summaryrefslogtreecommitdiff
path: root/meta-asrock
AgeCommit message (Collapse)AuthorFilesLines
2025-09-01meta-asrock: e3c246d4i: Disable DRM to reduce v6.12 binary sizeAndrew Jeffery1-0/+2
Shrinking vmlinux by disabling CONFIG_DRM allows the resulting FIT image to meet the size constraint imposed by its flash partition. CONFIG_DRM and CONFIG_DRM_ASPEED_GFX tend to be used for fbterm and rendering the host console managed by uart-render-controller. From inspection it's only IBM systems doing this, so assuming DRM is unused here, drop it from the kernel config. For comparision, size(1) reports: dev-6.6: text data bss dec hex filename 7102072 2406882 144352 9653306 934c3a vmlinux dev-6.12, CONFIG_DRM=y: text data bss dec hex filename 7472824 2541570 161216 10175610 9b447a vmlinux dev-6.12, CONFIG_DRM=n: text data bss dec hex filename 7085476 2380654 159008 9625138 92de32 vmlinux Suggested-by: Joel Stanley <joel@jms.id.au> Change-Id: Iea294faa302b75b3d8fa55cb68646a9cc2d55a6b Signed-off-by: Andrew Jeffery <andrew@codeconstruct.com.au>
2025-07-23meta-asrock: add x86-power-control configuration for X570D4URenze Nicolai2-0/+104
Change-Id: I3a369bf0eca1be49c362553fd4ad41d6f354e2b8 Signed-off-by: Renze Nicolai <renze@rnplus.nl>
2025-07-23meta-asrock:e3c246d4i:kernel: shrink image sizeJayanth Othayoth1-0/+4
During the migration to Linux kernel 6.12, the e3c246d4i platform began encountering flash image size overflows due to increased kernel size. ''' ERROR: obmc-phosphor-image-1.0-r0 do_generate_static: Image '/target/e3c246d4i/build/deploy/images/e3c246d4i/fitImage-obmc-phosphor-initramfs-e3c246d4i-e3c246d4i' is 52292 bytes too large! ''' To address this, the following debug options have been disabled: - CONFIG_FUNCTION_TRACER: Enables function-level instrumentation for tracing via ftrace. Disabling it significantly reduces kernel image size. Side effect: Removes support for function call tracing and performance profiling using ftrace. - CONFIG_DEBUG_LIST: Adds runtime checks for kernel linked list misuse. Disabling it reduces code size and runtime overhead. Side effect: Disables automatic detection of list corruption (e.g., double insertions or invalid deletions), which may make certain memory bugs harder to detect. These changes help reduce the kernel image size while retaining essential debugging capabilities like dynamic debug and symbol-based crash analysis. Change-Id: I26580a8a0cebe2076877090204b13275f9c0540e Signed-off-by: Jayanth Othayoth <ojayanth@gmail.com>
2025-07-23treewide: add whinlatter LAYERSERIES_COMPATPatrick Williams6-6/+6
Change-Id: I1d53510329749ff122c8a65157a72c8ff88bd306 Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
2025-05-24treewide: add walnascar LAYERSERIES_COMPATPatrick Williams6-6/+6
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I6291eeaf0ca7d4bd674bf0717f1234bfd757f6fe
2025-05-12dbus-sensors: enable mctpreactor only for mctp systemsPatrick Williams1-1/+0
We already have a distro-feature for MCTP support. Leverage this so that mctpreactor is only installed by default for systems which intend to utilize MCTP. Tested: Build yosemite4 and confirmed mctpreactor is still included. Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I5b61bbc1905d384adb442cc0aedbd52f8aebaf76
2025-03-29meta-asrock: e3c246d4i: Trim back dbus-sensors PACKAGECONFIGZev Weiss1-0/+9
e3c246d4i has started running out of space for its squashfs volume; let's clear up some space by removing some unneeded dbus-sensors daemons. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ib31140c8ef42054856e4b2b0b0afb9dab6e081a4
2024-12-20treewide: switch debug-tweaks to allow-root-loginPatrick Williams5-5/+5
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I2665e9e7e37242d546eafdb4d1c65dd0369ca635
2024-12-20treewide: add styhead to LAYERSERIES_COMPATPatrick Williams6-6/+6
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I8f74fcbb7bf66bb24fde1571ef09fac1095c7e59
2024-12-06styhead: update WORKDIR usageAndrew Geissler10-12/+12
The latest yocto release(styhead,5.1) has made a significant change[1] to the use of the WORKDIR parameter. This is not the complete change required but gets us going in the right direction. [1]: https://docs.yoctoproject.org/next/migration-guides/migration-5.1.html#workdir-changes Signed-off-by: Andrew Geissler <geissonator@yahoo.com> Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I4f184b361be69e7704a0657a4a00cdad981aff8a
2024-08-28update configurations for phosphor-led-managerAlexander Hansen4-0/+8
phosphor-led-manager no longer assumes any default led priority. The behavior was changed in https://github.com/openbmc/phosphor-led-manager/commit/5a9f651949c133ed0868d72734c4ac82c75a7656 and https://github.com/openbmc/phosphor-led-manager/commit/55badf7979f57d5f9ed2db726eebf35e84d0dac5 The led priority needs to be configured, so updating all configurations here. Before, led-manager assumed a default led priority of 'Blink'. So this patch is just making that explicit in the existing configurations. There should be no behavior change from this patch. Change-Id: I74aabb567ccf6682d56877f8bb42e636fc1746f6 Signed-off-by: Alexander Hansen <alexander.hansen@9elements.com>
2024-04-12meta-asrock: romed8hm3: Correct phosphor-led-manager bbappend filenameZev Weiss1-0/+0
Typo in the file name meant it wasn't getting applied. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I112be527fc90c2f85a531d5f0dc6bda1b94cb2e5
2024-04-12meta-asrock: Remove redundant phosphor-ipmi-kcs dependenciesZev Weiss3-3/+0
We've now got this specified in the platform-indepdent config in the meta-asrock layer, so there's no need to repeat it in each platform's packagegroup-asrock-apps. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I2c79abd77d2b936ae57efce74f7fc102d83dd452
2024-04-12meta-asrock: Deduplicate common platform configurationZev Weiss6-55/+23
All existing ASRock platforms use most of the same core packages (entity-manager, dbus-sensors, x86-power-control, phosphor-ipmi-kcs...), so we might as well define it all in one place instead of duplicating it for each machine. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ife82b12cf909de6bdd770943adb30327a22438e1
2024-04-06meta-asrock: initial skeleton of X570D4U supportRenze Nicolai8-0/+105
This is a Micro-ATX single-socket AMD X570 server board with an AST2500 BMC. Change-Id: If9c0cd2c0abfd6d106cdc37e144d71fb72b7cb56 Signed-off-by: Renze Nicolai <renze@rnplus.nl>
2024-03-25meta-asrock: e3c256d4i: Add bios-update config parametersZev Weiss2-0/+12
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I0940ff15a1880ececf3cfac9196f5df8f4f15ab7
2024-03-25meta-asrock: spc621d8hm3: Add bios-update config parametersZev Weiss2-0/+12
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I84f602039d16209bc8b579dd201fd85c49624bf3
2024-03-25meta-asrock: romed8hm3: Add bios-update config parametersZev Weiss3-0/+10
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I4d532d70e35a2944fcc74a52d22e6d6fec02aba1
2024-03-25meta-asrock: e3c246d4i: Add bios-update config parametersZev Weiss3-0/+13
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I66efb86b10401cba4c8a26b80e2b6aed4e6768b6
2024-03-25meta-asrock: Add bios-update infrastructureZev Weiss3-0/+230
On all supported ASRock platforms the BIOS (host firmware) update mechanism involves setting some GPIOs, binding a driver for an MTD flash device, writing the image to the flash, and then undoing the setup steps in reverse order. The exact GPIOs involved (and some parameters of the firmware image) vary a bit between systems; the platform-specific differences are distilled out into some config parameters loaded from /etc/default/bios-update. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Co-developed-by: Olivier Faurax <olivier.faurax@eu.equinix.com> Change-Id: I4ecf93a73359063e3639895b078057a637dff452
2024-03-22treewide: remove meta-poky usagePatrick Williams4-4/+0
Upstream poky made a commit that adds an `/etc/motd` with a strong warning that meta-poky is not intended to be used for production purposes. There isn't anything we use directly from meta-poky anymore, so clean up all references to it. See poky commit a226865c8683398b5f58628ba2ec5aee1ee6c19d for additional context. We could potentially switch from using the `poky` subtree to picking up OECore directly, but right now it also provides us a qualified set of OECore + bitbake. Tested: Built yosemite4 and confirmed `motd` output is absent when SSHing in. Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I9c08650d2278b3ba26794c009adad3e593767f58
2024-03-18meta-asrock: e3c256d4i: Work around hardware bug with NMI signalsZev Weiss1-3/+12
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I507f963b15106b686b04275b13137f1036afcd03
2024-02-09meta-asrock: romed8hm3: Add LED configZev Weiss2-0/+33
Enable the heartbeat & system-fault LEDs. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Id5f4952a61823f17bf3db764b092ef885f01caf2
2024-02-09meta-asrock: e3c246d4i: Add LED configZev Weiss2-0/+33
Enable the heartbeat & system-fault LEDs. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Id2507a05dbc8a17cd415a1a863a0b40319e43bad
2024-02-09meta-asrock: e3c246d4i: Add kernel config for PECI supportZev Weiss2-0/+10
We've had it enabled in the device-tree for a while now, but didn't have the driver support to actually make use of it configured; let's fix that. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I3104509a4e18de948b9f671d3feb0942f9efb0e3
2024-02-09meta-asrock: e3c246d4i: Enable MMC/SDHCI reset quirkZev Weiss1-0/+2
It seems e3c246d4i suffers from the same missing pull-up on the reset line that also affects e3c256d4i, so enable the quirk override to disable MMC/SDHCI driver support in u-boot as a workaround. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I411f3857110f6c7e9fefccfb330742daf417138c
2024-02-09meta-asrock: Generalize MMC/SDHCI reset quirkZev Weiss6-6/+12
As it turns out, e3c256d4i isn't the only board with this problem, so move the u-boot config fragment to meta-common and make it available via an override. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I7182a6a1dd12603fec1d0633edc60d018285914f
2024-01-23meta-asrock: Add common SOL settings overrideZev Weiss2-0/+9
Enable SOL on all ASRock platforms. The defaults for the xyz.openbmc_project.Ipmi.SOL interface are sufficient, but the /xyz/openbmc_project/ipmi/sol/eth0 settings object doesn't unless explicitly instantiated. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I176444f9e8f608667de514d450dca18ce46e2c42
2024-01-20meta-asrock: e3c256d4i: Add LED configZev Weiss2-0/+33
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I4125773f51c8318555577206c09ca1bdb99e626e
2024-01-20meta-asrock: e3c256d4i: Add IPMI dev_id.jsonZev Weiss2-0/+8
This replicates the values (ASRock manufacturer ID and such) returned by 'ipmitool mc info' on the factory BMC firmware. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I580281b91ee0b2cd00bf5e13f8c9497adf362c9e
2024-01-20meta-asrock: e3c256d4i: Add obmc-console configZev Weiss3-0/+12
While on all other currently supported ASRock platforms the Aspeed VUART provides the host serial console, we haven't had any success using it over eSPI on e3c256d4i, so instead we're using a pair of SUARTs connected via the UART mux. This unforunately requires leaving the AST2500's built-in Super-IO enabled, because the host firmware needs to access it shortly after power-on in order to configure its UART output, so we also have to open up some of the u-boot backdoor config options too. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ibdab607d504cebd956a7e23191c09cd0d5c44e50
2024-01-20meta-asrock: e3c256d4i: Enable and configure x86-power-controlZev Weiss3-0/+100
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ic6c5460f587ae141ed69e704e359152713171097
2024-01-20meta-asrock: Initial skeleton of e3c256d4i platform supportZev Weiss11-0/+113
This is a single-socket mini-ITX Xeon board with an AST2500 BMC, roughly an updated version of the e3c246d4i. Note though that because this platform uses eSPI instead of LPC, it currently requires some driver support that's not yet in the OpenBMC kernel to actually boot the host. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ib3a13e0ac9f4130c7965ce63f804eb169262fcd3
2024-01-13meta-asrock: spc621d8hm3: Add phosphor-led-manager configZev Weiss2-0/+33
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I8d21ddbe3d43c63576377e934739be3cec96268f
2024-01-13meta-asrock: spc621d8hm3: Add IPMI dev_id.jsonZev Weiss2-0/+8
This replicates the values returned by 'ipmitool mc info' on the factory (AMI) BMC firmware. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I1b4b76ecf5a69f11fe59a0dbf2d2522dac911cc8
2024-01-13meta-asrock: spc621d8hm3: Add obmc-console configZev Weiss2-0/+4
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I6242d255eb49b995c56309a6b3398c72ce87a8c7
2024-01-13meta-asrock: spc621d8hm3: Enable and configure x86-power-controlZev Weiss3-0/+100
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I1195e18cbe1957702fe31ab5eb2b8a14f71bbad2
2024-01-13meta-asrock: initial skeleton of spc621d8hm3 supportZev Weiss7-0/+103
This is a half-width single-socket Xeon server board with an AST2500 BMC. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I3362a992997d02d43f409e1df00c9f8098d05c5b
2023-11-25treewide: add scarthgap to LAYERSERIES_COMPAT2.16.0-devPatrick Williams3-3/+3
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I28ee50fa55ae47dd7fd8c99f8f6db8f5f6dfa53d
2023-11-24treewide: add nanbield to LAYERSERIES_COMPATPatrick Williams3-3/+3
Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I85272779478b66452acd11be93f5fea99e4c3a34
2023-10-11treewide: clean up bmcweb in package groupsPatrick Williams2-2/+0
bmcweb is already added as a default in the meta-phosphor package groups. A large number of machines have mistakenly also explicitly added a RDEPEND on assorted package groups, which is not necessary. Clean these up. Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: Ifd3726d18aab45475f80d054a4640196ac0b71d2
2023-10-11treewide: clean up webui selectionPatrick Williams2-2/+0
The webui-vue package is now enabled by default. Clean up all the meta layers to clean up the explicit enablement. Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Change-Id: I4895be56c70d2c8666dd96cec18d2c097f0c05a3
2023-09-26linux-aspeed: Move to Linux v6.5Joel Stanley2-2/+2
This moves the OpenBMC kernel to a v6.5 base for ASPEED. The 6.5 kernel moved all 32-bit ARM device trees, so our BMC device tress are now under an 'aspeed' subdirectory. This change means systems must describe the KERNEL_DEVICETREE with the aspeed directory prefix. There are 78 patches in the tree, with 40 of those patches not yet queued for merging in v6.6. The remaining out of tree patches include: Andrew Jeffery (6): dt-bindings: hwmon: pmbus: Add Maxim MAX31785 documentation pmbus (max31785): Add support for devicetree configuration pmbus (core): One-shot retries for failure to set page pmbus (max31785): Wrap all I2C accessors in one-shot failure handlers ARM: dts: aspeed: witherspoon: Update max31785 node ipmi: kcs_bmc: Add a "raw" character device interface Cédric Le Goater (1): /dev/mem: add a devmem kernel parameter to activate the device Eddie James (16): ARM: dts: aspeed: bonnell: Add reserved memory for TPM event log dt-bindings: soc: Add Aspeed XDMA Engine soc: aspeed: Add XDMA Engine Driver soc: aspeed: xdma: Add user interface soc: aspeed: xdma: Add reset ioctl soc: aspeed: xdma: Add trace events i2c: core: Add mux root adapter operations iio: si7020: Lock root adapter to wait for reset eeprom: ee1004: Enable devices on multiple busses dt-bindings: trivial-devices: Add Atmel AT30TSE004A serial eeprom eeprom: ee1004: Add OF matching support leds: pca955x: Refactor with helper functions and renaming leds: pca955x: Use pointers to driver data rather than I2C client leds: pca955x: Optimize probe led selection leds: pca955x: Add HW blink support leds: Ensure hardware blinking turns off when requested Jae Hyun Yoo (1): clk: ast2600: enable BCLK for PCI/PCIe bus always Joel Stanley (14): net: ftgmac100: Ensure tx descriptor updates are visible ARM: aspeed: Add debugfs directory ARM: soc: aspeed: Add secure boot controller support dt-bindings: trivial-devices: Remove Infineon SLB9673 TPM ARM: dts: nuvoton: npmc750-evb: Add default console tpm: tis-i2c: Add more compatible strings leds: pca955x: Revert "Remove the unused function pca95xx_num_led_regs()" arm64: configs: Add Nuvoton NPCM defconfig ARM: configs: aspeed: Add new FSI drivers ARM: config: aspeed_g5: Enable SSIF BMC driver ARM: config: aspeed: Remove FIRMWARE_MEMMAP ARM: config: aspeed: Add Ampere SMPro drivers ARM: config: Add openbmc defconfig ARM: config: openbmc: Add HPE GPX and Nuvoton 7xx Johannes Holland (1): dt-bindings: tpm: Add schema for TIS I2C devices Potin Lai (1): mtd: spi-nor: winbond: Add support for w25q01jvq Change-Id: Ib97af192391af6e71c96fe14e12cfc88c23d7a7d Signed-off-by: Joel Stanley <joel@jms.id.au>
2023-05-12Revert "Add socket-id for the first console"Andrew Jeffery2-2/+0
This reverts commit ec7cab9378f548e94ee4af83946b2583d1fc7194. ec7cab9378f5 ("Add socket-id for the first console") was only a partial fixup of the obmc-console configurations found in the tree. For historical reasons not all platforms that support multiple consoles install a client configuration file for all consoles. Instead they relied on some default behaviour that was removed in obmc-console's 4e7186918599 ("Fixed broken dbus interface for multiple consoles"). Applying the configuration changes in the manner of ec7cab9378f5 ("Add socket-id for the first console") resulted in an asymmetric configuration between the client and server, breaking some SSH SOL instances. As of ae2460d0b8e8 ("obmc-console: Provide a default value for `console-id`.") in obmc-console the requirement to specify `socket-id`[^1] is lifted. Instead, the configuration can choose to override a default value. This restores the original behaviour and unbreaks SSH SOL. [^1]: Now an alias of the `console-id` configuration key Websocket and IPMI consoles remain broken. A fix is under development that also takes a proper approach to supporting multiple consoles in bmcweb, and the same solution pattern can be applied in ipmid. More discussion of the problems involved can be found here: https://amboar.github.io/notes/2023/05/08/happenings-in-obmc-console.html Change-Id: I274284c791758f336da6d6301dc523bac2b5dd69 Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
2023-04-21Add socket-id for the first consoleNinad Palsule2-0/+2
This drop adds socket-id for first console. This was left out before but with the new design this parameter is required and enforced by obmc-console code. It is decided to set "socket-id = console0" initially because that is the current path used by bmcweb but in future you can change this value to whatever is appropriate. The console information is now available through the DBUS interface. Each console server register it's object path on DBUS. The leaf node of the object path is used to create a console path. For example in the following console0 the leaf node "console0" is used to make up the console path and the GUI will use console path "/console/console0" to connect to the console. busctl tree xyz.openbmc_project.Console.console0 `-/xyz `-/xyz/openbmc_project `-/xyz/openbmc_project/console `-/xyz/openbmc_project/console/console0 Here is a discord thread where new design was discussed. https://discord.com/channels/775381525260664832/1083551792094249051 Tested: Tested on the rainier system. Related commits: 1) phosphor-dbus-interface: https://gerrit.openbmc.org/c/openbmc/phosphor-dbus-interfaces/+/61486 2) obmc-console: https://gerrit.openbmc.org/c/openbmc/obmc-console/+/62496 3) bmcweb: https://gerrit.openbmc.org/c/openbmc/bmcweb/+/62525 Change-Id: I5fd06f3aac69872a3e5032d592a9b59711795cfe Signed-off-by: Ninad Palsule <ninadpalsule@us.ibm.com>
2023-01-26romed8hm3: Adjust kernel config for nct6779 supportZev Weiss2-0/+9
On the romed8hm3 board (as on some other ASRock AMD platforms) the host CPU's TSI interface is connected to a super-io chip instead of being run directly to the BMC, so in order to retrieve CPU temperature readings we have to go through the super-io chip's back-door i2c interface. This enables the driver support we need to do so. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: I5db6fb9e1a290f42af9561c2eb889ac00cddd104
2023-01-25e3c246d4i: Add phosphor-ipmi-configZev Weiss2-0/+9
This brings the manufacturer & product info returned by 'ipmitool mc info' into line with what the factory (AMI) firmware returns. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ia7eefefe2776f8b5693328698d969c638e9cff2b
2023-01-19romed8hm3: Add phosphor-ipmi-configZev Weiss2-0/+9
This brings the manufacturer & product info returned by 'ipmitool mc info' into line with what the factory (AMI) firmware returns. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ib9e2d1f7e6286083d4697ce5e4b539ca41b12be9
2023-01-19romed8hm3: Add obmc-console configZev Weiss2-0/+4
Like other ASRock Rack systems, romed8hm3 has the host console at 0x2f8 using IRQ 3, running at 115200 baud. Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ib05faf98329ae275ae36f57e3c413b1ffbf7839e
2023-01-19romed8hm3: Enable x86-power-control, add configZev Weiss3-0/+100
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Change-Id: Ifa6ea971eec3ceef9c2ba4d8ec8b1ef922d21d61
generated by cgit v1.2.3 (git 2.39.0) at 2026-09-08 18:28:33 +0300