[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-id: <004c01cf4e2d$1d84a1c0$588de540$@samsung.com>
Date: Wed, 02 Apr 2014 13:36:29 +0900
From: Jungseok Lee <jays.lee@...sung.com>
To: 'Konrad Rzeszutek Wilk' <konrad.wilk@...cle.com>
Cc: linux-kernel@...r.kernel.org, sungjinn.chung@...sung.com,
'linux-samsung-soc' <linux-samsung-soc@...r.kernel.org>
Subject: Re: [PATCH] swiotlb: Correct kernel command line handler
On Tuesday, April 01, 2014 9:46 PM, Konrad Rzeszutek Wilk wrote:
> On Tue, Apr 01, 2014 at 10:51:58AM +0900, Jungseok Lee wrote:
> > This patch corrects kernel command line handler when io_tlb_nslabs is
> > set to 0.
> >
> > A current implementation allocates default size memory (64MB) when
> > 0 is given to io_tlb_nslabs ("swiotlb=0") by kernel command line.
> > In other words, memory is allocated unintentionally.
>
> What are the other IOMMUs that end up taking over the DMA ops?
I think that no other IOMMUs take over it on ARM64 system, except Xen.
> Or do you end up with no DMA ops?
I focus on ARM64 system which uses swiotlb as default DMA ops.
If some ARM64 platforms would like to use memory only from
__get_free_pages in swiotlb_alloc_coherent function, an
interface would be needed to configure reserved memory size.
Thus, I add it to kernel command line handler.
Best Regards
Jungseok Lee
> >
> > Signed-off-by: Jungseok Lee <jays.lee@...sung.com>
> > Signed-off-by: Sungjinn Chung <sungjinn.chung@...sung.com>
> > ---
> > lib/swiotlb.c | 8 ++++++--
> > 1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/lib/swiotlb.c b/lib/swiotlb.c index b604b83..9fe9f71
> > 100644
> > --- a/lib/swiotlb.c
> > +++ b/lib/swiotlb.c
> > @@ -55,6 +55,8 @@
> >
> > int swiotlb_force;
> >
> > +static int swiotlb_no_use;
> > +
> > /*
> > * Used to do a quick range check in swiotlb_tbl_unmap_single and
> > * swiotlb_tbl_sync_single_*, to see if the memory was in fact
> > allocated by this @@ -102,6 +104,8 @@ setup_io_tlb_npages(char *str)
> > io_tlb_nslabs = simple_strtoul(str, &str, 0);
> > /* avoid tail segment of size < IO_TLB_SEGSIZE */
> > io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
> > + if (!io_tlb_nslabs)
> > + swiotlb_no_use = 1;
> > }
> > if (*str == ',')
> > ++str;
> > @@ -212,7 +216,7 @@ swiotlb_init(int verbose)
> > unsigned char *vstart;
> > unsigned long bytes;
> >
> > - if (!io_tlb_nslabs) {
> > + if (!swiotlb_no_use && !io_tlb_nslabs) {
> > io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
> > io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
> > }
> > @@ -244,7 +248,7 @@ swiotlb_late_init_with_default_size(size_t default_size)
> > unsigned int order;
> > int rc = 0;
> >
> > - if (!io_tlb_nslabs) {
> > + if (!swiotlb_no_use && !io_tlb_nslabs) {
> > io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
> > io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
> > }
> > --
> > 1.7.10.4
> >
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in the body of a message
> to majordomo@...r.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists