mmc: Set PROBE_PREFER_ASYNCHRONOUS for all host drivers
Subsystems (46)
cover mmc: Set PROBE_PREFER_ASYNCHRONOUS for all host drivers
As per discussion [1], it seems like it should be quite safe to turn on PROBE_PREFER_ASYNCHRONOUS for all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe").
I've broken this series into chunks based on LTS kernel releases to attempt to make it easier if someone wanted to cherry-pick it to an older kernel. While these cherry-picks won't be conflict free, there should only be trivial context conflicts and no problems with drivers that are totally missing. This is a bit of a compromise between a 1-big patch and a many-part patch series.
I have only tested this on a rk3399-kevin (sdhci-of-arasan) and a rk3288-veyron (dw_mmc-rockchip) device and only lightly. If this patch causes anyone problems those drivers should be marked with PROBE_FORCE_SYNCHRONOUS, debugged, and then go back to prefer asynchronous. Any problems are likely just a hidden bug that has been exposed by this change.
NOTE: in theory, it'd be nice if there was a KConfig option that we could flip that would turn on async probe everywhere (except for those that opt out by adding PROBE_FORCE_SYNCHRONOUS). My hope is that by adding this flag in more places it will become clear that this flag is working reliably and easier to move over when we're ready.
While coccinelle is too difficult for my feeble brain, I managed to whip up a pretty terrible python script to help with this. For your edification:
import os
import sys
import re
for filename in os.listdir("."):
found_plat = False
found_driver = False
output = []
for line in open(filename, "r").readlines():
output.append(line)
if "struct platform_driver" in line:
found_plat = True
if found_plat and re.search(r"\t\.driver\s*=\s*{", line):
found_driver = True
found_plat = False
mo = re.search(r"(\s*)\.name(\s*)=", line)
if found_driver and mo:
if mo.group(2) == " ":
space = " "
elif mo.group(2) == "\t":
# Best we can do
space = " "
elif mo.group(2).startswith("\t"):
# Guess that removing one tab is right space = mo.group(2)[1:]
else:
# Guess it's all spaces
space = mo.group(2)[7:] + " "
output.append("%s.probe_type%s= PROBE_PREFER_ASYNCHRONOUS,\n" % (mo.group(1), space)) found_driver = False
open(filename, "w").write("".join(output))
[1] https://lore.kernel.org/r/CAPDyKFq31bucJhP9hp1HSqh-qM2uNGHgDoyQpmbJf00nEf_T4Q at mail.gmail.com/
Douglas Anderson (6):
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.4
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.9
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.14
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.19
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v5.4
mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that are newer than 5.4
drivers/mmc/host/alcor.c | 1 + drivers/mmc/host/android-goldfish.c | 1 + drivers/mmc/host/atmel-mci.c | 1 + drivers/mmc/host/au1xmmc.c | 1 + drivers/mmc/host/bcm2835.c | 1 + drivers/mmc/host/cavium-octeon.c | 1 + drivers/mmc/host/davinci_mmc.c | 1 + drivers/mmc/host/dw_mmc-bluefield.c | 1 + drivers/mmc/host/dw_mmc-exynos.c | 1 + drivers/mmc/host/dw_mmc-hi3798cv200.c | 1 + drivers/mmc/host/dw_mmc-k3.c | 1 + drivers/mmc/host/dw_mmc-pltfm.c | 1 + drivers/mmc/host/dw_mmc-rockchip.c | 1 + drivers/mmc/host/dw_mmc-zx.c | 1 + drivers/mmc/host/jz4740_mmc.c | 1 + drivers/mmc/host/meson-gx-mmc.c | 1 + drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/meson-mx-sdio.c | 1 + drivers/mmc/host/moxart-mmc.c | 1 + drivers/mmc/host/mtk-sd.c | 1 + drivers/mmc/host/mvsdio.c | 1 + drivers/mmc/host/mxcmmc.c | 1 + drivers/mmc/host/mxs-mmc.c | 1 + drivers/mmc/host/omap.c | 1 + drivers/mmc/host/omap_hsmmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/pxamci.c | 1 + drivers/mmc/host/renesas_sdhi_internal_dmac.c | 1 + drivers/mmc/host/renesas_sdhi_sys_dmac.c | 1 + drivers/mmc/host/rtsx_pci_sdmmc.c | 1 + drivers/mmc/host/rtsx_usb_sdmmc.c | 1 + drivers/mmc/host/s3cmci.c | 1 + drivers/mmc/host/sdhci-acpi.c | 1 + drivers/mmc/host/sdhci-bcm-kona.c | 1 + drivers/mmc/host/sdhci-brcmstb.c | 1 + drivers/mmc/host/sdhci-cadence.c | 1 + drivers/mmc/host/sdhci-cns3xxx.c | 1 + drivers/mmc/host/sdhci-dove.c | 1 + drivers/mmc/host/sdhci-esdhc-imx.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-iproc.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-arasan.c | 1 + drivers/mmc/host/sdhci-of-aspeed.c | 2 ++ drivers/mmc/host/sdhci-of-at91.c | 1 + drivers/mmc/host/sdhci-of-dwcmshc.c | 1 + drivers/mmc/host/sdhci-of-esdhc.c | 1 + drivers/mmc/host/sdhci-of-hlwd.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + drivers/mmc/host/sdhci-omap.c | 1 + drivers/mmc/host/sdhci-pic32.c | 1 + drivers/mmc/host/sdhci-pxav2.c | 1 + drivers/mmc/host/sdhci-pxav3.c | 1 + drivers/mmc/host/sdhci-s3c.c | 1 + drivers/mmc/host/sdhci-sirf.c | 1 + drivers/mmc/host/sdhci-spear.c | 1 + drivers/mmc/host/sdhci-sprd.c | 1 + drivers/mmc/host/sdhci-st.c | 1 + drivers/mmc/host/sdhci-tegra.c | 1 + drivers/mmc/host/sdhci-xenon.c | 1 + drivers/mmc/host/sdhci_am654.c | 1 + drivers/mmc/host/sdhci_f_sdh30.c | 1 + drivers/mmc/host/sh_mmcif.c | 1 + drivers/mmc/host/sunxi-mmc.c | 1 + drivers/mmc/host/tmio_mmc.c | 1 + drivers/mmc/host/uniphier-sd.c | 1 + drivers/mmc/host/usdhi6rol0.c | 1 + drivers/mmc/host/wbsd.c | 1 + drivers/mmc/host/wmt-sdmmc.c | 1 + 69 files changed, 70 insertions(+)
Signature · 1 line
2.28.0.526.ge36021eeef-goog
PATCH 1/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.4
This is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.4 timeframe.
---
drivers/mmc/host/android-goldfish.c | 1 + drivers/mmc/host/atmel-mci.c | 1 + drivers/mmc/host/au1xmmc.c | 1 + drivers/mmc/host/davinci_mmc.c | 1 + drivers/mmc/host/dw_mmc-exynos.c | 1 + drivers/mmc/host/dw_mmc-k3.c | 1 + drivers/mmc/host/dw_mmc-pltfm.c | 1 + drivers/mmc/host/dw_mmc-rockchip.c | 1 + drivers/mmc/host/jz4740_mmc.c | 1 + drivers/mmc/host/moxart-mmc.c | 1 + drivers/mmc/host/mtk-sd.c | 1 + drivers/mmc/host/mvsdio.c | 1 + drivers/mmc/host/mxcmmc.c | 1 + drivers/mmc/host/mxs-mmc.c | 1 + drivers/mmc/host/omap.c | 1 + drivers/mmc/host/omap_hsmmc.c | 1 + drivers/mmc/host/pxamci.c | 1 + drivers/mmc/host/rtsx_pci_sdmmc.c | 1 + drivers/mmc/host/rtsx_usb_sdmmc.c | 1 + drivers/mmc/host/s3cmci.c | 1 + drivers/mmc/host/sdhci-acpi.c | 1 + drivers/mmc/host/sdhci-bcm-kona.c | 1 + drivers/mmc/host/sdhci-cns3xxx.c | 1 + drivers/mmc/host/sdhci-dove.c | 1 + drivers/mmc/host/sdhci-esdhc-imx.c | 1 + drivers/mmc/host/sdhci-iproc.c | 1 + drivers/mmc/host/sdhci-of-arasan.c | 1 + drivers/mmc/host/sdhci-of-at91.c | 1 + drivers/mmc/host/sdhci-of-esdhc.c | 1 + drivers/mmc/host/sdhci-of-hlwd.c | 1 + drivers/mmc/host/sdhci-pxav2.c | 1 + drivers/mmc/host/sdhci-pxav3.c | 1 + drivers/mmc/host/sdhci-s3c.c | 1 + drivers/mmc/host/sdhci-sirf.c | 1 + drivers/mmc/host/sdhci-spear.c | 1 + drivers/mmc/host/sdhci-st.c | 1 + drivers/mmc/host/sdhci-tegra.c | 1 + drivers/mmc/host/sdhci_f_sdh30.c | 1 + drivers/mmc/host/sh_mmcif.c | 1 + drivers/mmc/host/sunxi-mmc.c | 1 + drivers/mmc/host/tmio_mmc.c | 1 + drivers/mmc/host/usdhi6rol0.c | 1 + drivers/mmc/host/wbsd.c | 1 + drivers/mmc/host/wmt-sdmmc.c | 1 + 44 files changed, 44 insertions(+)
drivers/mmc/host/android-goldfish.c +1 -0
@@ -537,6 +537,7 @@ static struct platform_driver goldfish_mmc_driver = { .remove = goldfish_mmc_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, }, };
drivers/mmc/host/atmel-mci.c +1 -0
@@ -2668,6 +2668,7 @@ static struct platform_driver atmci_driver = { .remove = atmci_remove, .driver = { .name = "atmel_mci", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(atmci_dt_ids), .pm = &atmci_dev_pm_ops, },
drivers/mmc/host/au1xmmc.c +1 -0
@@ -1189,6 +1189,7 @@ static struct platform_driver au1xmmc_driver = { .resume = au1xmmc_resume, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, }, };
drivers/mmc/host/davinci_mmc.c +1 -0
@@ -1395,6 +1395,7 @@ static const struct dev_pm_ops davinci_mmcsd_pm = {static struct platform_driver davinci_mmcsd_driver = { .driver = { .name ="davinci_mmc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = davinci_mmcsd_pm_ops, .of_match_table = davinci_mmc_dt_ids, },
drivers/mmc/host/dw_mmc-exynos.c +1 -0
@@ -592,6 +592,7 @@ static struct platform_driver dw_mci_exynos_pltfm_driver = { .remove = dw_mci_exynos_remove, .driver = { .name = "dwmmc_exynos", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_exynos_match, .pm = &dw_mci_exynos_pmops, },
drivers/mmc/host/dw_mmc-k3.c +1 -0
@@ -473,6 +473,7 @@ static struct platform_driver dw_mci_k3_pltfm_driver = { .remove = dw_mci_pltfm_remove, .driver = { .name = "dwmmc_k3", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_k3_match, .pm = &dw_mci_k3_dev_pm_ops, },
drivers/mmc/host/dw_mmc-pltfm.c +1 -0
@@ -98,6 +98,7 @@ static struct platform_driver dw_mci_pltfm_driver = { .remove = dw_mci_pltfm_remove, .driver = { .name = "dw_mmc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_pltfm_match, .pm = &dw_mci_pltfm_pmops, },
drivers/mmc/host/dw_mmc-rockchip.c +1 -0
@@ -383,6 +383,7 @@ static struct platform_driver dw_mci_rockchip_pltfm_driver = { .remove = dw_mci_rockchip_remove, .driver = { .name = "dwmmc_rockchip", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_rockchip_match, .pm = &dw_mci_rockchip_dev_pm_ops, },
drivers/mmc/host/jz4740_mmc.c +1 -0
@@ -1125,6 +1125,7 @@ static struct platform_driver jz4740_mmc_driver = { .remove = jz4740_mmc_remove, .driver = { .name = "jz4740-mmc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(jz4740_mmc_of_match), .pm = pm_ptr(&jz4740_mmc_pm_ops), },
drivers/mmc/host/moxart-mmc.c +1 -0
@@ -717,6 +717,7 @@ static struct platform_driver moxart_mmc_driver = { .remove = moxart_remove, .driver = { .name = "mmc-moxart", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = moxart_mmc_match, }, };
drivers/mmc/host/mtk-sd.c +1 -0
@@ -2689,6 +2689,7 @@ static struct platform_driver mt_msdc_driver = { .remove = msdc_drv_remove, .driver = { .name = "mtk-msdc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = msdc_of_ids, .pm = &msdc_dev_pm_ops, },
drivers/mmc/host/mvsdio.c +1 -0
@@ -824,6 +824,7 @@ static struct platform_driver mvsd_driver = { .remove = mvsd_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = mvsdio_dt_ids, }, };
drivers/mmc/host/mxcmmc.c +1 -0
@@ -1244,6 +1244,7 @@ static struct platform_driver mxcmci_driver = { .id_table = mxcmci_devtype, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &mxcmci_pm_ops, .of_match_table = mxcmci_of_match, }
drivers/mmc/host/mxs-mmc.c +1 -0
@@ -726,6 +726,7 @@ static struct platform_driver mxs_mmc_driver = { .id_table = mxs_ssp_ids, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &mxs_mmc_pm_ops, .of_match_table = mxs_mmc_dt_ids, },
drivers/mmc/host/omap.c +1 -0
@@ -1504,6 +1504,7 @@ static struct platform_driver mmc_omap_driver = { .remove = mmc_omap_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(mmc_omap_match), }, };
drivers/mmc/host/omap_hsmmc.c +1 -0
@@ -2171,6 +2171,7 @@ static struct platform_driver omap_hsmmc_driver = { .remove = omap_hsmmc_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &omap_hsmmc_dev_pm_ops, .of_match_table = of_match_ptr(omap_mmc_of_match), },
drivers/mmc/host/pxamci.c +1 -0
@@ -811,6 +811,7 @@ static struct platform_driver pxamci_driver = { .remove = pxamci_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(pxa_mmc_dt_ids), }, };
drivers/mmc/host/rtsx_pci_sdmmc.c +1 -0
@@ -1471,6 +1471,7 @@ static struct platform_driver rtsx_pci_sdmmc_driver = { .id_table = rtsx_pci_sdmmc_ids, .driver = { .name = DRV_NAME_RTSX_PCI_SDMMC, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, }, };module_platform_driver (rtsx_pci_sdmmc_driver);
drivers/mmc/host/rtsx_usb_sdmmc.c +1 -0
@@ -1458,6 +1458,7 @@ static struct platform_driver rtsx_usb_sdmmc_driver = { .id_table = rtsx_usb_sdmmc_ids, .driver = { .name = "rtsx_usb_sdmmc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &rtsx_usb_sdmmc_dev_pm_ops, }, };
drivers/mmc/host/s3cmci.c +1 -0
@@ -1804,6 +1804,7 @@ MODULE_DEVICE_TABLE(platform, s3cmci_driver_ids);static struct platform_driver s3cmci_driver = { .driver = { .name ="s3c-sdi" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = s3cmci_dt_match, }, .id_table = s3cmci_driver_ids,
drivers/mmc/host/sdhci-acpi.c +1 -0
@@ -1027,6 +1027,7 @@ static const struct dev_pm_ops sdhci_acpi_pm_ops = {static struct platform_driver sdhci_acpi_driver = { .driver = { .name ="sdhci-acpi" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .acpi_match_table = sdhci_acpi_ids, .pm = &sdhci_acpi_pm_ops, },
drivers/mmc/host/sdhci-bcm-kona.c +1 -0
@@ -324,6 +324,7 @@ static int sdhci_bcm_kona_probe(struct platform_device *pdev)static struct platform_driver sdhci_bcm_kona_driver = { .driver = { .name ="sdhci-kona" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pltfm_pmops, .of_match_table = sdhci_bcm_kona_of_match, },
drivers/mmc/host/sdhci-cns3xxx.c +1 -0
@@ -98,6 +98,7 @@ static int sdhci_cns3xxx_probe(struct platform_device *pdev)static struct platform_driver sdhci_cns3xxx_driver = { .driver = { .name ="sdhci-cns3xxx" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pltfm_pmops, }, .probe = sdhci_cns3xxx_probe,
drivers/mmc/host/sdhci-dove.c +1 -0
@@ -105,6 +105,7 @@ MODULE_DEVICE_TABLE(of, sdhci_dove_of_match_table);static struct platform_driver sdhci_dove_driver = { .driver = { .name ="sdhci-dove" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pltfm_pmops, .of_match_table = sdhci_dove_of_match_table, },
drivers/mmc/host/sdhci-esdhc-imx.c +1 -0
@@ -1925,6 +1925,7 @@ static const struct dev_pm_ops sdhci_esdhc_pmops = {static struct platform_driver sdhci_esdhc_imx_driver = { .driver = { .name ="sdhci-esdhc-imx" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = imx_esdhc_dt_ids, .pm = &sdhci_esdhc_pmops, },
drivers/mmc/host/sdhci-iproc.c +1 -0
@@ -369,6 +369,7 @@ static int sdhci_iproc_probe(struct platform_device *pdev)static struct platform_driver sdhci_iproc_driver = { .driver = { .name ="sdhci-iproc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_iproc_of_match, .acpi_match_table =ACPI_PTR (sdhci_iproc_acpi_ids), .pm = &sdhci_pltfm_pmops,
drivers/mmc/host/sdhci-of-arasan.c +1 -0
@@ -1693,6 +1693,7 @@ static int sdhci_arasan_remove(struct platform_device *pdev)static struct platform_driver sdhci_arasan_driver = { .driver = { .name ="sdhci-arasan" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_arasan_of_match, .pm = &sdhci_arasan_dev_pm_ops, },
drivers/mmc/host/sdhci-of-at91.c +1 -0
@@ -465,6 +465,7 @@ static int sdhci_at91_remove(struct platform_device *pdev)static struct platform_driver sdhci_at91_driver = { .driver = { .name ="sdhci-at91" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_at91_dt_match, .pm = &sdhci_at91_dev_pm_ops, },
drivers/mmc/host/sdhci-of-esdhc.c +1 -0
@@ -1468,6 +1468,7 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)static struct platform_driver sdhci_esdhc_driver = { .driver = { .name ="sdhci-esdhc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_esdhc_of_match, .pm = &esdhc_of_dev_pm_ops, },
drivers/mmc/host/sdhci-of-hlwd.c +1 -0
@@ -80,6 +80,7 @@ MODULE_DEVICE_TABLE(of, sdhci_hlwd_of_match);static struct platform_driver sdhci_hlwd_driver = { .driver = { .name ="sdhci-hlwd" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_hlwd_of_match, .pm = &sdhci_pltfm_pmops, },
drivers/mmc/host/sdhci-pxav2.c +1 -0
@@ -226,6 +226,7 @@ static int sdhci_pxav2_probe(struct platform_device *pdev)static struct platform_driver sdhci_pxav2_driver = { .driver = { .name ="sdhci-pxav2" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (sdhci_pxav2_of_match), .pm = &sdhci_pltfm_pmops, },
drivers/mmc/host/sdhci-pxav3.c +1 -0
@@ -567,6 +567,7 @@ static const struct dev_pm_ops sdhci_pxav3_pmops = {static struct platform_driver sdhci_pxav3_driver = { .driver = { .name ="sdhci-pxav3" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (sdhci_pxav3_of_match), .pm = &sdhci_pxav3_pmops, },
drivers/mmc/host/sdhci-s3c.c +1 -0
@@ -784,6 +784,7 @@ static struct platform_driver sdhci_s3c_driver = { .id_table = sdhci_s3c_driver_ids, .driver = { .name = "s3c-sdhci", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(sdhci_s3c_dt_match), .pm = &sdhci_s3c_pmops, },
drivers/mmc/host/sdhci-sirf.c +1 -0
@@ -220,6 +220,7 @@ MODULE_DEVICE_TABLE(of, sdhci_sirf_of_match);static struct platform_driver sdhci_sirf_driver = { .driver = { .name ="sdhci-sirf" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_sirf_of_match, .pm = &sdhci_pltfm_pmops, },
drivers/mmc/host/sdhci-spear.c +1 -0
@@ -181,6 +181,7 @@ MODULE_DEVICE_TABLE(of, sdhci_spear_id_table);static struct platform_driver sdhci_driver = { .driver = { .name ="sdhci" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pm_ops, .of_match_table =of_match_ptr (sdhci_spear_id_table), },
drivers/mmc/host/sdhci-st.c +1 -0
@@ -521,6 +521,7 @@ static struct platform_driver sdhci_st_driver = { .remove = sdhci_st_remove, .driver = { .name = "sdhci-st", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &sdhci_st_pmops, .of_match_table = of_match_ptr(st_sdhci_match), },
drivers/mmc/host/sdhci-tegra.c +1 -0
@@ -1782,6 +1782,7 @@ static SIMPLE_DEV_PM_OPS(sdhci_tegra_dev_pm_ops, sdhci_tegra_suspend,static struct platform_driver sdhci_tegra_driver = { .driver = { .name ="sdhci-tegra" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_tegra_dt_match, .pm = &sdhci_tegra_dev_pm_ops, },
drivers/mmc/host/sdhci_f_sdh30.c +1 -0
@@ -219,6 +219,7 @@ MODULE_DEVICE_TABLE(acpi, f_sdh30_acpi_ids);static struct platform_driver sdhci_f_sdh30_driver = { .driver = { .name ="f_sdh30" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (f_sdh30_dt_ids), .acpi_match_table =ACPI_PTR (f_sdh30_acpi_ids), .pm = &sdhci_pltfm_pmops,
drivers/mmc/host/sh_mmcif.c +1 -0
@@ -1562,6 +1562,7 @@ static struct platform_driver sh_mmcif_driver = { .remove = sh_mmcif_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &sh_mmcif_dev_pm_ops, .of_match_table = sh_mmcif_of_match, },
drivers/mmc/host/sunxi-mmc.c +1 -0
@@ -1514,6 +1514,7 @@ static const struct dev_pm_ops sunxi_mmc_pm_ops = {static struct platform_driver sunxi_mmc_driver = { .driver = { .name ="sunxi-mmc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (sunxi_mmc_of_match), .pm = &sunxi_mmc_pm_ops, },
drivers/mmc/host/tmio_mmc.c +1 -0
@@ -213,6 +213,7 @@ static const struct dev_pm_ops tmio_mmc_dev_pm_ops = {static struct platform_driver tmio_mmc_driver = { .driver = { .name ="tmio-mmc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &tmio_mmc_dev_pm_ops, }, .probe = tmio_mmc_probe,
drivers/mmc/host/usdhi6rol0.c +1 -0
@@ -1890,6 +1890,7 @@ static struct platform_driver usdhi6_driver = { .remove = usdhi6_remove, .driver = { .name = "usdhi6rol0", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = usdhi6_of_match, }, };
drivers/mmc/host/wbsd.c +1 -0
@@ -1905,6 +1905,7 @@ static struct platform_driver wbsd_driver = { .resume = wbsd_platform_resume, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, }, };
drivers/mmc/host/wmt-sdmmc.c +1 -0
@@ -990,6 +990,7 @@ static struct platform_driver wmt_mci_driver = { .remove = wmt_mci_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = wmt_mci_pm_ops, .of_match_table = wmt_mci_dt_ids, },
Signature · 7 lines
2.28.0.526.ge36021eeef-goog _______________________________________________ Linux-rockchip mailing list [email protected] http://lists.infradead.org/mailman/listinfo/linux-rockchip
Re: [PATCH 1/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.4
Quoted 5 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.4 timeframe.Signed-off-by Douglas Anderson <[email protected]>
Positive comments from patch 3 apply here, too.
Quoted 1 line
drivers/mmc/host/sh_mmcif.c | 1 +
_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip
Re: [PATCH 1/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.4
In reply to Douglas AndersonQuoted 36 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.4 timeframe.Signed-off-by Douglas Anderson <[email protected]>---
drivers/mmc/host/android-goldfish.c | 1 + drivers/mmc/host/atmel-mci.c | 1 + drivers/mmc/host/au1xmmc.c | 1 + drivers/mmc/host/davinci_mmc.c | 1 + drivers/mmc/host/dw_mmc-exynos.c | 1 + drivers/mmc/host/dw_mmc-k3.c | 1 + drivers/mmc/host/dw_mmc-pltfm.c | 1 + drivers/mmc/host/dw_mmc-rockchip.c | 1 + drivers/mmc/host/jz4740_mmc.c | 1 + drivers/mmc/host/moxart-mmc.c | 1 + drivers/mmc/host/mtk-sd.c | 1 + drivers/mmc/host/mvsdio.c | 1 + drivers/mmc/host/mxcmmc.c | 1 + drivers/mmc/host/mxs-mmc.c | 1 + drivers/mmc/host/omap.c | 1 + drivers/mmc/host/omap_hsmmc.c | 1 + drivers/mmc/host/pxamci.c | 1 + drivers/mmc/host/rtsx_pci_sdmmc.c | 1 + drivers/mmc/host/rtsx_usb_sdmmc.c | 1 + drivers/mmc/host/s3cmci.c | 1 + drivers/mmc/host/sdhci-acpi.c | 1 + drivers/mmc/host/sdhci-bcm-kona.c | 1 + drivers/mmc/host/sdhci-cns3xxx.c | 1 + drivers/mmc/host/sdhci-dove.c | 1 + drivers/mmc/host/sdhci-esdhc-imx.c | 1 + drivers/mmc/host/sdhci-iproc.c | 1 + drivers/mmc/host/sdhci-of-arasan.c | 1 + drivers/mmc/host/sdhci-of-at91.c | 1 + drivers/mmc/host/sdhci-of-esdhc.c | 1 +
Quoted 8 lines
drivers/mmc/host/sdhci-of-hlwd.c | 1 + drivers/mmc/host/sdhci-pxav2.c | 1 + drivers/mmc/host/sdhci-pxav3.c | 1 + drivers/mmc/host/sdhci-s3c.c | 1 + drivers/mmc/host/sdhci-sirf.c | 1 + drivers/mmc/host/sdhci-spear.c | 1 + drivers/mmc/host/sdhci-st.c | 1 + drivers/mmc/host/sdhci-tegra.c | 1 +
I do see a failure on Tegra124, but comparing to recent test results from linux-next that seems to be a preexisting problem, so:
_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip
PATCH 2/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.9
In reply to Douglas AndersonThis is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.9 timeframe.
---
drivers/mmc/host/sdhci-brcmstb.c | 1 + drivers/mmc/host/sdhci-pic32.c | 1 + 2 files changed, 2 insertions(+)
drivers/mmc/host/sdhci-brcmstb.c +1 -0
@@ -326,6 +326,7 @@ MODULE_DEVICE_TABLE(of, sdhci_brcm_of_match);static struct platform_driver sdhci_brcmstb_driver = { .driver = { .name ="sdhci-brcmstb" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pltfm_pmops, .of_match_table =of_match_ptr (sdhci_brcm_of_match), },
drivers/mmc/host/sdhci-pic32.c +1 -0
@@ -241,6 +241,7 @@ MODULE_DEVICE_TABLE(of, pic32_sdhci_id_table);static struct platform_driver pic32_sdhci_driver = { .driver = { .name ="pic32-sdhci" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (pic32_sdhci_id_table), }, .probe = pic32_sdhci_probe,
Signature · 1 line
2.28.0.526.ge36021eeef-goog
PATCH 3/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.14
In reply to Douglas AndersonThis is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.14 timeframe.
---
drivers/mmc/host/bcm2835.c | 1 + drivers/mmc/host/cavium-octeon.c | 1 + drivers/mmc/host/dw_mmc-zx.c | 1 + drivers/mmc/host/meson-gx-mmc.c | 1 + drivers/mmc/host/renesas_sdhi_internal_dmac.c | 1 + drivers/mmc/host/renesas_sdhi_sys_dmac.c | 1 + drivers/mmc/host/sdhci-cadence.c | 1 + drivers/mmc/host/sdhci-xenon.c | 1 + 8 files changed, 8 insertions(+)
drivers/mmc/host/bcm2835.c +1 -0
@@ -1474,6 +1474,7 @@ static struct platform_driver bcm2835_driver = { .remove = bcm2835_remove, .driver = { .name = "sdhost-bcm2835", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = bcm2835_match, }, };
drivers/mmc/host/cavium-octeon.c +1 -0
@@ -327,6 +327,7 @@ static struct platform_driver octeon_mmc_driver = { .remove = octeon_mmc_remove, .driver = { .name = KBUILD_MODNAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = octeon_mmc_match, }, };
drivers/mmc/host/dw_mmc-zx.c +1 -0
@@ -222,6 +222,7 @@ static struct platform_driver dw_mci_zx_pltfm_driver = { .remove = dw_mci_pltfm_remove, .driver = { .name = "dwmmc_zx", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_zx_match, .pm = &dw_mci_zx_dev_pm_ops, },
drivers/mmc/host/meson-gx-mmc.c +1 -0
@@ -1264,6 +1264,7 @@ static struct platform_driver meson_mmc_driver = { .remove = meson_mmc_remove, .driver = { .name = DRIVER_NAME, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(meson_mmc_of_match), }, };
drivers/mmc/host/renesas_sdhi_internal_dmac.c +1 -0
@@ -357,6 +357,7 @@ static const struct dev_pm_ops renesas_sdhi_internal_dmac_dev_pm_ops = {static struct platform_driver renesas_internal_dmac_sdhi_driver = { .driver = { .name ="renesas_sdhi_internal_dmac" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &renesas_sdhi_internal_dmac_dev_pm_ops, .of_match_table = renesas_sdhi_internal_dmac_of_match, },
drivers/mmc/host/renesas_sdhi_sys_dmac.c +1 -0
@@ -463,6 +463,7 @@ static const struct dev_pm_ops renesas_sdhi_sys_dmac_dev_pm_ops = {static struct platform_driver renesas_sys_dmac_sdhi_driver = { .driver = { .name ="sh_mobile_sdhi" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &renesas_sdhi_sys_dmac_dev_pm_ops, .of_match_table = renesas_sdhi_sys_dmac_of_match, },
drivers/mmc/host/sdhci-cadence.c +1 -0
@@ -463,6 +463,7 @@ MODULE_DEVICE_TABLE(of, sdhci_cdns_match);static struct platform_driver sdhci_cdns_driver = { .driver = { .name ="sdhci-cdns" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_cdns_pm_ops, .of_match_table = sdhci_cdns_match, },
drivers/mmc/host/sdhci-xenon.c +1 -0
@@ -677,6 +677,7 @@ MODULE_DEVICE_TABLE(of, sdhci_xenon_dt_ids);static struct platform_driver sdhci_xenon_driver = { .driver = { .name ="xenon-sdhci" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_xenon_dt_ids, .pm = &sdhci_xenon_dev_pm_ops, },
Signature · 1 line
2.28.0.526.ge36021eeef-goog
Re: [PATCH 3/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.14
Quoted 3 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.14 timeframe.
The LTS granularity of patches is a nice idea! I will keep it in mind.
Quoted 2 lines
Signed-off-by Douglas Anderson <[email protected]>
In general, I like more coverage of async probing. It makes sense for SD/MMC, I think.
Quoted 2 lines
drivers/mmc/host/renesas_sdhi_internal_dmac.c | 1 + drivers/mmc/host/renesas_sdhi_sys_dmac.c | 1 +
For these two also:
The speedups somewhat match the expected values and no regressions when checksumming a large file.
PATCH 4/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.19
In reply to Douglas AndersonThis is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v4.19 timeframe.
---
drivers/mmc/host/dw_mmc-bluefield.c | 1 + drivers/mmc/host/dw_mmc-hi3798cv200.c | 1 + drivers/mmc/host/meson-mx-sdio.c | 1 + drivers/mmc/host/sdhci-of-dwcmshc.c | 1 + drivers/mmc/host/sdhci-omap.c | 1 + 5 files changed, 5 insertions(+)
drivers/mmc/host/dw_mmc-bluefield.c +1 -0
@@ -55,6 +55,7 @@ static struct platform_driver dw_mci_bluefield_pltfm_driver = { .remove = dw_mci_pltfm_remove, .driver = { .name = "dwmmc_bluefield", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_bluefield_match, .pm = &dw_mci_pltfm_pmops, },
drivers/mmc/host/dw_mmc-hi3798cv200.c +1 -0
@@ -200,6 +200,7 @@ static struct platform_driver dw_mci_hi3798cv200_driver = { .remove = dw_mci_hi3798cv200_remove, .driver = { .name = "dwmmc_hi3798cv200", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = dw_mci_hi3798cv200_match, }, };
drivers/mmc/host/meson-mx-sdio.c +1 -0
@@ -755,6 +755,7 @@ static struct platform_driver meson_mx_mmc_driver = { .remove = meson_mx_mmc_remove, .driver = { .name = "meson-mx-sdio", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(meson_mx_mmc_of_match), }, };
drivers/mmc/host/sdhci-of-dwcmshc.c +1 -0
@@ -214,6 +214,7 @@ MODULE_DEVICE_TABLE(of, sdhci_dwcmshc_dt_ids);static struct platform_driver sdhci_dwcmshc_driver = { .driver = { .name ="sdhci-dwcmshc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_dwcmshc_dt_ids, .pm = &dwcmshc_pmops, },
drivers/mmc/host/sdhci-omap.c +1 -0
@@ -1297,6 +1297,7 @@ static struct platform_driver sdhci_omap_driver = { .remove = sdhci_omap_remove, .driver = { .name = "sdhci-omap", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &sdhci_omap_dev_pm_ops, .of_match_table = omap_sdhci_match, },
Signature · 1 line
2.28.0.526.ge36021eeef-goog
PATCH 5/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v5.4
In reply to Douglas AndersonThis is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to be around in the v5.4 timeframe.
---
drivers/mmc/host/alcor.c | 1 + drivers/mmc/host/sdhci-of-aspeed.c | 2 ++ drivers/mmc/host/sdhci-sprd.c | 1 + drivers/mmc/host/sdhci_am654.c | 1 + drivers/mmc/host/uniphier-sd.c | 1 + 5 files changed, 6 insertions(+)
drivers/mmc/host/alcor.c +1 -0
@@ -1178,6 +1178,7 @@ static struct platform_driver alcor_pci_sdmmc_driver = { .id_table = alcor_pci_sdmmc_ids, .driver = { .name = DRV_NAME_ALCOR_PCI_SDMMC, + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .pm = &alcor_mmc_pm_ops }, };
drivers/mmc/host/sdhci-of-aspeed.c +2 -0
@@ -240,6 +240,7 @@ static const struct of_device_id aspeed_sdhci_of_match[] = {static struct platform_driver aspeed_sdhci_driver = { .driver = { .name ="sdhci-aspeed" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = aspeed_sdhci_of_match, }, .probe = aspeed_sdhci_probe,
@@ -318,6 +319,7 @@ MODULE_DEVICE_TABLE(of, aspeed_sdc_of_match);static struct platform_driver aspeed_sdc_driver = { .driver = { .name ="sd-controller-aspeed" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .pm = &sdhci_pltfm_pmops, .of_match_table = aspeed_sdc_of_match, },
drivers/mmc/host/sdhci-sprd.c +1 -0
@@ -787,6 +787,7 @@ static struct platform_driver sdhci_sprd_driver = { .remove = sdhci_sprd_remove, .driver = { .name = "sdhci_sprd_r11", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(sdhci_sprd_of_match), .pm = &sdhci_sprd_pm_ops, },
drivers/mmc/host/sdhci_am654.c +1 -0
@@ -745,6 +745,7 @@ static int sdhci_am654_remove(struct platform_device *pdev)static struct platform_driver sdhci_am654_driver = { .driver = { .name ="sdhci-am654" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_am654_of_match, }, .probe = sdhci_am654_probe,
drivers/mmc/host/uniphier-sd.c +1 -0
@@ -685,6 +685,7 @@ static struct platform_driver uniphier_sd_driver = { .remove = uniphier_sd_remove, .driver = { .name = "uniphier-sd", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = uniphier_sd_match, }, };
Signature · 1 line
2.28.0.526.ge36021eeef-goog
PATCH 6/6 mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that are newer than 5.4
In reply to Douglas AndersonThis is like commit 3d3451124f3d ("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to have been added after kernel 5.4.
---
drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + 5 files changed, 5 insertions(+)
drivers/mmc/host/meson-mx-sdhc-mmc.c +1 -0
@@ -903,6 +903,7 @@ static struct platform_driver meson_mx_sdhc_driver = { .remove = meson_mx_sdhc_remove, .driver = { .name = "meson-mx-sdhc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(meson_mx_sdhc_of_match), }, };
drivers/mmc/host/owl-mmc.c +1 -0
@@ -689,6 +689,7 @@ MODULE_DEVICE_TABLE(of, owl_mmc_of_match);static struct platform_driver owl_mmc_driver = { .driver = { .name ="owl_mmc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = owl_mmc_of_match, }, .probe = owl_mmc_probe,
drivers/mmc/host/sdhci-esdhc-mcf.c +1 -0
@@ -509,6 +509,7 @@ static int sdhci_esdhc_mcf_remove(struct platform_device *pdev)static struct platform_driver sdhci_esdhc_mcf_driver = { .driver = { .name ="sdhci-esdhc-mcf" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , }, .probe = sdhci_esdhc_mcf_probe, .remove = sdhci_esdhc_mcf_remove,
drivers/mmc/host/sdhci-milbeaut.c +1 -0
@@ -333,6 +333,7 @@ static int sdhci_milbeaut_remove(struct platform_device *pdev)static struct platform_driver sdhci_milbeaut_driver = { .driver = { .name ="sdhci-milbeaut" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (mlb_dt_ids), }, .probe = sdhci_milbeaut_probe,
drivers/mmc/host/sdhci-of-sparx5.c +1 -0
@@ -255,6 +255,7 @@ MODULE_DEVICE_TABLE(of, sdhci_sparx5_of_match);static struct platform_driver sdhci_sparx5_driver = { .driver = { .name ="sdhci-sparx5" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_sparx5_of_match, .pm = &sdhci_pltfm_pmops, },
Signature · 1 line
2.28.0.526.ge36021eeef-goog
Re: [PATCH 6/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that are newer than 5.4
Quoted 14 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to have been added after kernel 5.4.Signed-off-by Douglas Anderson <[email protected]>---
drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + 5 files changed, 5 insertions(+)
[snip]
drivers/mmc/host/sdhci-of-sparx5.c MODULE_DEVICE_TABLE(of, sdhci_sparx5_of_match); · line 261
Quoted 12 lines
drivers/mmc/host/sdhci-of-sparx5.c +1
-0@@ -255,6 +255,7 @@ MODULE_DEVICE_TABLE(of, sdhci_sparx5_of_match);static struct platform_driver sdhci_sparx5_driver = { .driver = { .name ="sdhci-sparx5" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_sparx5_of_match, .pm = &sdhci_pltfm_pmops, },
Signature · 2 lines
Lars Povlsen, Microchip
Re: [PATCH 6/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that are newer than 5.4
In reply to Douglas AndersonQuoted 13 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to have been added after kernel 5.4.Signed-off-by Douglas Anderson <[email protected]>---
drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + 5 files changed, 5 insertions(+)
[snip]
Quoted 8 lines
drivers/mmc/host/sdhci-esdhc-mcf.c +0
-0
drivers/mmc/host/sdhci-esdhc-mcf.c static int sdhci_esdhc_mcf_remove(struct platform_device *pdev) · line 515
Quoted 7 lines
drivers/mmc/host/sdhci-esdhc-mcf.c +1
-0@@ -509,6 +509,7 @@ static int sdhci_esdhc_mcf_remove(struct platform_device *pdev)static struct platform_driver sdhci_esdhc_mcf_driver = { .driver = { .name ="sdhci-esdhc-mcf" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , }, .probe = sdhci_esdhc_mcf_probe, .remove = sdhci_esdhc_mcf_remove,
Re: [PATCH 6/6] mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that are newer than 5.4
In reply to Douglas AndersonQuoted 4 lines · Douglas Anderson
This is like commit
3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe") but applied to a whole pile of drivers. This batch converts the drivers that appeared to have been added after kernel 5.4.
Backporting made easy ;)
Quoted 1 line
Signed-off-by Douglas Anderson <[email protected]>
For owl-mmc,
Thanks,
Mani
Quoted 72 lines
---
drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + 5 files changed, 5 insertions(+)drivers/mmc/host/meson-mx-sdhc-mmc.c +1
-0@@ -903,6 +903,7 @@ static struct platform_driver meson_mx_sdhc_driver = { .remove = meson_mx_sdhc_remove, .driver = { .name = "meson-mx-sdhc", + .probe_type = PROBE_PREFER_ASYNCHRONOUS, .of_match_table = of_match_ptr(meson_mx_sdhc_of_match), }, };drivers/mmc/host/owl-mmc.c +1
-0@@ -689,6 +689,7 @@ MODULE_DEVICE_TABLE(of, owl_mmc_of_match);static struct platform_driver owl_mmc_driver = { .driver = { .name ="owl_mmc" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = owl_mmc_of_match, }, .probe = owl_mmc_probe,drivers/mmc/host/sdhci-esdhc-mcf.c +1
-0@@ -509,6 +509,7 @@ static int sdhci_esdhc_mcf_remove(struct platform_device *pdev)static struct platform_driver sdhci_esdhc_mcf_driver = { .driver = { .name ="sdhci-esdhc-mcf" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , }, .probe = sdhci_esdhc_mcf_probe, .remove = sdhci_esdhc_mcf_remove,drivers/mmc/host/sdhci-milbeaut.c +1
-0@@ -333,6 +333,7 @@ static int sdhci_milbeaut_remove(struct platform_device *pdev)static struct platform_driver sdhci_milbeaut_driver = { .driver = { .name ="sdhci-milbeaut" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table =of_match_ptr (mlb_dt_ids), }, .probe = sdhci_milbeaut_probe,drivers/mmc/host/sdhci-of-sparx5.c +1
-0@@ -255,6 +255,7 @@ MODULE_DEVICE_TABLE(of, sdhci_sparx5_of_match);static struct platform_driver sdhci_sparx5_driver = { .driver = { .name ="sdhci-sparx5" , + .probe_type =PROBE_PREFER_ASYNCHRONOUS , .of_match_table = sdhci_sparx5_of_match, .pm = &sdhci_pltfm_pmops, },Signature · 1 line
2.28.0.526.ge36021eeef-goog
_______________________________________________
linux-amlogic mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-amlogic
Martin Blumenstingl
reviewerHi Douglas,
Quoted 6 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").
can this somehow change the order in which the MMC drivers end up loading? on Amlogic SoCs we have up to three MMC controllers, some SoCs even use two different MMC controller IPs (and therefore two different drivers).
so far the MMC controller naming (/dev/mmcblk* etc.) was consistent - can that change with this patch?
apologies if this has been discussed and answered anywhere
Best regards,
Martin
Hi Martin.
Quoted 10 lines · Martin Blumenstingl
Hi Douglas,
Quoted 6 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").
Quoted 7 lines
can this somehow change the order in which the MMC drivers end up loading? on Amlogic SoCs we have up to three MMC controllers, some SoCs even use two different MMC controller IPs (and therefore two different drivers).
so far the MMC controller naming (/dev/mmcblk*etc.) was consistent - can that change with this patch?
We could resolve this by setting up aliases for mmc nodes in the dts.
Quoted 5 lines
apologies if this has been discussed and answered anywhere
Best regards,
Martin
Best regards
-Anand
Quoted 23 lines · Anand Moon
Hi Martin.
Quoted 10 lines · Martin Blumenstingl
Hi Douglas,
Quoted 6 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").Quoted 7 lines
can this somehow change the order in which the MMC drivers end up loading? on Amlogic SoCs we have up to three MMC controllers, some SoCs even use two different MMC controller IPs (and therefore two different drivers).
so far the MMC controller naming (/dev/mmcblk*etc.) was consistent - can that change with this patch?
Quoted 1 line
We could resolve this by setting up aliases for mmc nodes in the dts.
Right now, only the dw_mmc driver family supports aliases for mmc nodes.
Quoted 8 lines
Quoted 5 lines
apologies if this has been discussed and answered anywhere
Best regards,
MartinBest regards
-Anand
Chen-Yu Tsai
maintainer(Resent from kernel.org)
Quoted 23 lines · Anand Moon
Hi Martin.
Quoted 10 lines · Martin Blumenstingl
Hi Douglas,
Quoted 6 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").Quoted 7 lines
can this somehow change the order in which the MMC drivers end up loading? on Amlogic SoCs we have up to three MMC controllers, some SoCs even use two different MMC controller IPs (and therefore two different drivers).
so far the MMC controller naming (/dev/mmcblk*etc.) was consistent - can that change with this patch?
Quoted 1 line
We could resolve this by setting up aliases for mmc nodes in the dts.
Right now, only the dw_mmc driver family supports aliases for mmc nodes.
Quoted 8 lines
Quoted 5 lines
apologies if this has been discussed and answered anywhere
Best regards,
MartinBest regards
-Anand
Quoted 26 lines · Chen-Yu Tsai
(Resent from kernel.org)
Quoted 22 lines · Anand Moon
Hi Martin.
Quoted 17 lines · Martin Blumenstingl
Hi Douglas,
Quoted 6 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").can this somehow change the order in which the MMC drivers end up loading? on Amlogic SoCs we have up to three MMC controllers, some SoCs even use two different MMC controller IPs (and therefore two different drivers).
so far the MMC controller naming (/dev/mmcblk*etc.) was consistent - can that change with this patch?
Consistency has never been guaranteed. Just imagine one of the mmc host drivers ending up lacking some of its resources during ->probe() and returning -EPROBE_DEFER.
UUID/PARTID has been the only way.
Quoted 2 lines
Quoted 2 lines
We could resolve this by setting up aliases for mmc nodes in the dts.
Yes, this is now possible due to the recently [1] applied patches for supporting mmc aliases.
Quoted 2 lines
Right now, only the
dw_mmcdriver family supports aliases for mmc nodes.
That's "mshc" specific (related to the old multiple slot support I think), but not affecting the mmc aliases, which the mmc core now is supporting.
So the mmc aliases are common for all mmc hosts. Please have a look at the new OF parsing in mmc_alloc_host().
[...]
Kind regards
Uffe
[1]
https://patchwork.kernel.org/patch/11747669/
https://patchwork.kernel.org/patch/11747671/
Quoted 148 lines · Douglas Anderson
As per discussion [1], it seems like it should be quite safe to turn on
PROBE_PREFER_ASYNCHRONOUSfor all sd/mmc host controllers. Let's give it a shot. For some discussion about this flag, see the commit message for commit3d3451124f3d("mmc: sdhci-msm: Prefer asynchronous probe").I've broken this series into chunks based on LTS kernel releases to attempt to make it easier if someone wanted to cherry-pick it to an older kernel. While these cherry-picks won't be conflict free, there should only be trivial context conflicts and no problems with drivers that are totally missing. This is a bit of a compromise between a 1-big patch and a many-part patch series.
I have only tested this on a rk3399-kevin (sdhci-of-arasan) and a rk3288-veyron (dw_mmc-rockchip) device and only lightly. If this patch causes anyone problems those drivers should be marked with
PROBE_FORCE_SYNCHRONOUS, debugged, and then go back to prefer asynchronous. Any problems are likely just a hidden bug that has been exposed by this change.NOTE: in theory, it'd be nice if there was a KConfig option that we could flip that would turn on async probe everywhere (except for those that opt out by adding
PROBE_FORCE_SYNCHRONOUS). My hope is that by adding this flag in more places it will become clear that this flag is working reliably and easier to move over when we're ready.While coccinelle is too difficult for my feeble brain, I managed to whip up a pretty terrible python script to help with this. For your edification:
import os
import sys
import refor filename in
os.listdir("."):
found_plat= False
found_driver= False
output = []
for line inopen(filename, "r").readlines():
output.append(line)if "struct
platform_driver" in line:
found_plat= True
iffound_platandre.search(r"\t\.driver\s*=\s*{", line):
found_driver= True
found_plat= False
mo =re.search(r"(\s*)\.name(\s*)=", line)
iffound_driverand mo:
ifmo.group(2)== " ":
space = " "
elifmo.group(2)== "\t":
# Best we can do
space = " "
elif mo.group(2).startswith("\t"):
# Guess that removing one tab is right space = mo.group(2)[1:]
else:
# Guess it's all spaces
space = mo.group(2)[7:] + " "
output.append("%s.probe_type%s= PROBE_PREFER_ASYNCHRONOUS,\n" % (mo.group(1), space))found_driver= False
open(filename, "w").write("".join(output))[1] https://lore.kernel.org/r/CAPDyKFq31bucJhP9hp1HSqh-qM2uNGHgDoyQpmbJf00nEf_T4Q at mail.gmail.com/
Douglas Anderson (6):
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that existed in v4.4
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that existed in v4.9
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that existed in v4.14
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that existed in v4.19
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that existed in v5.4
mmc: SetPROBE_PREFER_ASYNCHRONOUSfor drivers that are newer than 5.4drivers/mmc/host/alcor.c | 1 + drivers/mmc/host/android-goldfish.c | 1 + drivers/mmc/host/atmel-mci.c | 1 + drivers/mmc/host/au1xmmc.c | 1 + drivers/mmc/host/bcm2835.c | 1 + drivers/mmc/host/cavium-octeon.c | 1 + drivers/mmc/host/davinci_mmc.c | 1 + drivers/mmc/host/dw_mmc-bluefield.c | 1 + drivers/mmc/host/dw_mmc-exynos.c | 1 + drivers/mmc/host/dw_mmc-hi3798cv200.c | 1 + drivers/mmc/host/dw_mmc-k3.c | 1 + drivers/mmc/host/dw_mmc-pltfm.c | 1 + drivers/mmc/host/dw_mmc-rockchip.c | 1 + drivers/mmc/host/dw_mmc-zx.c | 1 + drivers/mmc/host/jz4740_mmc.c | 1 + drivers/mmc/host/meson-gx-mmc.c | 1 + drivers/mmc/host/meson-mx-sdhc-mmc.c | 1 + drivers/mmc/host/meson-mx-sdio.c | 1 + drivers/mmc/host/moxart-mmc.c | 1 + drivers/mmc/host/mtk-sd.c | 1 + drivers/mmc/host/mvsdio.c | 1 + drivers/mmc/host/mxcmmc.c | 1 + drivers/mmc/host/mxs-mmc.c | 1 + drivers/mmc/host/omap.c | 1 + drivers/mmc/host/omap_hsmmc.c | 1 + drivers/mmc/host/owl-mmc.c | 1 + drivers/mmc/host/pxamci.c | 1 + drivers/mmc/host/renesas_sdhi_internal_dmac.c | 1 + drivers/mmc/host/renesas_sdhi_sys_dmac.c | 1 + drivers/mmc/host/rtsx_pci_sdmmc.c | 1 + drivers/mmc/host/rtsx_usb_sdmmc.c | 1 + drivers/mmc/host/s3cmci.c | 1 + drivers/mmc/host/sdhci-acpi.c | 1 + drivers/mmc/host/sdhci-bcm-kona.c | 1 + drivers/mmc/host/sdhci-brcmstb.c | 1 + drivers/mmc/host/sdhci-cadence.c | 1 + drivers/mmc/host/sdhci-cns3xxx.c | 1 + drivers/mmc/host/sdhci-dove.c | 1 + drivers/mmc/host/sdhci-esdhc-imx.c | 1 + drivers/mmc/host/sdhci-esdhc-mcf.c | 1 + drivers/mmc/host/sdhci-iproc.c | 1 + drivers/mmc/host/sdhci-milbeaut.c | 1 + drivers/mmc/host/sdhci-of-arasan.c | 1 + drivers/mmc/host/sdhci-of-aspeed.c | 2 ++ drivers/mmc/host/sdhci-of-at91.c | 1 + drivers/mmc/host/sdhci-of-dwcmshc.c | 1 + drivers/mmc/host/sdhci-of-esdhc.c | 1 + drivers/mmc/host/sdhci-of-hlwd.c | 1 + drivers/mmc/host/sdhci-of-sparx5.c | 1 + drivers/mmc/host/sdhci-omap.c | 1 + drivers/mmc/host/sdhci-pic32.c | 1 + drivers/mmc/host/sdhci-pxav2.c | 1 + drivers/mmc/host/sdhci-pxav3.c | 1 + drivers/mmc/host/sdhci-s3c.c | 1 + drivers/mmc/host/sdhci-sirf.c | 1 + drivers/mmc/host/sdhci-spear.c | 1 + drivers/mmc/host/sdhci-sprd.c | 1 + drivers/mmc/host/sdhci-st.c | 1 + drivers/mmc/host/sdhci-tegra.c | 1 + drivers/mmc/host/sdhci-xenon.c | 1 + drivers/mmc/host/sdhci_am654.c | 1 + drivers/mmc/host/sdhci_f_sdh30.c | 1 + drivers/mmc/host/sh_mmcif.c | 1 + drivers/mmc/host/sunxi-mmc.c | 1 + drivers/mmc/host/tmio_mmc.c | 1 + drivers/mmc/host/uniphier-sd.c | 1 + drivers/mmc/host/usdhi6rol0.c | 1 + drivers/mmc/host/wbsd.c | 1 + drivers/mmc/host/wmt-sdmmc.c | 1 + 69 files changed, 70 insertions(+)
Quoted 3 lines
--
2.28.0.526.ge36021eeef-goog
An interesting idea about a patch per LTS release. I think it makes perfect sense in this type of case.
Let's give this a shot in next and see how it goes. So, applied for next, thanks!
Kind regards
Uffe