From: Krishna Reddy <vdumpa@nvidia.com>
To: catalin.marinas@arm.com, will.deacon@arm.com, vdumpa@nvidia.com,
linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
jglisse@redhat.com, linux-tegra@vger.kernel.org
Subject: [PATCH] arm64: mm: Set MAX_PHYSMEM_BITS based on ARM64_VA_BITS
Date: Thu, 9 Nov 2017 14:58:59 -0800 [thread overview]
Message-ID: <1510268339-21989-1-git-send-email-vdumpa@nvidia.com> (raw)
MAX_PHYSMEM_BITS greater than ARM64_VA_BITS is causing memory
access fault, when HMM_DMIRROR test is enabled.
In the failing case, ARM64_VA_BITS=39 and MAX_PHYSMEM_BITS=48.
HMM_DMIRROR test selects phys memory range from end based on
MAX_PHYSMEM_BITS and gets mapped into VA space linearly.
As VA space is 39-bit and phys space is 48-bit, this has caused
incorrect mapping and leads to memory access fault.
Limiting the MAX_PHYSMEM_BITS to ARM64_VA_BITS fixes the issue and is
the right thing instead of hard coding it as 48-bit always.
[ 3.378655] Unable to handle kernel paging request at virtual address 3befd000000
[ 3.378662] pgd = ffffff800a04b000
[ 3.378900] [3befd000000] *pgd=0000000081fa3003, *pud=0000000081fa3003, *pmd=0060000268200711
[ 3.378933] Internal error: Oops: 96000044 [#1] PREEMPT SMP
[ 3.378938] Modules linked in:
[ 3.378948] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 4.9.52-tegra-g91402fdc013b-dirty #51
[ 3.378950] Hardware name: quill (DT)
[ 3.378954] task: ffffffc1ebac0000 task.stack: ffffffc1eba64000
[ 3.378967] PC is at __memset+0x1ac/0x1d0
[ 3.378976] LR is at sparse_add_one_section+0xf8/0x174
[ 3.378981] pc : [<ffffff80084c212c>] lr : [<ffffff8008eda17c>] pstate: 404000c5
[ 3.378983] sp : ffffffc1eba67a40
[ 3.378993] x29: ffffffc1eba67a40 x28: 0000000000000000
[ 3.378999] x27: 000000000003ffff x26: 0000000000000040
[ 3.379005] x25: 00000000000003ff x24: ffffffc1e9f6cf80
[ 3.379010] x23: ffffff8009ecb2d4 x22: 000003befd000000
[ 3.379015] x21: ffffffc1e9923ff0 x20: 000000000003ffff
[ 3.379020] x19: 00000000ffffffef x18: ffffffffffffffff
[ 3.379025] x17: 00000000000024d7 x16: 0000000000000000
[ 3.379030] x15: ffffff8009cd8690 x14: ffffffc1e9f6c70c
[ 3.379035] x13: ffffffc1e9f6c70b x12: 0000000000000030
[ 3.379039] x11: 0000000000000040 x10: 0101010101010101
[ 3.379044] x9 : 0000000000000000 x8 : 000003befd000000
[ 3.379049] x7 : 0000000000000000 x6 : 000000000000003f
[ 3.379053] x5 : 0000000000000040 x4 : 0000000000000000
[ 3.379058] x3 : 0000000000000004 x2 : 0000000000ffffc0
[ 3.379063] x1 : 0000000000000000 x0 : 000003befd000000
[ 3.379064]
[ 3.379069] Process swapper/0 (pid: 1, stack limit = 0xffffffc1eba64028)
[ 3.379071] Call trace:
[ 3.379079] [<ffffff80084c212c>] __memset+0x1ac/0x1d0
[ 3.379085] [<ffffff8008ed5100>] __add_pages+0x130/0x2e0
[ 3.379093] [<ffffff8008211cf4>] hmm_devmem_pages_create+0x20c/0x310
[ 3.379100] [<ffffff8008211fcc>] hmm_devmem_add+0x1d4/0x270
[ 3.379128] [<ffffff80087111c8>] dmirror_probe+0x50/0x158
[ 3.379137] [<ffffff8008732590>] platform_drv_probe+0x60/0xc8
[ 3.379143] [<ffffff800872fbf4>] driver_probe_device+0x26c/0x420
[ 3.379149] [<ffffff800872fecc>] __driver_attach+0x124/0x128
[ 3.379155] [<ffffff800872d388>] bus_for_each_dev+0x88/0xe8
[ 3.379166] [<ffffff800872f248>] driver_attach+0x30/0x40
[ 3.379171] [<ffffff800872ec18>] bus_add_driver+0x1f8/0x2b0
[ 3.379177] [<ffffff8008730e38>] driver_register+0x68/0x100
[ 3.379183] [<ffffff80087324d4>] __platform_driver_register+0x5c/0x68
[ 3.379192] [<ffffff800951f918>] hmm_dmirror_init+0x88/0xc4
[ 3.379200] [<ffffff800808359c>] do_one_initcall+0x5c/0x170
[ 3.379208] [<ffffff80094e0dd0>] kernel_init_freeable+0x1b8/0x258
[ 3.379231] [<ffffff8008ed44f0>] kernel_init+0x18/0x108
[ 3.379236] [<ffffff80080832d0>] ret_from_fork+0x10/0x40
[ 3.379246] ---[ end trace 578db63bb139b8b8 ]---
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
---
arch/arm64/include/asm/sparsemem.h | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/arch/arm64/include/asm/sparsemem.h b/arch/arm64/include/asm/sparsemem.h
index 74a9d301819f..19ecd0b0f3a3 100644
--- a/arch/arm64/include/asm/sparsemem.h
+++ b/arch/arm64/include/asm/sparsemem.h
@@ -17,7 +17,13 @@
#define __ASM_SPARSEMEM_H
#ifdef CONFIG_SPARSEMEM
+
+#ifdef CONFIG_ARM64_VA_BITS
+#define MAX_PHYSMEM_BITS CONFIG_ARM64_VA_BITS
+#else
#define MAX_PHYSMEM_BITS 48
+#endif
+
#define SECTION_SIZE_BITS 30
#endif
--
2.1.4
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next reply other threads:[~2017-11-09 22:59 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-11-09 22:58 Krishna Reddy [this message]
2017-11-10 15:11 ` [PATCH] arm64: mm: Set MAX_PHYSMEM_BITS based on ARM64_VA_BITS Robin Murphy
2017-11-12 17:55 ` Jerome Glisse
2017-11-13 10:32 ` Suzuki K Poulose
2017-11-13 12:56 ` Robin Murphy
2017-11-13 16:51 ` Suzuki K Poulose
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=1510268339-21989-1-git-send-email-vdumpa@nvidia.com \
--to=vdumpa@nvidia.com \
--cc=catalin.marinas@arm.com \
--cc=jglisse@redhat.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-tegra@vger.kernel.org \
--cc=will.deacon@arm.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).