> -----Original Message-----
> From: Shreyansh Jain [mailto:[email protected]]
> Sent: Wednesday, April 4, 2018 11:14 AM
> To: Xu, Rosen <[email protected]>; [email protected]
> Cc: Doherty, Declan <[email protected]>; Richardson, Bruce
> <[email protected]>; Yigit, Ferruh <[email protected]>;
> Ananyev, Konstantin <[email protected]>; Zhang, Tianfei
> <[email protected]>; Wu, Hao <[email protected]>;
> [email protected]
> Subject: RE: [PATCH v5 0/3] Introduce Intel FPGA BUS
> 
> Hello Rosen,
> 
> > -----Original Message-----
> > From: Rosen Xu [mailto:[email protected]]
> > Sent: Wednesday, April 4, 2018 12:21 PM
> > To: [email protected]
> > Cc: [email protected]; [email protected]; Shreyansh
> > Jain <[email protected]>; [email protected];
> > [email protected]; [email protected];
> > [email protected]; [email protected]
> > Subject: [PATCH v5 0/3] Introduce Intel FPGA BUS
> >
> > Intel FPGA BUS in DPDK
> > -------------------------
> >
> > This patch set introduces Intel FPGA BUS support in DPDK.
> >
> > v5 updates:
> > ===========
> >  - Fixed SHARED LIB Build issue
> >  - Changed command name to IFPGA Rawdev name,
> >    so remove pci library datastruct and function.
> >  - Fixed PATCH v2/v3/v4 comments
> >
> 
> [...]
> 
> Primary problems I see with your patches:
> 1. They are not split enough. Still the patch 2/3 is dependent on 3/3.
> That mean, it would break the compilation. There is no simpler way to
> solve this except breaking the patch into multiple patches and slowly
> introducing each function/feature.
>  (One obvious way would be to have 3/3 as 2/3 and vice-versa - Not sure
> what that blocks).
> 
> 2. Documentation - there is none right now. Being a special use case for
> PCI, I think a lot of people would benefit if you can explain the comments
> about why iFPGA bus is required through documentation.
> 
> 3. Meson as requested by Bruce. Problem you will face is that rawdev
> doesn't yet have meson enabled. I will work on that. If you can still
> rework your patches for (1)+(2), I think meson enable over rawdev would be
> trivial.

I just spotted this and I've sent a patch for rawdev. It was pretty trivial. :-)
Please review and ack if you have the chance. The skeleton rawdev however, I 
haven't
done, so feel free to patch in that.

/Bruce

Reply via email to