On 6/3/21 7:33 AM, Raviteja Narayanam wrote:
[...]

quoted

quoted

+	if (i2c->dynamic) {
+		u8 bytes;
+		u16 val;
+
+		/* Clear and enable Rx full interrupt. */
+		xiic_irq_clr_en(i2c, XIIC_INTR_RX_FULL_MASK |
+				XIIC_INTR_TX_ERROR_MASK);
+
+		/*
+		 * We want to get all but last byte, because the TX_ERROR

IRQ

quoted

+		 * is used to indicate error ACK on the address, and
+		 * negative ack on the last received byte, so to not mix
+		 * them receive all but last.
+		 * In the case where there is only one byte to receive
+		 * we can check if ERROR and RX full is set at the same time
+		 */
+		rx_watermark = msg->len;
+		bytes = min_t(u8, rx_watermark, IIC_RX_FIFO_DEPTH);
+		bytes--;
+
+		xiic_setreg8(i2c, XIIC_RFD_REG_OFFSET, bytes);
+
+		local_irq_save(flags);
+		if (!(msg->flags & I2C_M_NOSTART))
+			/* write the address */
+			xiic_setreg16(i2c, XIIC_DTR_REG_OFFSET,
+				      i2c_8bit_addr_from_msg(msg) |
+				      XIIC_TX_DYN_START_MASK);
+
+		xiic_irq_clr_en(i2c, XIIC_INTR_BNB_MASK);
+
+		/* If last message, include dynamic stop bit with length */
+		val = (i2c->nmsgs == 1) ? XIIC_TX_DYN_STOP_MASK : 0;
+		val |= msg->len;
+
+		xiic_setreg16(i2c, XIIC_DTR_REG_OFFSET, val);
+		local_irq_restore(flags);

Is local_irq_save()/local_irq_restore() used here to prevent concurrent
access to the XIIC ?

These were used to fix the timing constraint between two register writes to the IP.
As we have discussed last time, these are removed in " [PATCH 08/10] i2c: xiic: Remove interrupt enable/disable in Rx path".
Now they are no longer needed since our IP is fixed.
For legacy IP versions, driver is switching to 'AXI I2C standard mode' of operation in
" [PATCH 07/10] i2c: xiic: Switch to Xiic standard mode for i2c-read".

I see, I would expect such fixes to be at the beginning of the series, 
so they can be picked into linux-stable.

In fact, now that you mention the bugfixes which I posted a year ago, 
they also fixed a multitude of locking issues on SMP in the xiic driver. 
Has this series been tested on SMP Zynq or ZynqMP extensively ?