Thread (2 messages) 2 messages, 2 authors, 2021-09-01

Re: [patch, v2] x86/pat: pass valid address to sanitize_phys()

From: Dan Williams <hidden>
Date: 2021-09-01 17:08:40
Also in: linux-mm, lkml

On Wed, Aug 11, 2021 at 2:06 PM Jeff Moyer [off-list ref] wrote:
The end address passed to memtype_reserve() is handed directly to
sanitize_phys().  However, end is exclusive and sanitize_phys() expects
an inclusive address.  If end falls at the end of the physical address
space, sanitize_phys() will return 0.  This can result in drivers
failing to load, and the following warning:

[    9.999440] mpt3sas version 29.100.01.00 loaded
[    9.999817] mpt3sas_cm0: 64 BIT PCI BUS DMA ADDRESSING SUPPORTED, total mem (65413664 kB)
[    9.999819] ------------[ cut here ]------------
[    9.999826] WARNING: CPU: 26 PID: 749 at arch/x86/mm/pat.c:354 reserve_memtype+0x262/0x450
[    9.999828] reserve_memtype failed: [mem 0x3ffffff00000-0xffffffffffffffff], req uncached-minus
[    9.999828] Modules linked in: mpt3sas(+) bnxt_en(+) ahci(+) crct10dif_pclmul crct10dif_common nvme crc32c_intel libahci nvme_core libata raid_class scsi_transport_sas devlink drm_panel_orientation_quirks nfit libnvdimm dm_mirror dm_region_hash dm_log dm_mod
[    9.999840] CPU: 26 PID: 749 Comm: systemd-udevd Not tainted 3.10.0-1077.el7_7.mpt3sas_test008.x86_64 #1
[    9.999842] Hardware name: Inspur SA5112M5/SA5112M5, BIOS 4.1.12 02/24/2021
[    9.999843] Call Trace:
[    9.999851]  [<ffffffffa497c4e4>] dump_stack+0x19/0x1b
[    9.999857]  [<ffffffffa429bc08>] __warn+0xd8/0x100
[    9.999859]  [<ffffffffa429bc8f>] warn_slowpath_fmt+0x5f/0x80
[    9.999861]  [<ffffffffa427b1f2>] reserve_memtype+0x262/0x450
[    9.999867]  [<ffffffffa4276254>] __ioremap_caller+0xf4/0x330
[    9.999872]  [<ffffffffc04620a1>] ? mpt3sas_base_map_resources+0x151/0xa60 [mpt3sas]
[    9.999875]  [<ffffffffa42764aa>] ioremap_nocache+0x1a/0x20
[    9.999879]  [<ffffffffc04620a1>] mpt3sas_base_map_resources+0x151/0xa60 [mpt3sas]
[    9.999884]  [<ffffffffa442656b>] ? __kmalloc+0x1eb/0x230
[    9.999889]  [<ffffffffc0465555>] mpt3sas_base_attach+0xf5/0xa50 [mpt3sas]
[    9.999894]  [<ffffffffc046af3c>] _scsih_probe+0x4ec/0xb00 [mpt3sas]
[    9.999901]  [<ffffffffa45d297a>] local_pci_probe+0x4a/0xb0
[    9.999903]  [<ffffffffa45d40c9>] pci_device_probe+0x109/0x160
[    9.999909]  [<ffffffffa46b7225>] driver_probe_device+0xc5/0x3e0
[    9.999910]  [<ffffffffa46b7623>] __driver_attach+0x93/0xa0
[    9.999912]  [<ffffffffa46b7590>] ? __device_attach+0x50/0x50
[    9.999914]  [<ffffffffa46b4dc5>] bus_for_each_dev+0x75/0xc0
[    9.999916]  [<ffffffffa46b6b9e>] driver_attach+0x1e/0x20
[    9.999918]  [<ffffffffa46b6640>] bus_add_driver+0x200/0x2d0
[    9.999920]  [<ffffffffa46b7cb4>] driver_register+0x64/0xf0
[    9.999922]  [<ffffffffa45d3905>] __pci_register_driver+0xa5/0xc0
[    9.999924]  [<ffffffffc049b000>] ? 0xffffffffc049afff
[    9.999928]  [<ffffffffc049b16e>] _mpt3sas_init+0x16e/0x1000 [mpt3sas]
[    9.999933]  [<ffffffffa420210a>] do_one_initcall+0xba/0x240
[    9.999940]  [<ffffffffa431e95a>] load_module+0x271a/0x2bb0
[    9.999946]  [<ffffffffa45b0600>] ? ddebug_proc_write+0x100/0x100
[    9.999948]  [<ffffffffa431eedf>] SyS_init_module+0xef/0x140
[    9.999954]  [<ffffffffa498fed2>] system_call_fastpath+0x25/0x2a
[    9.999955] ---[ end trace 6d6eea4438db89ef ]---
[    9.999957] ioremap reserve_memtype failed -22
[   10.000087] mpt3sas_cm0: unable to map adapter memory! or resource not found
[   10.000334] mpt3sas_cm0: failure at drivers/scsi/mpt3sas/mpt3sas_scsih.c:10597/_scsih_probe()!

(Note that this warning was from an older distribution kernel, so line
numbers and file names may not line up with the current tree.)

Fix this by passing the inclusive end address to sanitize_phys().

Fixes: 510ee090abc3 ("x86/mm/pat: Prepare {reserve, free}_memtype() for "decoy" addresses")
Signed-off-by: Jeff Moyer <redacted>
Reviewed-by: David Hildenbrand <redacted>
Reviewed-by: Dan Williams <redacted>
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help