From: Adriana Kobylak <[email protected]>
Add support to detect USB flash drives and create the /dev/sd* devices.
Also add support for vfat to support USB drives formatted as FAT32.
This support will be used to enable firmware updates via USB flash
drives where the firmware image is stored in the USB drive and it's
plugged into the BMC USB port.
Signed-off-by: Adriana Kobylak <[email protected]>
Tested-by: Adriana Kobylak <[email protected]>
---
arch/arm/configs/aspeed_g5_defconfig | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/arch/arm/configs/aspeed_g5_defconfig b/arch/arm/configs/aspeed_g5_defconfig
index bee9422919aa..1b0d82c64ad4 100644
--- a/arch/arm/configs/aspeed_g5_defconfig
+++ b/arch/arm/configs/aspeed_g5_defconfig
@@ -37,11 +37,9 @@ CONFIG_KEXEC=y
CONFIG_VFP=y
CONFIG_NEON=y
CONFIG_KERNEL_MODE_NEON=y
-CONFIG_FIRMWARE_MEMMAP=y
CONFIG_KPROBES=y
CONFIG_JUMP_LABEL=y
CONFIG_MODULES=y
-# CONFIG_BLK_DEV_BSG is not set
# CONFIG_BLK_DEBUG_FS is not set
# CONFIG_MQ_IOSCHED_DEADLINE is not set
# CONFIG_MQ_IOSCHED_KYBER is not set
@@ -98,6 +96,7 @@ CONFIG_NET_NCSI=y
CONFIG_DEVTMPFS=y
CONFIG_DEVTMPFS_MOUNT=y
# CONFIG_PREVENT_FIRMWARE_BUILD is not set
+CONFIG_FIRMWARE_MEMMAP=y
CONFIG_MTD=y
CONFIG_MTD_BLOCK=y
CONFIG_MTD_PARTITIONED_MASTER=y
@@ -111,6 +110,8 @@ CONFIG_BLK_DEV_LOOP=y
CONFIG_BLK_DEV_NBD=y
CONFIG_EEPROM_AT24=y
CONFIG_EEPROM_AT25=y
+CONFIG_SCSI=y
+CONFIG_BLK_DEV_SD=y
CONFIG_NETDEVICES=y
CONFIG_NETCONSOLE=y
# CONFIG_NET_VENDOR_ALACRITECH is not set
@@ -230,6 +231,7 @@ CONFIG_USB_DYNAMIC_MINORS=y
CONFIG_USB_EHCI_HCD=y
CONFIG_USB_EHCI_ROOT_HUB_TT=y
CONFIG_USB_EHCI_HCD_PLATFORM=y
+CONFIG_USB_STORAGE=y
CONFIG_USB_GADGET=y
CONFIG_USB_ASPEED_VHUB=y
CONFIG_USB_CONFIGFS=y
@@ -277,6 +279,7 @@ CONFIG_FSI_OCC=y
CONFIG_EXT4_FS=y
CONFIG_FANOTIFY=y
CONFIG_OVERLAY_FS=y
+CONFIG_VFAT_FS=y
CONFIG_TMPFS=y
CONFIG_JFFS2_FS=y
# CONFIG_JFFS2_FS_WRITEBUFFER is not set
@@ -292,6 +295,8 @@ CONFIG_PSTORE_PMSG=y
CONFIG_PSTORE_FTRACE=y
CONFIG_PSTORE_RAM=y
# CONFIG_NETWORK_FILESYSTEMS is not set
+CONFIG_NLS_CODEPAGE_437=y
+CONFIG_NLS_ISO8859_1=y
CONFIG_HARDENED_USERCOPY=y
CONFIG_FORTIFY_SOURCE=y
CONFIG_CRYPTO_HMAC=y
--
2.25.1
On Fri, Nov 12, 2021 at 12:29:31PM PST, Adriana Kobylak wrote:
>From: Adriana Kobylak <[email protected]>
>
>Add support to detect USB flash drives and create the /dev/sd* devices.
>Also add support for vfat to support USB drives formatted as FAT32.
>This support will be used to enable firmware updates via USB flash
>drives where the firmware image is stored in the USB drive and it's
>plugged into the BMC USB port.
>
Hmm, how common is it for BMCs to have a USB port? Perhaps it's more so
than I realize, but at least in my (admittedly somewhat limited)
experience I've yet to encounter one that does, so I'm wondering how
appropriate these options are for the aspeed-g5 defconfig if they might
just end up as a bunch of code that's never executed on most BMCs.
Zev
>Signed-off-by: Adriana Kobylak <[email protected]>
>Tested-by: Adriana Kobylak <[email protected]>
>---
> arch/arm/configs/aspeed_g5_defconfig | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
>diff --git a/arch/arm/configs/aspeed_g5_defconfig b/arch/arm/configs/aspeed_g5_defconfig
>index bee9422919aa..1b0d82c64ad4 100644
>--- a/arch/arm/configs/aspeed_g5_defconfig
>+++ b/arch/arm/configs/aspeed_g5_defconfig
>@@ -37,11 +37,9 @@ CONFIG_KEXEC=y
> CONFIG_VFP=y
> CONFIG_NEON=y
> CONFIG_KERNEL_MODE_NEON=y
>-CONFIG_FIRMWARE_MEMMAP=y
> CONFIG_KPROBES=y
> CONFIG_JUMP_LABEL=y
> CONFIG_MODULES=y
>-# CONFIG_BLK_DEV_BSG is not set
> # CONFIG_BLK_DEBUG_FS is not set
> # CONFIG_MQ_IOSCHED_DEADLINE is not set
> # CONFIG_MQ_IOSCHED_KYBER is not set
>@@ -98,6 +96,7 @@ CONFIG_NET_NCSI=y
> CONFIG_DEVTMPFS=y
> CONFIG_DEVTMPFS_MOUNT=y
> # CONFIG_PREVENT_FIRMWARE_BUILD is not set
>+CONFIG_FIRMWARE_MEMMAP=y
> CONFIG_MTD=y
> CONFIG_MTD_BLOCK=y
> CONFIG_MTD_PARTITIONED_MASTER=y
>@@ -111,6 +110,8 @@ CONFIG_BLK_DEV_LOOP=y
> CONFIG_BLK_DEV_NBD=y
> CONFIG_EEPROM_AT24=y
> CONFIG_EEPROM_AT25=y
>+CONFIG_SCSI=y
>+CONFIG_BLK_DEV_SD=y
> CONFIG_NETDEVICES=y
> CONFIG_NETCONSOLE=y
> # CONFIG_NET_VENDOR_ALACRITECH is not set
>@@ -230,6 +231,7 @@ CONFIG_USB_DYNAMIC_MINORS=y
> CONFIG_USB_EHCI_HCD=y
> CONFIG_USB_EHCI_ROOT_HUB_TT=y
> CONFIG_USB_EHCI_HCD_PLATFORM=y
>+CONFIG_USB_STORAGE=y
> CONFIG_USB_GADGET=y
> CONFIG_USB_ASPEED_VHUB=y
> CONFIG_USB_CONFIGFS=y
>@@ -277,6 +279,7 @@ CONFIG_FSI_OCC=y
> CONFIG_EXT4_FS=y
> CONFIG_FANOTIFY=y
> CONFIG_OVERLAY_FS=y
>+CONFIG_VFAT_FS=y
> CONFIG_TMPFS=y
> CONFIG_JFFS2_FS=y
> # CONFIG_JFFS2_FS_WRITEBUFFER is not set
>@@ -292,6 +295,8 @@ CONFIG_PSTORE_PMSG=y
> CONFIG_PSTORE_FTRACE=y
> CONFIG_PSTORE_RAM=y
> # CONFIG_NETWORK_FILESYSTEMS is not set
>+CONFIG_NLS_CODEPAGE_437=y
>+CONFIG_NLS_ISO8859_1=y
> CONFIG_HARDENED_USERCOPY=y
> CONFIG_FORTIFY_SOURCE=y
> CONFIG_CRYPTO_HMAC=y
>--
>2.25.1
>
On Sat, 13 Nov 2021 at 01:58, Zev Weiss <[email protected]> wrote:
> Hmm, how common is it for BMCs to have a USB port?
If it helps, the evb-ast2500 has a couple of USB ports, and I've
actually been trying to enable USB (additionally with CONFIG_HIDRAW)
so that I can do a demo that uses fwupd to update the firmware on the
attached keyboard or mouse.
Richard
On Sat, 13 Nov 2021, at 10:59, Zev Weiss wrote:
> On Fri, Nov 12, 2021 at 12:29:31PM PST, Adriana Kobylak wrote:
>>From: Adriana Kobylak <[email protected]>
>>
>>Add support to detect USB flash drives and create the /dev/sd* devices.
>>Also add support for vfat to support USB drives formatted as FAT32.
>>This support will be used to enable firmware updates via USB flash
>>drives where the firmware image is stored in the USB drive and it's
>>plugged into the BMC USB port.
>>
>
> Hmm, how common is it for BMCs to have a USB port? Perhaps it's more so
> than I realize, but at least in my (admittedly somewhat limited)
> experience I've yet to encounter one that does, so I'm wondering how
> appropriate these options are for the aspeed-g5 defconfig if they might
> just end up as a bunch of code that's never executed on most BMCs.
The intent with the Aspeed defconfigs was to enable options that
provide a broad coverage of features used by e.g. upstream OpenBMC
systems. This aids kernel maintenance for OpenBMC. The platform
configuration in OpenBMC should provide a kernel config that only
enables options relevant to that specific platform. With this in mind I
think it's fine that USB support is enabled in the defconfigs.
Andrew