Re: [PATCH] commit: skip already cleared parents in clear_commit_marks_1()

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

 



On 12/13/22 1:27 AM, René Scharfe wrote:
> Don't put clean parents on the pending list, as they and their ancestors
> don't need any treatment and would be skipped later anyway.  This saves
> the allocation and release of a commit list item in ca. 20% of the cases
> during a run of the test suite.

> -		while ((parents = parents->next))
> -			commit_list_insert(parents->item, plist);
> +		while ((parents = parents->next)) {
> +			if (parents->item->object.flags & mark)
> +				commit_list_insert(parents->item, plist);
> +		}

>From this context, it looks like this is a change of behavior (the
walk only continues to parents that have the flag, so if the flag
is not closed under reachability, such as the UNINTERESTING flag,
then we won't clear all flags) but above this we already skip a
commit that doesn't have the flag enabled.

One thing that is odd about this method is that it sets the
'commit' value to the first parent, and adds the later parents
to the list. It's a minor point, but that check for the flag
on the 'commit' is what prevents us from walking too far along
the first-parent history.

All this to say, this is a safe change with a nice upside.

Thanks,
-Stolee



[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux