(IDF 5.0.2) PSRAM detected but not available.

[MRIIOT]

I just noticed that even though PSRAM is detected, available ram to MP is not reflecting this. And this was discovered because we ran out of RAM crashing the ESP so I don't believe gc.mem_free is incorrect. I'm going to look through commits ahead and see if this was discovered as an issue but this is where I am now.

Free FLASH 11296.0KB, RAM 37.59375KB
GC: total: 64000, used: 25648, free: 38352, max new split: 4063232
 No. of 1-blocks: 639, 2-blocks: 68, max blk sz: 64, max free sz: 1631
qstr pool: n_pool=2, n_qstr=27, n_str_data_bytes=354, n_total_bytes=834

ESP-IDF commit: 5181de8ac5ec5e18f04f634da8ce173b7ef5ab73 espressif/esp-idf@5181de8
MP commit: 1dedb65e645f72afbd614431435208d687e16992 1dedb65

I (28) boot: ESP-IDF v5.0.2 2nd stage bootloader
I (28) boot: compile time 12:13:29
I (28) boot: chip revision: v1.1
I (31) boot.esp32: SPI Speed      : 40MHz
I (35) boot.esp32: SPI Mode       : DIO
I (40) boot.esp32: SPI Flash Size : 16MB
I (45) boot: Enabling RNG early entropy source...
I (50) boot: Partition Table:
I (54) boot: ## Label            Usage          Type ST Offset   Length
I (61) boot:  0 nvs              WiFi data        01 02 00009000 00004000
I (68) boot:  1 otadata          OTA data         01 00 0000d000 00002000
I (76) boot:  2 phy_init         RF data          01 01 0000f000 00001000
I (83) boot:  3 ota_0            OTA app          00 10 00010000 00270000
I (91) boot:  4 ota_1            OTA app          00 11 00280000 00270000
I (98) boot:  5 vfs              Unknown data     01 81 004f0000 00b10000
I (106) boot: End of partition table
I (110) esp_image: segment 0: paddr=00010020 vaddr=3f400020 size=53294h (340628) map
I (241) esp_image: segment 1: paddr=000632bc vaddr=3ff80063 size=00008h (     8) load
I (242) esp_image: segment 2: paddr=000632cc vaddr=3ffbdb60 size=05298h ( 21144) load
I (256) esp_image: segment 3: paddr=0006856c vaddr=40080000 size=07aach ( 31404) load
I (269) esp_image: segment 4: paddr=00070020 vaddr=400d0020 size=119758h (1152856) map
I (685) esp_image: segment 5: paddr=00189780 vaddr=40087aac size=18044h ( 98372) load
I (726) esp_image: segment 6: paddr=001a17cc vaddr=400c0000 size=00064h (   100) load
I (726) esp_image: segment 7: paddr=001a1838 vaddr=500007f8 size=00808h (  2056) load
I (749) boot: Loaded app from partition at offset 0x10000
I (749) boot: Disabling RNG early entropy source...
I (761) quad_psram: This chip is ESP32-D0WD
I (761) esp_psram: Found 8MB PSRAM device
I (761) esp_psram: Speed: 40MHz
I (765) esp_psram: PSRAM initialized, cache is in low/high (2-core) mode.
W (773) esp_psram: Virtual address not enough for PSRAM, map as much as we can. 4MB is mapped
I (782) cpu_start: Pro cpu up.
I (785) cpu_start: Starting app cpu, entry point is 0x40081400
I (0) cpu_start: App cpu up.
I (1678) esp_psram: SPI SRAM memory test OK
I (1687) cpu_start: Pro cpu start user code
I (1687) cpu_start: cpu freq: 240000000 Hz
I (1687) cpu_start: Application information:
I (1690) cpu_start: Compile time:     Oct 23 2023 12:13:15
I (1696) cpu_start: ELF file SHA256:  53f6065e4a06f288...
ELF file not found. You need to build & flash the project before running 'monitor', and the binary on the device must match the one in the build directory exactly. 
I (1702) cpu_start: ESP-IDF:          v5.0.2
I (1707) cpu_start: Min chip rev:     v1.0
I (1712) cpu_start: Max chip rev:     v3.99 
I (1717) cpu_start: Chip rev:         v1.1
I (1722) heap_init: Initializing. RAM available for dynamic allocation:
I (1729) heap_init: At 3FFAFF10 len 000000F0 (0 KiB): DRAM
I (1735) heap_init: At 3FFB6388 len 00001C78 (7 KiB): DRAM
I (1741) heap_init: At 3FFB9A20 len 00004108 (16 KiB): DRAM
I (1747) heap_init: At 3FFCA8B0 len 00015750 (85 KiB): DRAM
I (1754) heap_init: At 3FFE0440 len 00003AE0 (14 KiB): D/IRAM
I (1760) heap_init: At 3FFE4350 len 0001BCB0 (111 KiB): D/IRAM
I (1767) heap_init: At 4009FAF0 len 00000510 (1 KiB): IRAM
I (1773) esp_psram: Adding pool of 4096K of PSRAM memory to heap allocator
I (1782) spi_flash: detected chip: generic
I (1785) spi_flash: flash io: dio
W (1790) i2s(legacy): legacy i2s driver is deprecated, please migrate to use driver/i2s_std.h, driver/i2s_pdm.h or driver/i2s_tdm.h
W (1802) rmt(legacy): legacy driver is deprecated, please migrate to `driver/rmt_tx.h` and/or `driver/rmt_rx.h`
W (1812) ADC: legacy driver is deprecated, please migrate to `esp_adc/adc_oneshot.h`
I (1822) cpu_start: Starting scheduler on PRO CPU.
I (0) cpu_start: Starting scheduler on APP CPU.
I (1832) esp_psram: Reserving pool of 32K of internal memory for DMA/internal allocations

vfs = os.statvfs('/')
print('Free FLASH {}KB, RAM {}KB'.format((vfs[0]*vfs[3])/1024, gc.mem_free()/1024))
micropython.mem_info()
micropython.qstr_info()

[robert-hh]

GC: total: 64000, used: 25648, free: 38352, max new split: 4063232

The information to mem_info is somewhat confusing since a while, But the max new split figure tells, that 4 MB of RAM is available, You can try so by allocating a few large bytearrays, e.g. with a size of 1 MByte. That should work.

[MRIIOT]

I see what you mean. Will try that.

[MRIIOT]

You are right allocation works. So our NimBLE GATT Notify ENOMEM is unrelated.