[U-Boot] [Patch V2 1/4] pci/layerscape: add support for LS1043A PCIe LUT register access
Simon Glass
sjg at chromium.org
Fri Nov 6 01:17:41 CET 2015
Hi,
On 4 November 2015 at 22:00, Bin Meng <bmeng.cn at gmail.com> wrote:
> +Simon,
>
> Hi,
>
> On Thu, Nov 5, 2015 at 7:45 AM, Scott Wood <scottwood at freescale.com> wrote:
>> On Mon, 2015-11-02 at 19:15 +0800, Gong Qianyu wrote:
>>> diff --git a/drivers/pci/pcie_layerscape.c b/drivers/pci/pcie_layerscape.c
>>> index 4cee038..8471678 100644
>>> --- a/drivers/pci/pcie_layerscape.c
>>> +++ b/drivers/pci/pcie_layerscape.c
>>> @@ -13,6 +13,7 @@
>>> #include <malloc.h>
>>> #ifdef CONFIG_FSL_LAYERSCAPE
>>> #include <asm/arch/fdt.h>
>>> +#include <asm/arch/soc.h>
>>> #endif
>>>
>>> #ifndef CONFIG_SYS_PCI_MEMORY_BUS
>>> @@ -57,11 +58,6 @@
>>> #define PCIE_ATU_FUNC(x) (((x) & 0x7) << 16)
>>> #define PCIE_ATU_UPPER_TARGET 0x91C
>>>
>>> -/* LUT registers */
>>> -#define PCIE_LUT_BASE 0x80000
>>> -#define PCIE_LUT_LCTRL0 0x7F8
>>> -#define PCIE_LUT_DBG 0x7FC
>>> -
>>> #define PCIE_DBI_RO_WR_EN 0x8bc
>>>
>>> #define PCIE_LINK_CAP 0x7c
>>> @@ -157,12 +153,12 @@ static int ls_pcie_link_state(struct ls_pcie *pcie)
>>>
>>> return 1;
>>> }
>>> -#else
>>> +#elif defined(CONFIG_FSL_LAYERSCAPE)
>>> static int ls_pcie_link_state(struct ls_pcie *pcie)
>>> {
>>> u32 state;
>>>
>>> - state = readl(pcie->dbi + PCIE_LUT_BASE + PCIE_LUT_DBG) &
>>> + state = pex_lut_in32(pcie->dbi + PCIE_LUT_BASE + PCIE_LUT_DBG) &
>>> LTSSM_STATE_MASK;
>>> if (state < LTSSM_PCIE_L0) {
>>> debug("....PCIe link error. LTSSM=0x%02x.\n", state);
>>> @@ -466,16 +462,20 @@ static void ls_pcie_setup_ep(struct ls_pcie *pcie,
>>> struct ls_pcie_info *info)
>>>
>>> for (pf = 0; pf < PCIE_PF_NUM; pf++) {
>>> for (vf = 0; vf <= PCIE_VF_NUM; vf++) {
>>> +#ifdef CONFIG_FSL_LAYERSCAPE
>>> writel(PCIE_LCTRL0_VAL(pf, vf),
>>> pcie->dbi + PCIE_LUT_BASE +
>>> PCIE_LUT_LCTRL0);
>>> +#endif
>>
>> It looks really weird to have "#ifdef CONFIG_FSL_LAYERSCAPE" inside a file
>> whose name suggests is layerscape-specific. What chips use this file that
>> are not CONFIG_FSL_LAYERSCAPE?
>>
>> -Scott
>
> Can we start converting the driver to use DM PCI? We should not
> continue using legacy PCI codes since right now DM PCI is in a good
> shape. These #ifdef in the driver should really go into device tree as
> needed.
Agreed. In fact I have a series to add a few improvements, will try to
get that out in the next week.
Regards,
Simon
More information about the U-Boot
mailing list