hulk inclusion
category: bugfix
bugzilla: 187190, https://gitee.com/src-openeuler/kernel/issues/I5GWOV
CVE: NA

--------------------------------

This reverts commit b2f0e44f.

Because it will introduce following problem in ltp zram tests:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000600
PGD 0 P4D 0
Oops: 0002 [#1] SMP PTI
CPU: 28 PID: 172121 Comm: sh Kdump: loaded Tainted: G           OE    --------- -  - 4.18.0+ #2
Hardware name: Huawei RH2288H V3/BC11HGSA0, BIOS 5.15 05/21/2019
RIP: 0010:flush_disk+0x1d/0x50
RSP: 0018:ffffaf14a516fe20 EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff899e26bac380 RCX: 0000000000000000
RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff899e26bac380
RBP: ffff899e26bac380 R08: 00000000000006a9 R09: 0000000000000004
R10: ffff89cd878ff440 R11: 0000000000000001 R12: 0000000000000000
R13: ffff899e26bac398 R14: ffffaf14a516ff00 R15: ffff89cd8709c3e0
FS:  00007f78d6840740(0000) GS:ffff89fcbf480000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000600 CR3: 000000308afc0002 CR4: 00000000001606e0
Call Trace:
 revalidate_disk+0x57/0x80
 reset_store+0xaf/0x120 [zram]
 kernfs_fop_write+0x10f/0x190
 vfs_write+0xad/0x1a0
 ksys_write+0x52/0xc0
 do_syscall_64+0x5d/0x1d0
 entry_SYSCALL_64_after_hwframe+0x65/0xca

This is because "bdev->bd_disk" is not ensured to exist, just convert
"set_bit(BDEV_NEED_PART_SCAN, &bdev->bd_flags)" to
"set_bit(GD_NEED_PART_SCAN, &bdev->bd_disk->state)" is wrong.

The reason to backport it is that commit 2a57456c8973 ("block:
Fix warning in bd_link_disk_holder()") has a regression that part scan
is disabled in device_add_disk(), and this problem will be fixed in
later patch.
Signed-off-by: NYu Kuai <yukuai3@huawei.com>
Reviewed-by: NJason Yan <yanaijie@huawei.com>
Signed-off-by: NYongqiang Liu <liuyongqiang13@huawei.com>