Hi Weiwen,
On Mon, Jan 18, 2021 at 09:02:16PM +0800, 胡玮文 wrote:
>
> > 在 2021年1月16日,14:41,Gao Xiang 写道:
> >
> > From: Hu Weiwen
> >
> > When using EROFS to pack our dataset which consists of millions of
> > files, mkfs.erofs is very slow compared with mksquashfs.
> >
> > The bottleneck is
Hi Weiwen,
On Mon, Jan 18, 2021 at 08:29:23PM +0800, 胡玮文 wrote:
> Hi Xiang,
>
> Thank you, your patch is indeed clearer. Could you explain why you don’t like
> my added erofs_dbg? I found them very useful when debugging.
I do like your erofs_dbg, but we might need to shift(e.g. rename to
> 在 2021年1月16日,14:41,Gao Xiang 写道:
>
> From: Hu Weiwen
>
> When using EROFS to pack our dataset which consists of millions of
> files, mkfs.erofs is very slow compared with mksquashfs.
>
> The bottleneck is `erofs_balloc` and `erofs_mapbh` function, which
> iterate over all previously
Hi Xiang,
Thank you, your patch is indeed clearer. Could you explain why you don’t like
my added erofs_dbg? I found them very useful when debugging.
I’ve updated the commit message, with some fixes (see inline) in PATCH v5
(coming soon).
> 在 2021年1月16日,14:41,Gao Xiang 写道:
>
> From: Hu
From: Hu Weiwen
When using EROFS to pack our dataset which consists of millions of
files, mkfs.erofs is very slow compared with mksquashfs.
The bottleneck is `erofs_balloc` and `erofs_mapbh` function, which
iterate over all previously allocated buffer blocks, making the
complexity of the