Amiga freezes when DATASIZE>2558 (suprascsi.device)

[rjnorthrow]

Hardware/Firmware

• Amiga 500+ stock configuration with 2MiB CHIP, OS 2.04
• SupraDrive 500XP with 8MiB FAST, bus-powered from the Amiga
• BlueSCSI Pico 2W firmware BlueSCSI_V2_Universal_2026-08-17_4d0f96f.uf2

Software

• daynaport-amiga 1.5
• Roadshow 1.15

Findings

• SupraDrive 500XP / suprascsi.device works in AmigaNET mode with Vendor=AmigaNET/Product=SCSI/Link.
• Legacy mode doesn't work on this firmware (MAC address fetch fails).
• DATASIZE must be ≤ 2558 (2559+ hangs the machine) — consistent with polled, non-DMA I/O.
• All three MODE values function once DATASIZE is capped; occasional benign orphaned-ARP warnings under load.
• Sometimes stops responding once connected for a few minutes (see log)

Config

I have created an empty file in the root of the SDCard called AM6.hda

bluescsi.ini

[SCSI]
Debug=0
System="Generic"
EnableSCSI2=0
PrefetchBytes=0
WiFiSSID=[ssid]
WiFiPassword=[password]
[SCSI0]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI1]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI2]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI3]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI4]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI5]
Vendor=Blue
Product=Bluedrive
Version=1.0
SectorsPerTrack=32
HeadsPerCylinder=1
[SCSI6]
Vendor=AmigaNET
Product=SCSI/Link
Version=1.0

ENVARC:scsidayna.prefs

DEVICE=suprascsi.device
DEVICEID=-1
PRIORITY=0
MODE=2
AUTOCONNECT=0
SSID=
KEY=
DATASIZE=2048
DEBUG=1
AGGRESSIVE=0

Log

Platform: BlueSCSI (Pico2/W)
FW Version: 2026.08.17-rel Aug 17 2026 16:08:16
Flash chip size: 4096 kB
I2C Supported
=== Bootloader Log ===
[0ms] Bootloader version: Aug 17 2026 16:08:02 BlueSCSI
[0ms] Looking for firmware: BlueSCSI_Pico_2_DaynaPORT*.bin
[196ms] Bootloader continuing to main firmware
=== End Bootloader Log ===
Reclock Pico 2/2W based boards to standardized speed
Initial Clock set to 150MHz
Reclocking the MCU to 155MHz
Setting the SDIO clock to 26MHz
After reclocking, system reports clock set to 155MHz
 
=== SD Card Info ===
SD card detected, exFAT volume size: 60923 MB
SD Name: SDU1 , MID: 0x74 (Transcend), OID: "J`" (0x4A 0x60)
-- EnableSCSI2 = No
-- PrefetchBytes = 0

=== Global Config ===
Reading configuration from bluescsi.ini
-- EnableSCSI2 = No
-- PrefetchBytes = 0
Using system preset "Generic"
-- WiFiSSID = [wifi ssid]
-- WiFiPassword = [set]

=== Finding images ===
[SCSI0]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI1]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI2]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI3]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI4]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI5]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
[SCSI6]
-- Vendor = AmigaNET
-- Product = SCSI/Link
=== Finding images in / ===
[SCSI3]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD3.hda for ID:3 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled
[SCSI5]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD5.hda for ID:5 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled
[SCSI4]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD4.hda for ID:4 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled
[SCSI6]
-- Version = 1.0
== Opening /AM6.hda for ID:6 LUN:0
---- Configuring as network based on image name
[SCSI1]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD1 - OCS ECS Games A-M.hda for ID:1 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled
[SCSI2]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD2 - OCS ECS Games N-Z.hda for ID:2 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled
[SCSI0]
-- SectorsPerTrack = 32
-- HeadsPerCylinder = 1
-- Vendor = Blue
-- Product = Bluedrive
== Opening /HD0 - Roger new.hda for ID:0 LUN:0
---- Configuring as disk drive
---- Read prefetch disabled

== Platform supports ROM drive up to 2848 kB
 
=== Configured SCSI Devices ===
ID: 0, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 1, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 2, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 3, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 4, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 5, BlockSize: 512, Type: Fixed, Quirks: None, Size: 1048576kB
ID: 6, Type: Network
 
=== Network Initialization ===
Wi-Fi MAC: 00:80:19:3A:F9:70
Connecting to Wi-Fi SSID "BrightonFibre_2DAF" with WPA/WPA2 PSK

Clock set to: 155MHz
Initialization complete!
Disabling status LED
INFO: Pico Voltage: 3.295V.
Wi-Fi connection to BrightonFibre_2DAF failed
Attempting Wi-Fi reconnection to "BrightonFibre_2DAF" (attempt 1/5)
Successfully connected to Wi-Fi SSID "BrightonFibre_2DAF"
--------------
WATCHDOG TIMEOUT, attempting bus reset
Platform: BlueSCSI
FW Version: 2026.08.17-rel Aug 17 2026 16:08:16
GPIO states: out 0x16B3FEFF oe 0x0ADC8FFF in 0x17B3FAFF
scsiDev.cdb: 0x08 0x00 0x40 0x0C 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 
scsiDev.phase: 6
SCSI DMA state: IDLE
Current buffer: 0x00000000/0x00000000, next 0x00000000 bytes
SyncOffset: 0 SyncPeriod 0
PIO Parity SM: tx_fifo 0, rx_fifo 0, pc 28, instr 0x000080A0
PIO Data SM: tx_fifo 0, rx_fifo 0, pc 23, instr 0x000090E0
PIO Sync SM: tx_fifo 0, rx_fifo 0, pc 0, instr 0x00000000
DMA CH A: ctrl: 0x0002C011 count: 0x00000000
DMA CH B: ctrl: 0x000AE009 count: 0x00000000
DMA CH C: ctrl: 0x0004E005 count: 0x00000000
DMA CH D: ctrl: 0x00052009 count: 0x00000000
GPIO states: 0x17B3FAFF
STACK 0x20081F08:    0x20081F88 0x00000000 0x00000000 0x20052ACC
STACK 0x20081F18:    0x2004CB4C 0x0000000C 0x5DA00000 0x2004C948
STACK 0x20081F28:    0x5DC00000 0x04F54710 0x400E0014 0x43280035
STACK 0x20081F38:    0x00000000 0x10000000 0x62707361 0xFFFFFFE9
STACK 0x20081F48:    0x00080000 0x40028000 0xD0000000 0x11DADAEE
STACK 0x20081F58:    0x00000000 0x2001BCED 0x2001BD84 0x29100000
STACK 0x20081F68:    0x00000000 0x00000000 0x00000000 0x00000000
STACK 0x20081F78:    0x00000000 0x00000000 0x00000000 0x00000000
--------------
WATCHDOG TIMEOUT, attempting bus reset
Platform: BlueSCSI
FW Version: 2026.08.17-rel Aug 17 2026 16:08:16
GPIO states: out 0x16B3FEFF oe 0x0ADC8FFF in 0x17B3FAFF
scsiDev.cdb: 0x08 0x00 0x40 0x0A 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 
scsiDev.phase: 6
SCSI DMA state: IDLE
Current buffer: 0x00000000/0x00000000, next 0x00000000 bytes
SyncOffset: 0 SyncPeriod 0
PIO Parity SM: tx_fifo 0, rx_fifo 0, pc 28, instr 0x000080A0
PIO Data SM: tx_fifo 0, rx_fifo 0, pc 23, instr 0x000090E0
PIO Sync SM: tx_fifo 0, rx_fifo 0, pc 0, instr 0x00000000
DMA CH A: ctrl: 0x0002C011 count: 0x00000000
DMA CH B: ctrl: 0x000AE009 count: 0x00000000
DMA CH C: ctrl: 0x0004E005 count: 0x00000000
DMA CH D: ctrl: 0x00052009 count: 0x00000000
GPIO states: 0x17B3FAFF
STACK 0x20081F08:    0x20081F88 0x00000000 0x00000000 0x20052ACC
STACK 0x20081F18:    0x2004CB4C 0x0000000C 0x5DA00000 0x2004C948
STACK 0x20081F28:    0x5DC00000 0x04F54710 0x400E0014 0x43280035
STACK 0x20081F38:    0x00000000 0x10000000 0x62707361 0xFFFFFFE9
STACK 0x20081F48:    0x00080000 0x40028000 0xD0000000 0xF67F5FE0
STACK 0x20081F58:    0x00000000 0x2001BCED 0x2001BD4A 0x89100000
STACK 0x20081F68:    0x00000000 0x00000000 0x00000000 0x00000000
STACK 0x20081F78:    0x00000000 0x00000000 0x00000000 0x00000000
--------------
WATCHDOG TIMEOUT, attempting bus reset
Platform: BlueSCSI
FW Version: 2026.08.17-rel Aug 17 2026 16:08:16
GPIO states: out 0x16B3FEFF oe 0x0ADC8FFF in 0x17B3FAFF
scsiDev.cdb: 0x1C 0x08 0xE9 0x0A 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 
scsiDev.phase: 6
SCSI DMA state: IDLE
Current buffer: 0x00000000/0x00000000, next 0x00000000 bytes
SyncOffset: 0 SyncPeriod 0
PIO Parity SM: tx_fifo 0, rx_fifo 0, pc 28, instr 0x000080A0
PIO Data SM: tx_fifo 0, rx_fifo 0, pc 23, instr 0x000090E0
PIO Sync SM: tx_fifo 0, rx_fifo 0, pc 0, instr 0x00000000
DMA CH A: ctrl: 0x0002C011 count: 0x00000000
DMA CH B: ctrl: 0x000AE009 count: 0x00000000
DMA CH C: ctrl: 0x0004E005 count: 0x00000000
DMA CH D: ctrl: 0x00052009 count: 0x00000000
GPIO states: 0x17B3FAFF
STACK 0x20081F08:    0x20081F88 0x00000000 0x00000000 0x20052ACC
STACK 0x20081F18:    0x2004CB4C 0x0000000C 0x5DA00000 0x2004C948
STACK 0x20081F28:    0x5DC00000 0x04F54710 0x400E0014 0x43280035
STACK 0x20081F38:    0x00000000 0x10000000 0x62707361 0xFFFFFFE9
STACK 0x20081F48:    0x00080000 0x40028000 0xD0000000 0x17B3FAFF
STACK 0x20081F58:    0x00000000 0x2001BCED 0x2001BD46 0x69100000
STACK 0x20081F68:    0x00000000 0x00000000 0x00000000 0x00000000
STACK 0x20081F78:    0x00000000 0x00000000 0x00000000 0x00000000
--------------
WATCHDOG TIMEOUT, attempting bus reset
Platform: BlueSCSI
FW Version: 2026.08.17-rel Aug 17 2026 16:08:16
GPIO states: out 0x16B3FEFF oe 0x0ADC8FFF in 0x17B3FAFF
scsiDev.cdb: 0x1C 0x08 0xE9 0x0A 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 
scsiDev.phase: 6
SCSI DMA state: IDLE
Current buffer: 0x00000000/0x00000000, next 0x00000000 bytes
SyncOffset: 0 SyncPeriod 0
PIO Parity SM: tx_fifo 0, rx_fifo 0, pc 28, instr 0x000080A0
PIO Data SM: tx_fifo 0, rx_fifo 0, pc 23, instr 0x000090E0
PIO Sync SM: tx_fifo 0, rx_fifo 0, pc 0, instr 0x00000000
DMA CH A: ctrl: 0x0002C011 count: 0x00000000
DMA CH B: ctrl: 0x000AE009 count: 0x00000000
DMA CH C: ctrl: 0x0004E005 count: 0x00000000
DMA CH D: ctrl: 0x00052009 count: 0x00000000
GPIO states: 0x17B3FAFF
STACK 0x20081F08:    0x20081F88 0x00000000 0x00000000 0x20052ACC
STACK 0x20081F18:    0x2004CB4C 0x0000000C 0x5DA00000 0x2004C948
STACK 0x20081F28:    0x5DC00000 0x04F54710 0x400E0014 0x43280035
STACK 0x20081F38:    0x00000000 0x10000000 0x62707361 0xFFFFFFE9
STACK 0x20081F48:    0x00080000 0x40028000 0xD0000000 0x00000000
STACK 0x20081F58:    0x00000000 0x2001BCED 0x2001BD44 0x69100000
STACK 0x20081F68:    0x00000000 0x00000000 0x00000000 0x00000000
STACK 0x20081F78:    0x00000000 0x00000000 0x00000000 0x00000000
scsi_accel_rp2040_stopWrite() timeout
SCSI DMA state: WRITE_DONE
Current buffer: 0x00000000/0x00000000, next 0x00000000 bytes
SyncOffset: 0 SyncPeriod 0
PIO Parity SM: tx_fifo 0, rx_fifo 0, pc 28, instr 0x000080A0
PIO Data SM: tx_fifo 3, rx_fifo 0, pc 27, instr 0x0000201A
PIO Sync SM: tx_fifo 0, rx_fifo 0, pc 0, instr 0x00000000
DMA CH A: ctrl: 0x0002C011 count: 0x00000000
DMA CH B: ctrl: 0x040AE009 count: 0x00000001
DMA CH C: ctrl: 0x0004E005 count: 0x00000000
DMA CH D: ctrl: 0x00052009 count: 0x00000000
GPIO states: 0x16B7FAFF

[rjnorthrow]

I tried the fixes suggested in #405 and they didn't work so that bears out your theory that this is something different

[rjnorthrow]

I tried asking Claude to help. It came up with the following patches, however they did not fix the problem but may help.

It concluded that the problem is that, when bluescsi sends the final data request after a large transfer (which is accepted by the suprascsi), it waits in vain for the suprascsi driver to reply with an ACK, hence the freeze.

It also suggested incrementing dataPhaseDelayUs = <delay> along with another patch, but that didn't help after getting up to 16000

--- a/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp
+++ b/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp
@@ -250,11 +250,18 @@
             {
                 dataPhaseDelay = 400; // Legacy default (Akai S1000/S3000)
             }
-            if (dataPhaseDelay > 0 && (phase == DATA_IN || phase == DATA_OUT))
+            if (dataPhaseDelay > 0 && (phase == DATA_IN || phase == DATA_OUT || phase == STATUS))
             {
                 // Akai S1000/S3000 seems to need extra delay before changing to data phase
                 // after a command. The code in BlueSCSI_disk.cpp tries to do this while waiting
                 // for SD card, to avoid any extra latency.
+                //
+                // Also applied to STATUS entry: some legacy hosts (e.g. SupraDrive
+                // 500XP's suprascsi.device) fail to respond to the STATUS byte's REQ
+                // after a large preceding DATA_IN phase, apparently because the host
+                // is still processing the just-received data and hasn't returned to
+                // polling the bus in time. A short settle delay before driving STATUS
+                // gives it time to catch up.
                 s2s_delay_ns(dataPhaseDelay * 1000);
             }
         }

Here are the four patches it came up with initially

diff --git a/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp b/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp
index 594dc2d1..5e65139c 100644
--- a/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp
+++ b/lib/BlueSCSI_platform_RP2MCU/scsiPhy.cpp
@@ -225,7 +225,23 @@ extern "C" uint32_t scsiEnterPhaseImmediate(int phase)
     {
         // ANSI INCITS 362-2002 SPI-3 10.7.1:
         // Phase changes are not allowed while REQ or ACK is asserted.
-        while (likely(!scsiDev.resetFlag) && SCSI_IN(ACK)) {}
+        // This wait has no protocol-defined timeout, but some older/quirky
+        // initiators fail to release ACK on the final byte of a large data
+        // phase, which previously hung here indefinitely. Bound the wait so
+        // BlueSCSI can at least recover and log the condition instead of
+        // hard-freezing the whole board.
+        {
+            uint32_t ackWaitStart = platform_millis();
+            while (likely(!scsiDev.resetFlag) && SCSI_IN(ACK))
+            {
+                if ((uint32_t)(platform_millis() - ackWaitStart) > 5000)
+                {
+                    logmsg("scsiEnterPhaseImmediate() ACK release timeout, phase ", phase);
+                    scsiDev.resetFlag = 1;
+                    break;
+                }
+            }
+        }
 
         // Data phase entry delay - configurable for legacy device compatibility
         {
@@ -346,10 +362,22 @@ void scsiEnterBusFree(void)
 /* Transmit to host */
 /********************/
 
+// These single-byte handshake waits previously had no timeout beyond
+// scsiDev.resetFlag, which nothing here sets on its own. If an initiator
+// fails to complete a handshake step (e.g. doesn't pulse ACK for the
+// STATUS byte after a large preceding data phase), these could hang the
+// board indefinitely. Bound them, logging which pin/direction stalled.
 #define SCSI_WAIT_ACTIVE(pin) \
   if (!SCSI_IN(pin)) { \
     if (!SCSI_IN(pin)) { \
-      while(!SCSI_IN(pin) && !scsiDev.resetFlag); \
+      uint32_t waitStart_##pin##_active = platform_millis(); \
+      while(!SCSI_IN(pin) && !scsiDev.resetFlag) { \
+        if ((uint32_t)(platform_millis() - waitStart_##pin##_active) > 5000) { \
+          logmsg("SCSI_WAIT_ACTIVE(" #pin ") timeout"); \
+          scsiDev.resetFlag = 1; \
+          break; \
+        } \
+      } \
     } \
   }
 
@@ -359,13 +387,27 @@ void scsiEnterBusFree(void)
 
 #define SCSI_WAIT_ACTIVE_EDGE(pin) \
   if (!CHECK_EDGE(SCSI_IN_ ## pin)) { \
-    while(!SCSI_IN(pin) && !CHECK_EDGE(SCSI_IN_ ## pin) && !scsiDev.resetFlag); \
+    uint32_t waitStart_##pin##_edge = platform_millis(); \
+    while(!SCSI_IN(pin) && !CHECK_EDGE(SCSI_IN_ ## pin) && !scsiDev.resetFlag) { \
+      if ((uint32_t)(platform_millis() - waitStart_##pin##_edge) > 5000) { \
+        logmsg("SCSI_WAIT_ACTIVE_EDGE(" #pin ") timeout"); \
+        scsiDev.resetFlag = 1; \
+        break; \
+      } \
+    } \
   }
 
 #define SCSI_WAIT_INACTIVE(pin) \
   if (SCSI_IN(pin)) { \
     if (SCSI_IN(pin)) { \
-      while(SCSI_IN(pin) && !scsiDev.resetFlag); \
+      uint32_t waitStart_##pin##_inactive = platform_millis(); \
+      while(SCSI_IN(pin) && !scsiDev.resetFlag) { \
+        if ((uint32_t)(platform_millis() - waitStart_##pin##_inactive) > 5000) { \
+          logmsg("SCSI_WAIT_INACTIVE(" #pin ") timeout"); \
+          scsiDev.resetFlag = 1; \
+          break; \
+        } \
+      } \
     } \
   }
 
diff --git a/lib/SCSI2SD/src/firmware/scsi.c b/lib/SCSI2SD/src/firmware/scsi.c
index bc01ac8e..63c55699 100755
--- a/lib/SCSI2SD/src/firmware/scsi.c
+++ b/lib/SCSI2SD/src/firmware/scsi.c
@@ -1400,6 +1400,26 @@ void scsiPoll(void)
                else
                {
                        process_Status();
+
+                       // Complete MESSAGE_IN immediately, in this same poll cycle,
+                       // rather than waiting for the next scsiPoll() call. Otherwise
+                       // bluescsi_main_loop() work (network poll, etc.) can run in the
+                       // gap between STATUS completing and MESSAGE_IN/BUS_FREE
+                       // completing, stalling the SCSI phase. Guarded on phase ==
+                       // MESSAGE_IN because the OMTI linked-command quirk leaves
+                       // phase == COMMAND instead, and must not be short-circuited.
+                       if (scsiDev.phase == MESSAGE_IN)
+                       {
+                               scsiDev.atnFlag |= scsiStatusATN();
+                               if (scsiDev.atnFlag)
+                               {
+                                       process_MessageOut();
+                               }
+                               else
+                               {
+                                       process_MessageIn(1);
+                               }
+                       }
                }
        break;
 
diff --git a/src/BlueSCSI.cpp b/src/BlueSCSI.cpp
index 2f511af6..d3bad914 100644
--- a/src/BlueSCSI.cpp
+++ b/src/BlueSCSI.cpp
@@ -1592,12 +1592,18 @@ extern "C" void bluescsi_main_loop(void)
 
   platform_reset_watchdog();
   platform_poll();
-  diskEjectButtonUpdate(true);
-  blink_poll();
+
+  // These only run when idle (BUS_FREE), so they can no longer intrude on
+  // an in-progress SCSI phase. The watchdog kick above stays unconditional.
+  if (scsiDev.phase == BUS_FREE)
+  {
+    diskEjectButtonUpdate(true);
+    blink_poll();
 
 #ifdef BLUESCSI_NETWORK
-  platform_network_poll();
+    platform_network_poll();
 #endif // BLUESCSI_NETWORK
+  }
 
 #ifdef PLATFORM_HAS_INITIATOR_MODE
   if (is_initiator)
@@ -1618,7 +1624,7 @@ extern "C" void bluescsi_main_loop(void)
     // SCSI requests, so normally we only want to save during a phase where
     // the host is waiting for us. But for debugging issues where no requests
     // come through or a request hangs, it's useful to force saving of log.
-    if (scsiDev.phase == STATUS || (g_log_debug && (uint32_t)(platform_millis() - last_request_time) > 2000))
+    if (scsiDev.phase == BUS_FREE || (g_log_debug && (uint32_t)(platform_millis() - last_request_time) > 2000))
     {
       save_logfile();
       last_request_time = platform_millis();