RE: another pmem variant V2

From: Elliott, Robert (Server Storage)
Date: Wed Apr 01 2015 - 15:34:43 EST


> -----Original Message-----
> From: linux-kernel-owner@xxxxxxxxxxxxxxx [mailto:linux-kernel-
> owner@xxxxxxxxxxxxxxx] On Behalf Of Christoph Hellwig
> Sent: Thursday, March 26, 2015 3:33 AM
> To: linux-nvdimm@xxxxxxxxxxx; linux-fsdevel@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; x86@xxxxxxxxxx
> Cc: ross.zwisler@xxxxxxxxxxxxxxx; axboe@xxxxxxxxx; boaz@xxxxxxxxxxxxx
> Subject: another pmem variant V2
>

I triggered a paging error in the memcpy call for a block read
from system-udevd (actually in a modified memcpy() for the cache
attribute experiments).

1. This triggered an illegal schedule() call from an atomic context.
The call trace is shown below.

2. memcpy() doesn't provide exception handling or error reporting.
Some functions like do so, like __copy_user_nocache in
arch/x85/lib/copy_user_nocache_64.S.

Should pmem only use functions that do so, if available on the
architecture?

pmem_rw_page can pass along the return value from the copy function.
pmem_make_request can report the error, if any, via bio_endio.


Call trace
==========
[62117.317216] BUG: scheduling while atomic: systemd-udevd/22135/0x00000001
[62117.317232] Modules linked in: pmem ip6table_filter ip6_tables iptable_filter ip_tables ebtable_nat ebtables sg vfat fat x86_pkg_temp_thermal coretemp kvm_intel kvm crc32c_intel ghash_clmulni_intel aesni_intel aes_x86_64 glue_helper lrw gf128mul ablk_helper cryptd xhci_pci hpilo xhci_hcd sb_edac edac_core microcode iTCO_wdt iTCO_vendor_support hpwdt ioatdma shpchp pcspkr lpc_ich mfd_core i2c_i801 wmi pcc_cpufreq dca acpi_cpufreq uinput nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs exportfs sr_mod cdrom sd_mod bnx2x tg3 ahci mdio libahci ptp pps_core hpsa libcrc32c dm_mirror dm_region_hash dm_log dm_mod ipv6 autofs4 [last unloaded: pmem]
[62117.317233] CPU: 31 PID: 22135 Comm: systemd-udevd Tainted: G D 4.0.0-rc6+ #7
[62117.317234] Hardware name: HP ProLiant DL380 Gen9
[62117.317235] ffff88047f3f3ac0 ffff8804241db2e8 ffffffff815a8866 00000000ff86ff86
[62117.317236] ffff8804241dbfd8 ffff8804241db2f8 ffffffff815a4b45 ffff8804241db348
[62117.317237] ffffffff815ab893 ffff880457091050 ffff88047f3fbb20 0000000000000000
[62117.317237] Call Trace:
[62117.317240] [<ffffffff815a8866>] dump_stack+0x45/0x57
[62117.317245] [<ffffffff815a4b45>] __schedule_bug+0x46/0x54
[62117.317247] [<ffffffff815ab893>] __schedule+0x793/0x870
[62117.317251] [<ffffffff815ac0f0>] ? bit_wait+0x50/0x50
[62117.317252] [<ffffffff815ab9a7>] schedule+0x37/0x90
[62117.317253] [<ffffffff815ae0cc>] schedule_timeout+0x1dc/0x260
[62117.317258] [<ffffffff810bab5e>] ? ktime_get+0x3e/0xa0
[62117.317259] [<ffffffff815ab06c>] io_schedule_timeout+0xac/0x140
[62117.317261] [<ffffffff815ac126>] bit_wait_io+0x36/0x50
[62117.317262] [<ffffffff815abeeb>] __wait_on_bit_lock+0x4b/0xb0
[62117.317263] [<ffffffff81136f62>] ? find_get_entries+0xe2/0x130
[62117.317265] [<ffffffff811342ec>] __lock_page+0xac/0xb0
[62117.317269] [<ffffffff81090830>] ? autoremove_wake_function+0x40/0x40
[62117.317276] [<ffffffff8114322f>] truncate_inode_pages_range+0x3af/0x620
[62117.317278] [<ffffffff8128e837>] ? cpumask_next_and+0x37/0x50
[62117.317279] [<ffffffff811c6d80>] ? __brelse+0x40/0x40
[62117.317283] [<ffffffff810c8add>] ? smp_call_function_many+0x5d/0x280
[62117.317284] [<ffffffff8128e929>] ? free_cpumask_var+0x9/0x10
[62117.317285] [<ffffffff810c8ebd>] ? on_each_cpu_cond+0xbd/0x160
[62117.317286] [<ffffffff811c6d80>] ? __brelse+0x40/0x40
[62117.317288] [<ffffffff811434b5>] truncate_inode_pages+0x15/0x20
[62117.317289] [<ffffffff811cab13>] kill_bdev+0x33/0x40
[62117.317291] [<ffffffff811cbfa8>] __blkdev_put+0x68/0x210
[62117.317293] [<ffffffff811cca20>] blkdev_put+0x50/0x130
[62117.317294] [<ffffffff811ccbb5>] blkdev_close+0x25/0x30
[62117.317296] [<ffffffff811969f7>] __fput+0xe7/0x220
[62117.317298] [<ffffffff81196b7e>] ____fput+0xe/0x10
[62117.317302] [<ffffffff8106c554>] task_work_run+0xc4/0xe0
[62117.317306] [<ffffffff810533f8>] do_exit+0x2d8/0xb10
[62117.317308] [<ffffffff810a4b6c>] ? kmsg_dump+0x9c/0xc0
[62117.317312] [<ffffffff8100634e>] oops_end+0x8e/0xd0
[62117.317313] [<ffffffff815a428f>] no_context+0x2d4/0x334
[62117.317314] [<ffffffff815a435c>] __bad_area_nosemaphore+0x6d/0x1c6
[62117.317317] [<ffffffff811506c0>] ? zone_statistics+0x80/0xa0
[62117.317319] [<ffffffff815a44c8>] bad_area_nosemaphore+0x13/0x15
[62117.317321] [<ffffffff81043ea1>] __do_page_fault+0x91/0x430
[62117.317322] [<ffffffff8104424c>] do_page_fault+0xc/0x10
[62117.317324] [<ffffffff815b0a62>] page_fault+0x22/0x30
[62117.317325] [<ffffffffa0078302>] ? pmem_do_bvec.isra.6+0x212/0x3f0 [pmem]
[62117.317326] [<ffffffffa0078523>] pmem_rw_page+0x43/0x60 [pmem]
[62117.317328] [<ffffffff81293148>] ? __radix_tree_preload+0x38/0xa0
[62117.317329] [<ffffffff811ca9de>] bdev_read_page+0x2e/0x40
[62117.317330] [<ffffffff811d131f>] do_mpage_readpage+0x51f/0x6c0
[62117.317331] [<ffffffff8114211e>] ? lru_cache_add+0xe/0x10
[62117.317332] [<ffffffff811d159b>] mpage_readpages+0xdb/0x130
[62117.317333] [<ffffffff811ca990>] ? I_BDEV+0x10/0x10
[62117.317334] [<ffffffff811ca990>] ? I_BDEV+0x10/0x10
[62117.317336] [<ffffffff811cb14d>] blkdev_readpages+0x1d/0x20
[62117.317336] [<ffffffff811404d4>] __do_page_cache_readahead+0x194/0x210
[62117.317337] [<ffffffff811408e5>] force_page_cache_readahead+0x75/0xb0
[62117.317338] [<ffffffff81140963>] page_cache_sync_readahead+0x43/0x50
[62117.317339] [<ffffffff81136161>] generic_file_read_iter+0x431/0x630
[62117.317341] [<ffffffff811cb4e7>] blkdev_read_iter+0x37/0x40
[62117.317342] [<ffffffff8119466e>] new_sync_read+0x7e/0xb0
[62117.317343] [<ffffffff81195838>] __vfs_read+0x18/0x50
[62117.317344] [<ffffffff811958f6>] vfs_read+0x86/0x140
[62117.317345] [<ffffffff811959f6>] SyS_read+0x46/0xb0
[62117.317346] [<ffffffff810eac94>] ? __audit_syscall_entry+0xb4/0x110
[62117.317348] [<ffffffff815aeff2>] system_call_fastpath+0x12/0x17
[62121.618505] note: systemd-udevd[22133] exited with preempt_count 1


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/