Re: [PATCH 2/5] jffs2: Add LZO compression support to jffs2

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 2007-05-04 at 14:36 +0530, Satyam Sharma wrote:
> On 5/1/07, Richard Purdie <[email protected]> wrote:
> > +++ b/fs/jffs2/compr_lzo.c
> > [...]
> > +static void *lzo_mem;
> > +static void *lzo_compress_buf;
> > +static DEFINE_MUTEX(deflate_mutex);
> > +
> > +static void free_workspace(void)
> > +{
> > +       vfree(lzo_mem);
> > +       vfree(lzo_compress_buf);
> > +}
> > +
> > +static int __init alloc_workspace(void)
> > +{
> > +       lzo_mem = vmalloc(LZO1X_MEM_COMPRESS);
> > +       lzo_compress_buf = vmalloc(lzo1x_worst_compress(PAGE_SIZE));
> > +
> > +       if (!lzo_mem || !lzo_compress_buf) {
> > +               printk(KERN_WARNING "Failed to allocate lzo deflate workspace\n");
> > +               free_workspace();
> > +               return -ENOMEM;
> > +       }
> > +
> > +       return 0;
> > +}
> > +
> > +static int jffs2_lzo_compress(unsigned char *data_in, unsigned char *cpage_out,
> > +                             uint32_t *sourcelen, uint32_t *dstlen, void *model)
> > +{
> > +       unsigned long compress_size;
> > +       int ret;
> > +
> > +       mutex_lock(&deflate_mutex);
> > +       ret = lzo1x_1_compress(data_in, *sourcelen, lzo_compress_buf, &compress_size, lzo_mem);
> > +       mutex_unlock(&deflate_mutex);
> 
> Considering we do have to memcpy() the entire compressed result to the
> destination output buffer later anyway (note that
> fs/jffs2/compr_zlib.c doesn't need to do that), do we really gain much
> by avoiding vmalloc() and vfree() in jffs2_lzo_compress() itself and
> keeping the workspace buffers pre-allocated? I ask because I always
> found these global static workspace buffers ugly, and all the
> associated code + mutex could go away if we make them local to
> jffs2_lzo_compress() -- as long as it doesn't hurt performance
> terribly, of course.

memcpy is relatively fast and I'd expect continually allocing and
freeing buffers to have a significant overhead compared with that and
impact performance. The current approach means you don't get ENOMEM
errors in the uncompress/compress paths either.

Regards,

Richard


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux