Re: [Lsf-pc] [LSF/MM/BPF TOPIC] Parallelizing filesystem writeback

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

 



> I think xarray here is overkill. I'd just make this a plain array:
> 
> 	struct bdi_writeback_ctx **bdi_wb_ctx_arr;
> 
> which will get allocated with nr_wb_ctx entries during bdi_init(). Also I'd
> make default_ctx just be entry at index 0 in this array. I'm undecided
> whether it will be clearer to just drop default_ctx altogether or keep it
> and set:
> 
> 	struct bdi_writeback_ctx *default_ctx = bdi_wb_ctx_arr[0];
> 
> on init so I'll leave that up to you.

Killing default_ctx completely seems like a better choice here. This
allows us to unconditionally use the for_each_bdi_wb_ctx loop, even for
the existing single writeback case,  by simply relying on bdi_wb_ctx[0].
It also eliminates the need for is_parallel, simplifying the design.

Thanks for the detailed review, Jan! Your suggestions look good to me
and help streamline the plumbing. I'll incorporate them in the next
version.

Thanks,
Anuj Gupta  





[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [NTFS 3]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [NTFS 3]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux