jwboyer pushed to kernel (f20). "Backport patch to fix tg3 deadlock (rhbz 1207789)"

[notifications at fedoraproject.org]

>From 38b65a5673d50ee78562fd7fd9f54cf8b327f8a9 Mon Sep 17 00:00:00 2001
From: Josh Boyer <jwboyer at fedoraproject.org>
Date: Wed, 1 Apr 2015 20:30:14 -0400
Subject: Backport patch to fix tg3 deadlock (rhbz 1207789)


diff --git a/kernel.spec b/kernel.spec
index 2f2658a..e9a0759 100644
--- a/kernel.spec
+++ b/kernel.spec
@@ -778,6 +778,9 @@ Patch26175: xen-pciback-Don-t-disable-PCI_COMMAND-on-PCI-device-.patch
 #rhbz 1203913
 Patch26176: sunrpc-make-debugfs-file-creation-failure-non-fatal.patch
 
+#rhbz 1207789
+Patch26177: tg3-Hold-tp-lock-before-calling-tg3_halt-from-tg3_in.patch
+
 # END OF PATCH DEFINITIONS
 
 %endif
@@ -1525,6 +1528,9 @@ ApplyPatch xen-pciback-Don-t-disable-PCI_COMMAND-on-PCI-device-.patch
 #rhbz 1203913
 ApplyPatch sunrpc-make-debugfs-file-creation-failure-non-fatal.patch
 
+#rhbz 1207789
+ApplyPatch tg3-Hold-tp-lock-before-calling-tg3_halt-from-tg3_in.patch
+
 # END OF PATCH APPLICATIONS
 
 %endif
@@ -2337,6 +2343,7 @@ fi
 #                 ||     ||
 %changelog
 * Wed Apr 01 2015 Josh Boyer <jwboyer at fedoraproject.org>
+- Backport patch to fix tg3 deadlock (rhbz 1207789)
 - Fix gssproxy (rhbz 1203913)
 - CVE-2015-2150 xen: NMIs triggerable by guests (rhbz 1196266 1200397)
 
diff --git a/tg3-Hold-tp-lock-before-calling-tg3_halt-from-tg3_in.patch b/tg3-Hold-tp-lock-before-calling-tg3_halt-from-tg3_in.patch
new file mode 100644
index 0000000..b65d0a0
--- /dev/null
+++ b/tg3-Hold-tp-lock-before-calling-tg3_halt-from-tg3_in.patch
@@ -0,0 +1,44 @@
+From: "Jun'ichi Nomura \\(NEC\\)" <j-nomura at ce.jp.nec.com>
+Date: Thu, 12 Feb 2015 01:26:24 +0000
+Subject: [PATCH] tg3: Hold tp->lock before calling tg3_halt() from
+ tg3_init_one()
+
+tg3_init_one() calls tg3_halt() without tp->lock despite its assumption
+and causes deadlock.
+If lockdep is enabled, a warning like this shows up before the stall:
+
+  [ BUG: bad unlock balance detected! ]
+  3.19.0test #3 Tainted: G            E
+  -------------------------------------
+  insmod/369 is trying to release lock (&(&tp->lock)->rlock) at:
+  [<ffffffffa02d5a1d>] tg3_chip_reset+0x14d/0x780 [tg3]
+  but there are no more locks to release!
+
+tg3_init_one() doesn't call tg3_halt() under normal situation but
+during kexec kdump I hit this problem.
+
+Fixes: 932f19de ("tg3: Release tp->lock before invoking synchronize_irq()")
+Signed-off-by: Jun'ichi Nomura <j-nomura at ce.jp.nec.com>
+Signed-off-by: David S. Miller <davem at davemloft.net>
+---
+ drivers/net/ethernet/broadcom/tg3.c | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/drivers/net/ethernet/broadcom/tg3.c b/drivers/net/ethernet/broadcom/tg3.c
+index 96bf01ba32dd..05ae12690117 100644
+--- a/drivers/net/ethernet/broadcom/tg3.c
++++ b/drivers/net/ethernet/broadcom/tg3.c
+@@ -17868,8 +17868,10 @@ static int tg3_init_one(struct pci_dev *pdev,
+ 	 */
+ 	if ((tr32(HOSTCC_MODE) & HOSTCC_MODE_ENABLE) ||
+ 	    (tr32(WDMAC_MODE) & WDMAC_MODE_ENABLE)) {
++		tg3_full_lock(tp, 0);
+ 		tw32(MEMARB_MODE, MEMARB_MODE_ENABLE);
+ 		tg3_halt(tp, RESET_KIND_SHUTDOWN, 1);
++		tg3_full_unlock(tp);
+ 	}
+ 
+ 	err = tg3_test_dma(tp);
+-- 
+2.1.0
+
-- 
cgit v0.10.2


	http://pkgs.fedoraproject.org/cgit/kernel.git/commit/?h=f20&id=38b65a5673d50ee78562fd7fd9f54cf8b327f8a9