Re: Degraded performance in cat + patch

[Jim Meyering]

Pádraig Brady wrote:
> I was thinking a bit more about the patch just pushed.
> It sets the buffer size to 8*st_blksize which seems a
> little arbitrary, and also max buffer size is set to
> 32KiB even if the file system has a larger st_blksize.
> I'm not sure this is desired?
>
> How about making 32KiB the minimum as in the attached?
> The patch also changes `split` and `copy` to use the
> same IO size as `cat`.
...
> +#define IO_BLKSIZE(statbuf) (MAX(32*1024, ST_BLKSIZE(statbuf)))

That looks better.  Thanks.

But please make it a static inline function, not a macro.
That will be more debugger-friendly.