hi Russell & Ulf
On 2/21/19 3:03 PM, Russell King - ARM Linux admin wrote:
On Thu, Feb 21, 2019 at 02:38:36PM +0100, Ludovic BARRE wrote:
hi Russell & Ulf
On 2/21/19 11:30 AM, Russell King - ARM Linux admin wrote:
On Thu, Feb 21, 2019 at 10:27:39AM +0000, Russell King - ARM Linux admin wrote:
On Thu, Feb 21, 2019 at 11:10:49AM +0100, Ludovic Barre wrote:
From: Ludovic Barre <ludovic.barre@xxxxxx>
This patch series introduces a bitmap of hardware quirks that require
some special action. This should reduce the number of boolean
into variant structure.
And adds quirk bit to define sdmmc specific transfer modes.
Please find some other way to deal with these differences. As far as
I'm concerned, introducing a quirk bitmask such as what was done in
sdhci is a complete disaster and leads to long-term maintanability
problems.
We already have a way to deal with variants in mmci.
... to finish what I was saying ...
and I think that:
if (variant->blksz_datactrl16)
datactrl = variant->datactrl_dpsm_enable | (data->blksz << 16);
else if (variant->blksz_datactrl4)
datactrl = variant->datactrl_dpsm_enable | (data->blksz << 4);
else
datactrl = variant->datactrl_dpsm_enable | blksz_bits << 4;
ought to become a variant function call which returns the appropriate
datactrl value. This would shrink the amount of variant testing in this
path, and also means that going forward we aren't facing an endlessly
increasing number of tests here.
For blksz_datactrl case:
We could create an inline function for datactrl16 and blksz_datactrl4
which returns the appropriate datactrl value (specific for ux500v2 and
qcom). This function could be register in mmci_host_ops structure.
Yes, this is what I'm proposing (except for the "inline" bit which
seems meaningless if it's called via the mmci_host_ops structure.)
I'm also proposing that it shouldn't just be the blksz that's
returned but anything that the variant needs to take account of,
including the stm transfer mode.
Ulf, are you alright with this callback approach (just to be sure that
every body is align, before send a patch)?
This mmci_host_ops callback could return datactrl config to
start data (defined by variant).
in mmci_start_data function we could call a common function which call a
hook if defined.
int mmci_dblksz(struct mmci_host *host)
As this is returning a register value, "u32" would be more appropriate
than "int".
{
if (host->ops && host->ops->dblksz)
return host->ops->dblk(host);
/* default data block size definition */
blksz_bits = ffs(data->blksz) - 1;
return blksz_bits << 4;
}
what do you think about it?
I don't see any reason not to make the call unconditional and have every
variant supply an appropriate function pointer. IMHO that keeps stuff
cleaner.
After, I'm afraid to multiply callback function in mmci_host_ops.
For stm32 transfer mode:
ditto, a callback function or I keep a boolean?
BR
Ludo
Ludovic Barre (2):
mmc: mmci: introduce a quirks property into variant struct
mmc: mmci: add quirk property to add stm32 transfer mode
drivers/mmc/host/mmci.c | 11 +++++++++++
drivers/mmc/host/mmci.h | 9 +++++++++
2 files changed, 20 insertions(+)
--
2.7.4
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
