From c39a8362df8c09b9c66bb9c5295dc26373244fed Mon Sep 17 00:00:00 2001 From: David Hildenbrand Date: Mon, 26 Nov 2018 09:57:34 +0000 Subject: [PATCH 35/35] pc-dimm: turn alignment assert into check RH-Author: David Hildenbrand Message-id: <20181126095734.30666-1-david@redhat.com> Patchwork-id: 83163 O-Subject: [RHEL-8.0 qemu-kvm PATCH] pc-dimm: turn alignment assert into check Bugzilla: 1630116 RH-Acked-by: Pankaj Gupta RH-Acked-by: Dr. David Alan Gilbert RH-Acked-by: Igor Mammedov BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1630116 Brew: https://brewweb.engineering.redhat.com/brew/taskinfo?taskID=19276337 Upstream: 4d8938a05db15dea2c86c4ab9c5f872f160d2188 Branch: rhel8/master-2.12.0 The start of the address space indicates which maximum alignment is supported by our machine (e.g. ppc, x86 1GB). This is helpful to catch fragmenting guest physical memory in strange fashions. Right now we can crash QEMU by e.g. (there might be easier examples) qemu-system-x86_64 -m 256M,maxmem=20G,slots=2 \ -object memory-backend-file,id=mem0,size=8192M,mem-path=/dev/zero,align=8192M \ -device pc-dimm,id=dimm1,memdev=mem0 Backport conflicts: hw/mem/memory-device.c: The memory device refactoring is part of 3.0, so fix it in previous pc-dimm code. Note: The upstream patch missed a "x" (0% .. vs. 0x% ..), which was fixed in 7c63ba2055a0 ("memory-device: fix alignment error message"), however as this is not a clean cherry pick, I'm fixing it right away (like in the RHEL7.6 backport). Signed-off-by: David Hildenbrand Message-Id: <20180607154705.6316-2-david@redhat.com> Reviewed-by: Michael S. Tsirkin Reviewed-by: Igor Mammedov Signed-off-by: Paolo Bonzini (cherry picked from commit 4d8938a05db15dea2c86c4ab9c5f872f160d2188) Signed-off-by: David Hildenbrand Signed-off-by: Danilo C. L. de Paula --- hw/mem/pc-dimm.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/hw/mem/pc-dimm.c b/hw/mem/pc-dimm.c index 51350d9..a9d7c51 100644 --- a/hw/mem/pc-dimm.c +++ b/hw/mem/pc-dimm.c @@ -298,14 +298,19 @@ uint64_t pc_dimm_get_free_addr(uint64_t address_space_start, uint64_t new_addr, ret = 0; uint64_t address_space_end = address_space_start + address_space_size; - g_assert(QEMU_ALIGN_UP(address_space_start, align) == address_space_start); - if (!address_space_size) { error_setg(errp, "memory hotplug is not enabled, " "please add maxmem option"); goto out; } + /* address_space_start indicates the maximum alignment we expect */ + if (QEMU_ALIGN_UP(address_space_start, align) != address_space_start) { + error_setg(errp, "the alignment (0x%" PRIx64 ") is not supported", + align); + goto out; + } + if (hint && QEMU_ALIGN_UP(*hint, align) != *hint) { error_setg(errp, "address must be aligned to 0x%" PRIx64 " bytes", align); -- 1.8.3.1