[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PULL 21/29] xen/pt: Sync up the dev.config and data va
From: |
Konrad Rzeszutek Wilk |
Subject: |
Re: [Qemu-devel] [PULL 21/29] xen/pt: Sync up the dev.config and data values. |
Date: |
Tue, 15 Sep 2015 09:55:28 -0400 |
User-agent: |
Mutt/1.5.23 (2014-03-12) |
On Tue, Sep 15, 2015 at 02:28:51PM +0100, Stefano Stabellini wrote:
> On Tue, 15 Sep 2015, Konrad Rzeszutek Wilk wrote:
> > On Tue, Sep 15, 2015 at 11:07:02AM +0100, Stefano Stabellini wrote:
> > > CC Konrad
> > >
> > > On Mon, 14 Sep 2015, Paolo Bonzini wrote:
> > > > On 10/09/2015 19:15, Stefano Stabellini wrote:
> > > > > +
> > > > > + switch (reg->size) {
> > > > > + case 1: rc = xen_host_pci_get_byte(&s->real_device, offset,
> > > > > (uint8_t *)&val);
> > > >
> > > > A bit ugly, and it relies on the host being little endian.
> > > >
> > > > > + break;
> > > > > + case 2: rc = xen_host_pci_get_word(&s->real_device, offset,
> > > > > (uint16_t *)&val);
> > > >
> > > > Same here.
> > >
> > > cpu_to_le32?
> > >
> > > But in practice, Xen being little endian only, I doubt that
> > > xen_pt_config_init.c
> > > would actually work on be.
> > >
> > >
> > > > > + break;
> > > > > + case 4: rc = xen_host_pci_get_long(&s->real_device, offset,
> > > > > &val);
> > > > > + break;
> > > > > + default: assert(1);
> > > >
> > > > This should be assert(0) or, better, abort().
> >
> > OK. Stefano, do you want me to:
> >
> > 1). Rebase the patches on top of your tag?
>
> Patches are already upstream, so no worries about rebasing.
>
>
> > 2). Send an follow up patch to change this to abort()? (and wherever else
> > I used
> > assert(..)?
>
> Yes, that would be good.
>
>
> > 3). Wait till Paolo is done going through the patchset and then revisit 1)
> > or 2)?
>
> I don't know if Paolo has any more comments.
>
> Regarding his previous comment on little-endian, as I wrote I am not
> sure if sprinkling around some cpu_to_le32 would actually make
> hw/xen/xen_pt_config_init.c much better. I'll leave that to you.
I got some other outstanding things I need to get done so will be
as most lazy as possible in regards to this and just send you a patch :-)
- [Qemu-devel] [PULL 13/29] xen/HVM: atomically access pointers in bufioreq handling, (continued)
- [Qemu-devel] [PULL 13/29] xen/HVM: atomically access pointers in bufioreq handling, Stefano Stabellini, 2015/09/10
- [Qemu-devel] [PULL 12/29] xen-hvm: When using xc_domain_add_to_physmap also include errno when reporting, Stefano Stabellini, 2015/09/10
- [Qemu-devel] [PULL 22/29] xen/pt: Check if reg->init function sets the 'data' past the reg->size, Stefano Stabellini, 2015/09/10
- [Qemu-devel] [PULL 28/29] xen/pt: Check for return values for xen_host_pci_[get|set] in init, Stefano Stabellini, 2015/09/10
- [Qemu-devel] [PULL 21/29] xen/pt: Sync up the dev.config and data values., Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 14/29] xen/pt: Update comments with proper function name., Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 25/29] xen/pt: Log xen_host_pci_get/set errors in MSI code., Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 24/29] xen/pt: Log xen_host_pci_get in two init functions, Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 18/29] xen/pt/msi: Add the register value when printing logging and error messages, Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 23/29] xen/pt: Remove XenPTReg->data field., Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 11/29] xen, gfx passthrough: add opregion mapping, Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 15/29] xen/pt: Make xen_pt_msi_set_enable static, Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 16/29] xen/pt: xen_host_pci_config_read returns -errno, not -1 on failure, Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 27/29] xen/pt: Move bulk of xen_pt_unregister_device in its own routine., Stefano Stabellini, 2015/09/10
[Qemu-devel] [PULL 20/29] xen/pt: Use xen_host_pci_get_[byte|word] instead of dev.config, Stefano Stabellini, 2015/09/10