devicetree.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Peter Geis <pgwipeout@gmail.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: Shawn Lin <shawn.lin@rock-chips.com>,
	Simon Xue <xxm@rock-chips.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	linux-pci@vger.kernel.org,
	"open list:ARM/Rockchip SoC..."
	<linux-rockchip@lists.infradead.org>,
	devicetree@vger.kernel.org, Rob Herring <robh+dt@kernel.org>,
	Johan Jonker <jbx6244@gmail.com>,
	Heiko Stuebner <heiko@sntech.de>
Subject: Re: [PATCH v7] PCI: rockchip: Add Rockchip RK356X host controller driver
Date: Thu, 29 Apr 2021 08:16:48 -0400	[thread overview]
Message-ID: <CAMdYzYobyyROToMXng25NQEgSC7OpYLmo=iM9eDuPbSfkKuKgA@mail.gmail.com> (raw)
In-Reply-To: <20210429115451.GA2287423@infradead.org>

On Thu, Apr 29, 2021 at 7:55 AM Christoph Hellwig <hch@infradead.org> wrote:
>
> On Thu, Apr 29, 2021 at 07:26:40AM -0400, Peter Geis wrote:
> > Good Morning! I'm trying to implement the MSI workaround in as sane a
> > method as possible.
> > Do you have a better way to have the kernel only allocate memory in
> > the lower regions?
> > (Without disabling everything above 4G altogether)
> > This chip only supports a maximum of 8G.
> > (I know, why they did it while maintaining a 32bit bus is beyond me).
>
>
> Please use dma_alloc_coherent to allocate the memory, the dma_addr_t
> return value is the address to be fed to the hardware.  Before doing
> that set the desired mask.  DMA_BIT_MASK(33) would get your 8GB.

Thanks!
Unfortunately this isn't actually DMA allocation, they were using
GFP_DMA32 as a hack to allocate regular kernel memory in the 32 bit
range.
It's in drivers/irqchip/irq-gic-v3-its.c for reference.
The functions are simply kcalloc(), kzalloc(), kzalloc_node(),
alloc_pages_node(), and alloc_pages().
I'd prefer not to have to rewrite this driver's entire memory
allocation system for one errata.

I'm following the code path for dma_alloc_coherent to see if anything
sticks out to me though.

  parent reply	other threads:[~2021-04-29 12:17 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-14  7:03 [PATCH v7] PCI: rockchip: Add Rockchip RK356X host controller driver Simon Xue
2021-04-28 16:20 ` Peter Geis
     [not found]   ` <5af0f6f8-bc29-f50e-ca14-94049b7d17ed@rock-chips.com>
2021-04-29  1:42     ` Peter Geis
2021-04-29  6:41       ` Christoph Hellwig
     [not found]         ` <CAMdYzYrbfD-jFC3uPVPkhjXWZGeaUw5X=0w1Q=m-6OuhEGOKkQ@mail.gmail.com>
     [not found]           ` <20210429115451.GA2287423@infradead.org>
2021-04-29 12:16             ` Peter Geis [this message]
     [not found]       ` <3bf6f7ba-d5cb-4857-5fa8-bc4fa0b7e5f6@rock-chips.com>
2021-04-29 11:21         ` Peter Geis

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAMdYzYobyyROToMXng25NQEgSC7OpYLmo=iM9eDuPbSfkKuKgA@mail.gmail.com' \
    --to=pgwipeout@gmail.com \
    --cc=bhelgaas@google.com \
    --cc=devicetree@vger.kernel.org \
    --cc=hch@infradead.org \
    --cc=heiko@sntech.de \
    --cc=jbx6244@gmail.com \
    --cc=linux-pci@vger.kernel.org \
    --cc=linux-rockchip@lists.infradead.org \
    --cc=lorenzo.pieralisi@arm.com \
    --cc=robh+dt@kernel.org \
    --cc=shawn.lin@rock-chips.com \
    --cc=xxm@rock-chips.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).