1:  2669060f036c ! 1:  5512493c52c9 wifi: rtw88: sdio: Fix unhandled RX request interrupt storm
    @@ Commit message
         In rtw_sdio_handle_interrupt(), the HISR status register is cleared using
         Write-1-to-Clear (W1C) semantics. However, the driver masks out the
         REG_SDIO_HISR_RX_REQUEST bit in the local 'hisr' variable before writing
    -    it back, causing a 0 to be written to that bit.
    +    it back, causing a 0 to be written to that bit. This prevents the RX
    +    request interrupt from being acknowledged and cleared in hardware, trapping
    +    the CPU core in an infinite interrupt storm loop (starving the RCU preempt
    +    kthread and locking up the system).
     
    -    This prevents the RX request interrupt from being acknowledged and
    -    cleared by the hardware, trapping the CPU core in an infinite interrupt
    -    storm loop upon receiving packets and triggering RCU stalls and system
    -    lockups.
    +    Masking out the bit in software was originally intended to allow
    +    budget-limited polling without losing interrupts. However, because the
    +    status register was never acknowledged, the hardware interrupt line
    +    remained asserted continuously.
     
    -    Remove the masking of REG_SDIO_HISR_RX_REQUEST so that the interrupt
    -    status is correctly written back as a 1 and acknowledged.
    +    Resolve this by adopting the standard interrupt masking pattern:
    +    1. Mask interrupts via rtw_sdio_disable_interrupt(rtwdev) upon entry.
    +    2. Acknowledge pending status bits via rtw_write32(rtwdev, REG_SDIO_HISR,
    +       hisr) without clearing REG_SDIO_HISR_RX_REQUEST from the writeback mask.
    +    3. Service the pending interrupt events (rtw_sdio_rx_isr).
    +    4. Re-enable interrupts via rtw_sdio_enable_interrupt(rtwdev).
    +
    +    Any packet arriving over the air while interrupts are masked latches
    +    REG_SDIO_HISR_RX_REQUEST in hardware. Unmasking HIMR restores the interrupt
    +    mask and immediately re-asserts the SDIO interrupt line (DAT[1]), causing
    +    ksdioirqd to run another pass and drain the new packets.
     
         Fixes: 65371a3f14e7 ("wifi: rtw88: sdio: Add HCI implementation for SDIO based chipsets")
         Cc: stable@vger.kernel.org
    -    Assisted-by: LLM
         Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
     
      ## drivers/net/wireless/realtek/rtw88/sdio.c ##
     @@ drivers/net/wireless/realtek/rtw88/sdio.c: static void rtw_sdio_handle_interrupt(struct sdio_func *sdio_func)
    + 	rtwsdio->irq_thread = current;
    + 
    + 	hisr = rtw_read32(rtwdev, REG_SDIO_HISR);
    ++	if (!hisr)
    ++		goto out;
    ++
    ++	rtw_sdio_disable_interrupt(rtwdev);
    ++	rtw_write32(rtwdev, REG_SDIO_HISR, hisr);
      
      	if (hisr & REG_SDIO_HISR_TXERR)
      		rtw_sdio_tx_err_isr(rtwdev);
    @@ drivers/net/wireless/realtek/rtw88/sdio.c: static void rtw_sdio_handle_interrupt
      		rtw_sdio_rx_isr(rtwdev);
     -	}
      
    - 	rtw_write32(rtwdev, REG_SDIO_HISR, hisr);
    +-	rtw_write32(rtwdev, REG_SDIO_HISR, hisr);
    ++	/* Unmasking HIMR re-asserts the IRQ line if new packets arrived */
    ++	rtw_sdio_enable_interrupt(rtwdev);
    + 
    ++out:
    + 	rtwsdio->irq_thread = NULL;
    + }