Re: [PATCH v4 2/5] r8169: Coalesce RTL8411b PHY power-down recovery calls to reduce spinlock stalls

[Jacob Keller]

On 10/29/2023 4:04 AM, Mirsad Goran Todorovac wrote:
> On RTL8411b the RX unit gets confused if the PHY is powered-down.
> This was reported in [0] and confirmed by Realtek. Realtek provided
> a sequence to fix the RX unit after PHY wakeup.
> 
> A series of about 130 r8168_mac_ocp_write() calls is performed to
> program the RTL registers for recovery.
> 
> With about 130 of these sequential calls to r8168_mac_ocp_write() this looks like
> a lock storm that will stall all of the cores and CPUs on the same memory controller
> for certain time I/O takes to finish.
> 
> In a sequential case of RTL register programming, a sequence of writes to the RTL
> registers can be coalesced under a same raw spinlock. This can dramatically decrease
> the number of bus stalls in a multicore or multi-CPU system:
> 
>     static void rtl_hw_start_8411_2(struct rtl8169_private *tp)
>     {
> 
> 	...
> 
> 	/* The following Realtek-provided magic fixes an issue with the RX unit
> 	 * getting confused after the PHY having been powered-down.
> 	 */
> 
> 	static const struct recover_8411b_info init_zero_seq[] = {
> 	    { 0xFC28, 0x0000 }, { 0xFC2A, 0x0000 }, { 0xFC2C, 0x0000 },
> 	    ...
> 	};
> 
> 	static const struct recover_8411b_info recover_seq[] = {
> 	    { 0xF800, 0xE008 }, { 0xF802, 0xE00A }, { 0xF804, 0xE00C },
> 	    ...
> 	};
> 
> 	static const struct recover_8411b_info final_seq[] = {
> 	    { 0xFC2A, 0x0743 }, { 0xFC2C, 0x0801 }, { 0xFC2E, 0x0BE9 },
> 	    ...
> 	};
> 
> 	r8168_mac_ocp_write_seq(tp, init_zero_seq);
> 	mdelay(3);
> 	r8168_mac_ocp_write(tp, 0xFC26, 0x0000);
> 	r8168_mac_ocp_write_seq(tp, recover_seq);
> 	r8168_mac_ocp_write(tp, 0xFC26, 0x8000);
> 	r8168_mac_ocp_write_seq(tp, final_seq);
>     }
> 
> The hex data is preserved intact through s/r8168_mac_ocp_write[(]tp,/{ / and s/[)];/ },/
> functions that only changed the function names and the ending of the line, so the actual
> hex data is unchanged.
> 
> Note that the reason for the introduction of the original commit
> was to enable recovery of the RX unit on the RTL8411b which was confused by the
> powered-down PHY. This sequence of r8168_mac_ocp_write() calls amplifies the problem
> into a series of about 500+ memory bus locks, most waiting for the main MMIO memory
> read-modify-write under a LOCK. The memory barrier in RTL_W32 should suffice for
> the programming sequence to reach RTL NIC registers.
> 
> [0] https://bugzilla.redhat.com/show_bug.cgi?id=1692075
> 
> Fixes: fe4e8db0392a6 ("r8169: fix issue with confused RX unit after PHY power-down on RTL8411b")
> Cc: Heiner Kallweit <hkallweit1@xxxxxxxxx>
> Cc: Marco Elver <elver@xxxxxxxxxx>
> Cc: nic_swsd@xxxxxxxxxxx
> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx>
> Cc: Eric Dumazet <edumazet@xxxxxxxxxx>
> Cc: Jakub Kicinski <kuba@xxxxxxxxxx>
> Cc: Paolo Abeni <pabeni@xxxxxxxxxx>
> Cc: netdev@xxxxxxxxxxxxxxx
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> Link: https://lore.kernel.org/lkml/20231028005153.2180411-1-mirsad.todorovac@xxxxxxxxxxxx/
> Link: https://lore.kernel.org/lkml/20231028110459.2644926-1-mirsad.todorovac@xxxxxxxxxxxx/
> Signed-off-by: Mirsad Goran Todorovac <mirsad.todorovac@xxxxxxxxxxxx>
> ---

Reviewed-by: Jacob Keller <jacob.e.keller@xxxxxxxxx>