[PATCH v2 00/46] dm: Print device name in dev_xxx like Linux
Tom Rini
trini at konsulko.com
Wed Sep 30 15:01:39 CEST 2020
On Tue, Sep 15, 2020 at 10:44:36AM -0400, Sean Anderson wrote:
> This series adds some additional information to dev_xxx output. This requires
> the first argument to dev_xxx to be a struct udevice. The first argument has
> never been used in U-Boot, so many drivers have some invalid calls. To fix these
> drivers, the following strategies were generally followed
>
> * If there is a udevice already passed to the function, use it.
> * If there is a udevice contained in a struct passed to the function, use it.
> When there are multiple possible devices, I have tried to use whatever "makes
> sense" given the content of the message.
> * If there is no udevice passed to the function either directly or indirectly,
> but all the callers of said function can access a udevice, modify the function
> signature to pass in a udevice and use that.
> * If the driver does not use DM, convert the log statements to log_xxx.
> * If the driver uses DM only some of the time, use ifdefs to select an
> appropriate log function.
>
> I have done a little bit of cleanup beyond these strategies, but for the most
> part I have tried to *only* to the minimum necessary. Many drivers could use a
> follow-up patch to convert their printf()s and debug()s to an appropriate log
> function.
>
> I have tried to CC relevant maintainers/reviewers only for patches they
> maintain/review. If I have left someone out, please let me know and I will add
> them.
>
> CI is passing at [1]. This does not necessarily mean that this series is free of
> bugs, so I would appreciate testing on hardware. During my refactoring, I
> noticed a few bugs like
>
> struct udevice *dev;
>
> if (some_condition)
> dev_err(dev, "Something went wrong\n");
>
> dev = some_function();
>
> I have fixed such errors when I have encountered them, but they will not always
> warn or error on build.
>
> Removal of the duplicate definitions in <linux/compat.h> will be done in a
> follow-up patch due to the already-large size of this series.
>
> To test this series on sandbox, apply the following options in addition to
> sandbox_defconfig
>
> CONFIG_LOGLEVEL=8
> CONFIG_LOG=n
> CONFIG_CMD_LOG=n
> CONFIG_CMD_TPM=n
>
> Note that there is a soft dependency on [2] if you would like to test this patch
> with CONFIG_LOG and a higher LOG_LEVEL than LOGL_INFO.
>
> [1] https://dev.azure.com/seanga2/u-boot/_build/results?buildId=34&view=results
> [2] https://patchwork.ozlabs.org/project/uboot/list/?series=201343
>
> Changes in v2:
> - Support logging with struct device as well as struct udevice. A lot of
> drivers, especially USB gadgets, log with devices and not udevices. There
> is no major reason why they can't use udevice, but big changes like that
> are outside the scope of this series.
> - Add some comments to __dev_printk and dev_printk_emit
> - Handle struct device as well as struct udevice
> - Match format strings for the NULL path to the regular path. This reduces the
> amount of duplicated strings.
> - Print the device name before the driver name
> - Many new patches added to fix build bugs
>
> Sean Anderson (46):
> dm: syscon: Fix calling dev_dbg with an uninitialized device
> firmware: ti_sci: Fix not calling dev_err with a device
> i2c: mxc: Fix dev_err being called on a nonexistant variable
> mtd: nand: pxa3xx: Fix not calling dev_xxx with a device
> mtd: nand: sunxi: Fix not calling dev_err with a device
> mtd: spi: Include dm.h in spi-nor-core.c
> mtd: spi: Fix logging in spi-nor-tiny
> mtd: spi-nand: Fix not calling dev_err with a device
> mmc: Add mmc_dev()
> mmc: bcm2835-host: Fix not calling dev_dbg with a device
> mmc: mtk-sd: Fix not calling dev_err with a device
> mailbox: k3: Fix not calling dev_err with a device
> nand: atmel: Fix not calling dev_xxx with a device
> nand: brcmnand: Fix not calling dev_err() with a device
> nand: vybrid: Re-introduce vf610_nfc.dev
> net: bcm6368: Fix not calling dev_info with a device
> net: mdio: Fix not calling dev_dbg with a device
> net: mvneta: Fix not always calling dev_err with a device
> net: mvneta: Convert netdev_xxx to dev_xxx
> net: mvpp2: Fix not calling dev_xxx with a device
> net: mvpp2: Convert netdev_xxx to dev_xxx
> linux/compat.h: Remove netdev_xxx functions
> net: sunxi: Fix not calling dev_xxx with a device
> net: sun8i_emac: Fix not calling dev_xxx with a device
> net: ti: cpsw: Fix not calling dev_dbg with a device
> phy: marvell: Fix not calling dev_err with a device
> phy: rockchip: Fix not calling dev_err with a device
> phy: sun4i-usb: Fix not calling dev_err with a device
> phy: ti: Fix not calling dev_err with a device
> phy: usbphyc: Fix not calling dev_err with a device
> remoteproc: Remove unused function rproc_elf_sanity_check
> remoteproc: k3-r5: Fix not calling dev_xxx with a device
> remoteproc: k3: Fix not calling dev_xxx with a device
> soc: qualcomm: Fix not calling dev_err with a device
> spi: sunxi: Fix not calling dev_err with a device
> spi: zynqmp_gqspi: Fix not calling dev_err with a device
> sysreset: ti: Fix not calling dev_err with a device
> usb: cdns3: Fix not calling dev_xxx with a device
> usb: dwc2: Fix not calling dev_xxx with a device
> usb: dwc3: Fix not calling dev_xxx with a device
> usb: dwc3: ti: Fix not calling dev_err with a device
> usb: dwc3: Don't include asm-generic/io.h
> usb: musb-new: sunxi: Fix not calling dev_err with a device
> video: stm32: Fix not calling dev_xxx with a device
> dm: Use symbolic constants for log levels in dev_xxx
> dm: Print device name in dev_xxx like Linux
>
> drivers/core/syscon-uclass.c | 2 +-
> drivers/firmware/ti_sci.c | 23 ++--
> drivers/i2c/mxc_i2c.c | 7 +-
> drivers/mailbox/k3-sec-proxy.c | 6 +-
> drivers/mmc/bcm2835_sdhost.c | 34 +++---
> drivers/mmc/mtk-sd.c | 24 ++--
> drivers/mtd/nand/raw/atmel_nand.c | 69 +++++++-----
> drivers/mtd/nand/raw/brcmnand/brcmnand.c | 20 +---
> drivers/mtd/nand/raw/pxa3xx_nand.c | 30 ++---
> drivers/mtd/nand/raw/sunxi_nand.c | 45 ++++----
> drivers/mtd/nand/raw/vf610_nfc.c | 38 ++++---
> drivers/mtd/nand/spi/core.c | 8 +-
> drivers/mtd/spi/spi-nor-core.c | 1 +
> drivers/mtd/spi/spi-nor-tiny.c | 21 +++-
> drivers/net/bcm6368-eth.c | 3 +-
> drivers/net/mvneta.c | 56 +++++-----
> drivers/net/mvpp2.c | 87 ++++++++-------
> drivers/net/sun8i_emac.c | 9 +-
> drivers/net/sunxi_emac.c | 5 +-
> drivers/net/ti/cpsw.c | 6 +
> drivers/phy/allwinner/phy-sun4i-usb.c | 12 +-
> drivers/phy/marvell/comphy_core.c | 6 +-
> drivers/phy/phy-stm32-usbphyc.c | 2 +-
> drivers/phy/phy-ti-am654.c | 4 +-
> drivers/phy/rockchip/phy-rockchip-pcie.c | 14 +--
> drivers/phy/rockchip/phy-rockchip-typec.c | 6 +-
> drivers/remoteproc/k3_system_controller.c | 9 +-
> drivers/remoteproc/rproc-elf-loader.c | 16 ---
> drivers/remoteproc/ti_k3_r5f_rproc.c | 24 ++--
> drivers/smem/msm_smem.c | 2 +-
> drivers/spi/spi-sunxi.c | 6 +-
> drivers/spi/zynqmp_gqspi.c | 6 +-
> drivers/sysreset/sysreset-ti-sci.c | 3 +-
> drivers/usb/cdns3/ep0.c | 5 +-
> drivers/usb/cdns3/gadget.c | 3 +-
> drivers/usb/dwc3/core.c | 15 +--
> drivers/usb/dwc3/dwc3-generic.c | 1 -
> drivers/usb/dwc3/ep0.c | 1 +
> drivers/usb/dwc3/gadget.c | 23 ++--
> drivers/usb/dwc3/ti_usb_phy.c | 4 +-
> drivers/usb/host/dwc2.c | 39 ++++---
> drivers/usb/musb-new/sunxi.c | 9 +-
> drivers/video/dw_mipi_dsi.c | 24 ++--
> include/dm/device_compat.h | 127 ++++++++++++++++------
> include/linux/compat.h | 19 ----
> include/mmc.h | 2 +
> include/remoteproc.h | 13 ---
> net/mdio-uclass.c | 4 +-
> 48 files changed, 486 insertions(+), 407 deletions(-)
> Tested-by: Patrick Delaunay <patrick.delaunay at st.com>
I'm about to apply this series to -next and I'm cc'ing the board
maintainers list. A feature of this series is that a large number of
drivers that had put in information / error / etc messages that had
intended to be seen (presumably) by the user and were being discarded at
link/compile time before are included now. buildman has a set of flags
to tell you what exact functions grew between commits and if there's
messages that should have their log level changed, we really should do
that.
--
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20200930/168c88c8/attachment.sig>
More information about the U-Boot
mailing list