Re: [PATCH v2] diskdump: Optimize the boot time

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

 



On Mon, Apr 11, 2022 at 12:30 PM HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab@xxxxxxx> wrote:
Hi Lianbo,

-----Original Message-----
>       diff --git a/defs.h b/defs.h
>       index 81ac049..1e8360d 100644
>       --- a/defs.h
>       +++ b/defs.h
>       @@ -4531,6 +4531,26 @@ struct machine_specific {
>        #define NUM_IN_BITMAP(bitmap, x) (bitmap[(x)/BITS_PER_LONG] & NUM_TO_BIT(x))
>        #define SET_BIT(bitmap, x) (bitmap[(x)/BITS_PER_LONG] |= NUM_TO_BIT(x))
>
>       +static inline unsigned int __const_hweight8(unsigned long w)
>       +{
>       +       return
>       +               (!!((w) & (1ULL << 0))) +
>       +               (!!((w) & (1ULL << 1))) +
>       +               (!!((w) & (1ULL << 2))) +
>       +               (!!((w) & (1ULL << 3))) +
>       +               (!!((w) & (1ULL << 4))) +
>       +               (!!((w) & (1ULL << 5))) +
>       +               (!!((w) & (1ULL << 6))) +
>       +               (!!((w) & (1ULL << 7)));
>       +}
>       +
>       +#define __const_hweight16(w) (__const_hweight8(w)  + __const_hweight8((w)  >> 8))
>       +#define __const_hweight32(w) (__const_hweight16(w) + __const_hweight16((w) >> 16))
>       +#define __const_hweight64(w) (__const_hweight32(w) + __const_hweight32((w) >> 32))
>       +
>       +#define hweight32(w) __const_hweight32(w)
>       +#define hweight64(w) __const_hweight64(w)
>       +
>
>
>
> No need to move the above code from sbitmap.c to defs.h, a simple way is to implement a new function in
> sbitmap.c and add its definition in defs.h, that will
> be easy to call it in diskdump.c. For example:
>
> diff --git a/defs.h b/defs.h
> index 81ac0498dac7..0c5115e71f1c 100644
> --- a/defs.h
> +++ b/defs.h
> @@ -5894,6 +5894,7 @@ typedef bool (*sbitmap_for_each_fn)(unsigned int idx, void *p);
>  void sbitmap_for_each_set(const struct sbitmap_context *sc,
>   sbitmap_for_each_fn fn, void *data);
>  void sbitmap_context_load(ulong addr, struct sbitmap_context *sc);
> +unsigned long get_hweight64(unsigned long w);
>
>  /* sbitmap_queue helpers */
>  typedef bool (*sbitmapq_for_each_fn)(unsigned int idx, ulong addr, void *p);
> diff --git a/sbitmap.c b/sbitmap.c
> index 286259f71d64..628cc00c0b6b 100644
> --- a/sbitmap.c
> +++ b/sbitmap.c
> @@ -71,6 +71,11 @@ static inline unsigned int __const_hweight8(unsigned long w)
>
>  #define BIT(nr) (1UL << (nr))
>
> +unsigned long get_hweight64(unsigned long w)
> +{
> + return hweight64(w);
> +}
> +
>  static inline unsigned long min(unsigned long a, unsigned long b)
>  {
>   return (a < b) ? a : b;
>
>
> //diskdump.c
> ...
> dd->valid_pages[i] += get_hweight64(tmp);
>
> ...
>
> How about the above suggestions? Shijie and Kazu.

Thanks for the suggestion, but personally I don't think it has more
benefits than moving them.  What is the good point?

 It has fewer code changes. The bitmap operation can be maintained together
 in the sbitmap.c and won't be scattered elsewhere.

In the future, some new functions may be still extended for the bitmap operations in the sbitmap.c, that will avoid adding more bitmap operations to defs.h.

That's my concern. If that is not a problem, the v2 will be fine to me. :-)

It has extra function calls and the difference of callee between
hweight64() in sbitmap.c and get_hweight64() in diskdump.c, IMO
it's not simple.
 
You are right. It has extra function calls.

Thanks.
Lianbo

Thanks,
Kazu


--
Crash-utility mailing list
Crash-utility@xxxxxxxxxx
https://listman.redhat.com/mailman/listinfo/crash-utility
Contribution Guidelines: https://github.com/crash-utility/crash/wiki

[Index of Archives]     [Fedora Development]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [KDE Users]     [Fedora Tools]

 

Powered by Linux