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

Re: [Minios-devel] [UNIKRAFT PATCHv4 17/43] plat/include: Define macros for Arm64 to access registers

Hey Wei, hey Julien,

On 12.07.2018 13:17, Julien Grall wrote:
On 12/07/18 03:41, Wei Chen wrote:
Hi Julien,

Hi Wei,

-----Original Message-----
From: Julien Grall <julien.grall@xxxxxxx>
Sent: 2018年7月11日 21:16
To: Wei Chen <Wei.Chen@xxxxxxx>; minios-devel@xxxxxxxxxxxxxxxxxxxx;
Cc: Kaly Xin <Kaly.Xin@xxxxxxx>; nd <nd@xxxxxxx>
Subject: Re: [Minios-devel] [UNIKRAFT PATCHv4 17/43] plat/include: Define
macros for Arm64 to access registers

On 11/07/18 07:28, Wei Chen wrote:
Hi Julien,

Hi Wei,

Yes, the only device we're using on Arm is PL011, so it's the only one
user of this today.

I meant for outb/inb. Not REG_READ*() helpers. So do you have any users
of outb/inb today for Arm?

I also prefer the IOREG name for the macros.

Yes, the common code like PCI bus will use it. Even I know the PCI bus
could not work properly on Arm. But I still want the PCI code can be
compiled on Arm.

What's the rationale of compiling PCI code if it is wrong? Mostly like PCI code will use through 0xcf8/0xcfc that does not have equivalent on Arm. Instead, we will need to add support for ECAM.

Anyway that's the maintainers call here. However, you should at least not implement outb/inb the way you did. This is a call to misuse them. Instead you should implement using BUG() or ASSERT(). So any use of those outb will be caught later on.

We should ignore outb/inb/etc. for now. The PCI code is not ready for ARM yet and I would suggest to make it ready with another patch series. With this one, we should in fact introduce a proper API for accessing PCI device registers. inb/outb/... should be replaced.




Minios-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.