 /usr/sprockets/bin/pciprog
> > 3351126604 126976 /usr/sprockets/bin/pkunzip
> > 3351126604 126976 /usr/sprockets/bin/pkzip
> > 2166142790 53390 /usr/sprockets/bin/probeload
> > 694305221 41144 /usr/sprockets/bin/progflash
> > 1141035887 445257 /usr/sprockets/bin/pv
> > 974426301 3882 /usr/sprockets/bin/runnetscape
> > 1879287959 1801 /usr/sprockets/bin/running
> > 2273866242 74107 /usr/sprockets/bin/runroot
> > 2616587298 79 /usr/sprockets/bin/runTouch
> > 4222109660 875 /usr/sprockets/bin/samba
> > 1912440105 24633 /usr/sprockets/bin/sbscrsvr
> > 2558012300 3904 /usr/sprockets/bin/screensaverControl
> > 4256537430 1780 /usr/sprockets/bin/sd-set_parms
> > 1151924416 688 /usr/sprockets/bin/sendSignal
> > 2744646624 32829 /usr/sprockets/bin/sessionDisallow
> > 1227674200 1346161 /usr/sprockets/bin/sessionMgr
> > 3332668870 98791 /usr/sprockets/bin/sessionWrapper
> > 1477797090 82487 /usr/sprockets/bin/setlight
> > 3759253613 1122 /usr/sprockets/bin/sharity
> > 3313588847 3448000 /usr/sprockets/bin/sharityd
> > 1569390236 24630 /usr/sprockets/bin/shmLock
> > 3791448153 317300 /usr/sprockets/bin/smbclient
> > 1018298877 506532 /usr/sprockets/bin/smbd
> > 4009501079 45140 /usr/sprockets/bin/smtpclient
> > 991996913 144424 /usr/sprockets/bin/starhw
> > 3124681867 4311 /usr/sprockets/bin/startVNC
> > 1468091312 919 /usr/sprockets/bin/startVNCviewer
> > 304853428 1994350 /usr/sprockets/bin/sw_admin
> > 304853428 1994350 /usr/sprockets/bin/sw_autoinstall
> > 1455142555 123406 /usr/sprockets/bin/sw_copyfdset
> > 2157175922 111102 /usr/sprockets/bin/sw_depot2bin
> > 304853428 1994350 /usr/sprockets/bin/sw_install
> > 304853428 1994350 /usr/sprockets/bin/sw_list
> > 1842505403 7100 /usr/sprockets/bin/sw_makefdset
> > 304853428 1994350 /usr/sprockets/bin/sw_remove
> > 1743298644 2342 /usr/sprockets/bin/sw_rmprevious
> > 1582083858 20548 /usr/sprockets/bin/sw_vendorfdset
> > 178332129 1794 /usr/sprockets/bin/sysInfo
> > 3657811479 1574 /usr/sprockets/bin/tgz
> > 4018370132 36951 /usr/sprockets/bin/tp_license
> > 1649141155 11843 /usr/sprockets/bin/updsharitycfg
> > 2980018004 20480 /usr/sprockets/bin/ux2dos
> > 3745508150 1318 /usr/sprockets/bin/uz
> > 1348652714 32864 /usr/sprockets/bin/vncpasswd
> > 4059637159 12757 /usr/sprockets/bin/vncserver
> > 4128771573 366263 /usr/sprockets/bin/vncviewer
> > 285011189 1526819 /usr/sprockets/bin/vp
> > 589630079 387 /usr/sprockets/bin/vpexit
> > 528762950 1132 /usr/sprockets/bin/xcopy
> > 1793380080 521816 /usr/sprockets/bin/xscreensaver
> > 24739470 145832 /usr/sprockets/bin/xscreensaver-command
> > 4079099643 2167483 /usr/sprockets/bin/Xvnc
> > 49651643 41091 /usr/sprockets/bin/yakdebug
> > 2578809639 127914 /usr/sprockets/bin/yakhw
> > 306877946 189649 /usr/sprockets/bin/yakserver
>
>
>
--
Krzysztof Blaszkowski
next prev parent reply other threads:[~2016-05-26 14:44 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-25 21:27 freevxfs Krzysztof Błaszkowski
2016-05-26 11:50 ` freevxfs Carlos Maiolino
2016-05-26 14:44 ` Krzysztof Błaszkowski [this message]
-- strict thread matches above, loose matches on Subject: below --
2016-05-22 15:13 freevxfs Krzysztof Błaszkowski
2016-05-23 8:23 ` freevxfs Carlos Maiolino
2016-05-24 11:50 ` freevxfs Krzysztof Błaszkowski
2016-05-23 8:36 ` freevxfs Christoph Hellwig
2016-05-24 8:48 ` freevxfs Krzysztof Błaszkowski
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1464273846.17980.14.camel@linux-q3cb.site \
--to=kb@sysmikro.com.pl \
--cc=cmaiolino@redhat.com \
--cc=linux-fsdevel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).