All of lore.kernel.org
 help / color / mirror / Atom feed
From: Marcelo Manzo <marcelomanzo@gmail.com>
To: qemu-devel@nongnu.org, qemu-arm@nongnu.org
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
	"Philippe Mathieu-Daudé" <philmd@oss.qualcomm.com>,
	"Marcelo Manzo" <marcelomanzo@gmail.com>,
	"Philippe Mathieu-Daudé" <philmd@mailo.com>,
	"Jason Wang" <jasowangio@gmail.com>
Subject: [PATCH v2 15/19] hw/net/bcm2838_genet: fix PHY autonegotiation-restart deadlock
Date: Tue, 11 Aug 2026 10:35:52 -0400	[thread overview]
Message-ID: <20260811143557.7862-16-marcelomanzo@gmail.com> (raw)
In-Reply-To: <20260811143557.7862-1-marcelomanzo@gmail.com>

The guest PHY driver requests an autonegotiation restart by setting
BMCR.ANRESTART, then polls BMCR waiting for the hardware to clear
that bit as an acknowledgment. bcm2838_genet_mdio_cmd() called
FIELD_DP16() to clear ANRESTART in phy_reg_data but discarded its
return value -- FIELD_DP16() does not modify its argument in place,
it returns a new value -- so the bit was never actually cleared in
the register the guest reads back. The guest driver then spun
forever waiting for an acknowledgment that would never come,
appearing to hang the whole boot.

Assign the FIELD_DP16() result back to phy_reg_data, and also call
bcm2838_genet_phy_update_link() once the restart is handled so link
status actually gets refreshed, matching the surrounding code's own
comment ("Initiate auto-negotiation once it has been restarted").

Found by booting a real guest against this series and tracing the
MDIO command register with targeted debug prints: the guest was
observed polling BMCR indefinitely, always reading back the
ANRESTART bit still set. Confirmed fixed across repeated boots: link
comes up and autonegotiation completes every time.

Signed-off-by: Marcelo Manzo <marcelomanzo@gmail.com>
---
 hw/net/bcm2838_genet.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/hw/net/bcm2838_genet.c b/hw/net/bcm2838_genet.c
index 746e416e53..dfdc5ac728 100644
--- a/hw/net/bcm2838_genet.c
+++ b/hw/net/bcm2838_genet.c
@@ -20,6 +20,8 @@
 #include "hw/net/bcm2838_genet.h"
 #include "trace.h"
 
+static void bcm2838_genet_phy_update_link(BCM2838GenetState *s);
+
 /* GENET layouts */
 REG32(GENET_SYS_REV_CTRL,              0)
 FIELD(GENET_SYS_REV_CTRL, GPHY_REV,    0, 16)
@@ -392,10 +394,15 @@ static uint64_t bcm2838_genet_mdio_cmd(BCM2838GenetState *s, uint64_t cmd)
                 if (phy_reg_id == BCM2838_GENET_PHY_BMCR) {
                     /* Initiate auto-negotiation once it has been restarted */
                     if (anrestart == 1) {
-                        FIELD_DP16(phy_reg_data, GENET_PHY_BMCR, ANRESTART, 0);
+                        phy_reg_data = FIELD_DP16(phy_reg_data,
+                                                  GENET_PHY_BMCR,
+                                                  ANRESTART, 0);
                     }
                 }
                 *phy_reg = phy_reg_data;
+                if (phy_reg_id == BCM2838_GENET_PHY_BMCR && anrestart == 1) {
+                    bcm2838_genet_phy_update_link(s);
+                }
             }
         }
     }
-- 
2.47.1



  parent reply	other threads:[~2026-08-11 14:38 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-11 14:35 [PATCH v2 00/19] hw/arm/raspi4b: working PCIe and GENET (real networking) Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 01/19] hw/arm/bcm2838_pcie: add BCM2838 PCIe Root Complex Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 02/19] hw/arm/bcm2838_pcie: add BCM2838 PCIe host Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 03/19] hw/arm/bcm2838: enable BCM2838 PCIe host bridge Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 04/19] hw/net/bcm2838_genet: add GENET stub device Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 05/19] hw/net/bcm2838_genet: add GENET register structs, part 1/4 Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 06/19] hw/net/bcm2838_genet: add GENET register structs, part 2/4 Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 07/19] hw/net/bcm2838_genet: add GENET register structs, part 3/4 Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 08/19] hw/net/bcm2838_genet: add GENET register structs, part 4/4 Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 09/19] hw/net/bcm2838_genet: add GENET register access macros Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 10/19] hw/net/bcm2838_genet: implement GENET register ops Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 11/19] hw/net/bcm2838_genet: implement GENET MDIO Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 12/19] hw/net/bcm2838_genet: implement GENET TX path Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 13/19] hw/net/bcm2838_genet: implement GENET RX path Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 14/19] hw/arm/bcm2838: enable BCM2838 GENET controller Marcelo Manzo
2026-08-11 14:35 ` Marcelo Manzo [this message]
2026-08-11 14:35 ` [PATCH v2 16/19] hw/net/bcm2838_genet: fix bogus RX checksum reporting Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 17/19] hw/net/bcm2838_genet: fix TX ring activation check Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 18/19] docs/system/arm/raspi: move PCIe/GENET from missing to implemented Marcelo Manzo
2026-08-11 14:35 ` [PATCH v2 19/19] tests/functional/aarch64: add raspi4b GENET networking test Marcelo Manzo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260811143557.7862-16-marcelomanzo@gmail.com \
    --to=marcelomanzo@gmail.com \
    --cc=jasowangio@gmail.com \
    --cc=peter.maydell@linaro.org \
    --cc=philmd@mailo.com \
    --cc=philmd@oss.qualcomm.com \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.