ruby: Fixed a deadlock when restoring a checkpoint with garnet

Information
Submitter:	Emilio Castillo
Repository:	gem5
Branch:
Bugs:
Depends On:
Reviewers
Groups:	Default
People:

Description

This patch fixes a problem where in Garnet, the enqueue time in the VCallocator and the SWallocator
which is of type Cycles was being stored inside a variable with int type.

This lead to a known problem restoring checkpoints with garnet & the fixed pipeline enabled.
That value was really big and didn't fit in the variable overflowing it, therefore some conditions on the VC allocation stage & the SW allocation stage
were not met and the packets didn't advance through the network, leading to a deadlock panic right after the checkpoint was restored.

Testing Done

Originally the checkpoint was obtained in the atomic mode by doing some like this:
build/X86/gem5.opt configs/example/fs.py -n 16 --script=canneal_16c_simsmall.rcS --kernel=x86_64-vmlinux-2.6.28.4-smp  --disk-image=x86root.img

The m5checkpoint instruction is somewhere the binary i'm executing.

The checkpoint was restored with:

build/X86_MESI_CMP_directory/gem5.opt configs/example/ruby_fs.py -n 16 --kernel=x86_64-vmlinux-2.6.28.4-smp  --disk-image=x86root.img -r1 --num-l2cache=16 --num-dir=16 --topo=Mesh --mesh-rows=4 --restore-with=timing --garnet=fixed

Description:

		This patch fixes a problem where in Garnet, the enqueue time in the VCallocator and the SWallocator
		which is of type Cycles was being stored inside a variable with int type.

		This lead to a known problem restoring checkpoints with garnet & the fixed pipeline enabled.
		That value was really big and didn't fit in the variable overflowing it, therefore some conditions on the VC allocation stage & the SW allocation stage
		were not met and the packets didn't advance through the network, leading to a deadlock panic right after the checkpoint was restored.
-
-		Originally the checkpoint was obtained in the atomic mode by doing some like this:
-		build/X86/gem5.opt configs/example/fs.py -n 16 --script=canneal_16c_simsmall.rcS --kernel=x86_64-vmlinux-2.6.28.4-smp --disk-image=x86root.img
-
-		The m5checkpoint instruction is somewhere the binary i'm executing.
-
-		The checkpoint was restored with:
-
-		build/X86_MESI_CMP_directory/gem5.opt configs/example/ruby_fs.py -n 16 --kernel=x86_64-vmlinux-2.6.28.4-smp --disk-image=x86root.img -r1 --num-l2cache=16 --num-dir=16 --topo=Mesh --mesh-rows=4 --restore-with=timing --garnet=fixed

Testing Done:

	+
	+
	+	Originally the checkpoint was obtained in the atomic mode by doing some like this:
	+	build/X86/gem5.opt configs/example/fs.py -n 16 --script=canneal_16c_simsmall.rcS --kernel=x86_64-vmlinux-2.6.28.4-smp --disk-image=x86root.img
	+
	+	The m5checkpoint instruction is somewhere the binary i'm executing.
	+
	+	The checkpoint was restored with:
	+
	+	build/X86_MESI_CMP_directory/gem5.opt configs/example/ruby_fs.py -n 16 --kernel=x86_64-vmlinux-2.6.28.4-smp --disk-image=x86root.img -r1 --num-l2cache=16 --num-dir=16 --topo=Mesh --mesh-rows=4 --restore-with=timing --garnet=fixed

Good catch.

Nice to see a new contributor.

Please format the "keyword: summary" according to http://www.gem5.org/Commit_Access and I would say it's good to go.

Emilio Castillo Oct. 30, 2013, 6:36 a.m. (Oct. 30, 2013, 6:36 a.m.)

I'm trying to do the hg push but it's giving me "ssl required" errors, can you please give me a hand on this? 
Thanks a lot for your time.

Andreas Hansson Oct. 30, 2013, 6:43 a.m. (Oct. 30, 2013, 6:43 a.m.)

You will not be able to push. Nilay can hopefully help?

For the patch itself, officially the summary (including the keyword) should be < 65 char.

Summary:

-	Fixed a problem in the garnet fixed pipeline router that could lead to a deadlock when restoring a checkpoint with ruby & garnet enabled.
+	ruby: Fixed a bug that could lead to a deadlock when restoring a checkpoint with ruby & garnet enabled.

Ship It!

Ship It!

Summary:

-	ruby: Fixed a bug that could lead to a deadlock when restoring a checkpoint with ruby & garnet enabled.
+	ruby: Fixed a deadlock when restoring a checkpoint with garnet

You have a pending review.

Review Board 2.0.15

This change has been marked as submitted.

Screenshots

Files

Description:

Testing Done:

Summary:

Summary:

Status: Closed (submitted)