[PATCH] swiotlb: set IO TLB segment size via cmdline

[Roman Skakun]

Hi Jan,

>>>  In order to be sure to catch all uses like this one (including ones
>>>  which make it upstream in parallel to yours), I think you will want
>>>  to rename the original IO_TLB_SEGSIZE to e.g. IO_TLB_DEFAULT_SEGSIZE.
>>
>> I don't understand your point. Can you clarify this?
>
> There's a concrete present example: I have a patch pending adding
> another use of IO_TLB_SEGSIZE. This use would need to be replaced
> like you do here in several places. The need for the additional
> replacement would be quite obvious (from a build failure) if you
> renamed the manifest constant. Without renaming, it'll take
> someone running into an issue on a live system, which I consider
> far worse. This is because a simple re-basing of one of the
> patches on top of the other will not point out the need for the
> extra replacement, nor would a test build (with both patches in
> place).

It's reasonable.
I will change the original IO_TLB_SEGSIZE to IO_TLB_DEFAULT_SEGSIZE in the
next patch series.

Thanks.

ср, 15 сент. 2021 г. в 16:50, Jan Beulich <jbeulich at suse.com>:
>
> On 15.09.2021 15:37, Roman Skakun wrote:
> >>> From: Roman Skakun <roman_skakun at epam.com>
> >>>
> >>> It is possible when default IO TLB size is not
> >>> enough to fit a long buffers as described here [1].
> >>>
> >>> This patch makes a way to set this parameter
> >>> using cmdline instead of recompiling a kernel.
> >>>
> >>> [1] https://www.xilinx.com/support/answers/72694.html
> >>
> >>  I'm not convinced the swiotlb use describe there falls under "intended
> >>  use" - mapping a 1280x720 framebuffer in a single chunk?
> >
> > I had the same issue while mapping DMA chuck ~4MB for gem fb when
> > using xen vdispl.
> > I got the next log:
> > [ 142.030421] rcar-fcp fea2f000.fcp: swiotlb buffer is full (sz:
> > 3686400 bytes), total 32768 (slots), used 32 (slots)
> >
> > It happened when I tried to map bounce buffer, which has a large size.
> > The default size if 128(IO_TLB_SEGSIZE) * 2048(IO_TLB_SHIFT) = 262144
> > bytes, but we requested 3686400 bytes.
> > When I change IO_TLB_SEGSIZE to 2048. (2048(IO_TLB_SEGSIZE)  *
> > 2048(IO_TLB_SHIFT) = 4194304bytes).
> > It makes possible to retrieve a bounce buffer for requested size.
> > After changing this value, the problem is gone.
>
> But the question remains: Why does the framebuffer need to be mapped
> in a single giant chunk?
>
> >>  In order to be sure to catch all uses like this one (including ones
> >>  which make it upstream in parallel to yours), I think you will want
> >>  to rename the original IO_TLB_SEGSIZE to e.g. IO_TLB_DEFAULT_SEGSIZE.
> >
> > I don't understand your point. Can you clarify this?
>
> There's a concrete present example: I have a patch pending adding
> another use of IO_TLB_SEGSIZE. This use would need to be replaced
> like you do here in several places. The need for the additional
> replacement would be quite obvious (from a build failure) if you
> renamed the manifest constant. Without renaming, it'll take
> someone running into an issue on a live system, which I consider
> far worse. This is because a simple re-basing of one of the
> patches on top of the other will not point out the need for the
> extra replacement, nor would a test build (with both patches in
> place).
>
> Jan
>


-- 
Best Regards, Roman.