[PATCH v8] phy: mediatek: tphy: add debugfs files
Vinod Koul
vkoul at kernel.org
Tue Feb 14 06:00:09 PST 2023
On 13-02-23, 08:22, Chunfeng Yun (云春峰) wrote:
> On Mon, 2023-02-13 at 16:21 +0800, chunfeng yun wrote:
> > On Fri, 2023-02-10 at 21:01 +0530, Vinod Koul wrote:
> > > On 10-02-23, 16:58, Chunfeng Yun wrote:
> > > > These debugfs files are mainly used to make eye diagram test
> > > > easier,
> > > > especially helpful to do HQA test for a new IC without efuse
> > > > enabled.
> > > >
> > > > Signed-off-by: Chunfeng Yun <chunfeng.yun at mediatek.com>
> > > > ---
> > > > v8: abandon patch to create phy debug root file
> > > > /sys/kernel/debug/phy/ suggested by Vinod
> > >
> > > Sorry that was not my suggestion
> >
> > Sorry, I misunderstood you means
> > >
> > > >
> > > > v6~v7: no changes
> > > >
> > > > v5: using common debugfs config CONFIG_DEBUG_FS
> > > >
> > > > v4: fix build warning of sometimes uninitialized variable
> > > >
> > > > v3: fix typo of "debugfs" suggested by AngeloGioacchino
> > > >
> > > > v2: add CONFIG_PHY_MTK_TPHY_DEBUGFS suggested by AngeloGioacchino
> > > > ---
> > > > drivers/phy/mediatek/phy-mtk-tphy.c | 405
> > > > +++++++++++++++++++++++++++-
> > > > 1 file changed, 404 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/phy/mediatek/phy-mtk-tphy.c
> > > > b/drivers/phy/mediatek/phy-mtk-tphy.c
> > > > index e906a82791bd..65a95c3726bf 100644
> > > > --- a/drivers/phy/mediatek/phy-mtk-tphy.c
> > > > +++ b/drivers/phy/mediatek/phy-mtk-tphy.c
> > > > @@ -7,6 +7,7 @@
> > > >
> > > > #include <dt-bindings/phy/phy.h>
> > > > #include <linux/clk.h>
> > > > +#include <linux/debugfs.h>
> > > > #include <linux/delay.h>
> > > > #include <linux/iopoll.h>
> > > > #include <linux/mfd/syscon.h>
> > > > @@ -264,6 +265,8 @@
> > > >
> > > > #define TPHY_CLKS_CNT 2
> > > >
> > > > +#define USER_BUF_LEN(count) min_t(size_t, 8, (count))
> > > > +
> > > > enum mtk_phy_version {
> > > > MTK_PHY_V1 = 1,
> > > > MTK_PHY_V2,
> > > > @@ -310,6 +313,7 @@ struct mtk_phy_instance {
> > > > struct clk_bulk_data clks[TPHY_CLKS_CNT];
> > > > u32 index;
> > > > u32 type;
> > > > + struct dentry *dbgfs;
> > > > struct regmap *type_sw;
> > > > u32 type_sw_reg;
> > > > u32 type_sw_index;
> > > > @@ -332,10 +336,391 @@ struct mtk_tphy {
> > > > const struct mtk_phy_pdata *pdata;
> > > > struct mtk_phy_instance **phys;
> > > > int nphys;
> > > > + struct dentry *dbgfs_root;
> > > > int src_ref_clk; /* MHZ, reference clock for slew rate
> > > > calibrate */
> > > > int src_coef; /* coefficient for slew rate calibrate */
> > > > };
> > >
> > > I asked the struct dentry *debugfs be part of struct phy. Debugfs
> > > would
> > > be created by core but rather than exported, it should be populated
> > > in
> > > phy and you can use it here in the driver
> > >
> > > >
> >
> > Ok, I'll do the following steps:
> >
> > 0. put a struct dentry *debugfs in struct phy;
> > 1. define a static struct dentry *phy_debugfs_root in phy-core.c,
> > then
> > create a root directory (/sys/kernel/debug/phy/) for phy
> > subsystem
> > in phy_core_init();
> > 2. create directory for each phy under the root directory
> > (/sys/kernel/debug/phy/);
> > this step has two ways:
> > A: create the directory @debugfs of struct phy in phy_create() by
> > default;
That sounds good to me
> > B: provide some api for user to create/destory the directory
> > @debugfs or others when need;
> > 3. add private debugfs files in @debugfs of struct phy;
> >
> > For the case in tphy driver, it provides some struct phy, the driver
> > may want to create debugfs files which not belongs to any struct phy
> > in
> > the root directory (/sys/kernel/debug/phy/), then need provide a api.
> >
> > Do you have any suggestions?
> > Which way do you prefer to use in step 2?
> >
> > Thanks a lot;
> >
> > > <skip>
> > > > +
> > > > static struct platform_driver mtk_tphy_driver = {
> > > > .probe = mtk_tphy_probe,
> > > > + .remove = mtk_tphy_remove,
> > > > .driver = {
> > > > .name = "mtk-tphy",
> > > > .of_match_table = mtk_tphy_id_table,
> > > > --
> > > > 2.18.0
> > >
> > >
--
~Vinod
More information about the Linux-mediatek
mailing list