driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

When IMP goes errors, the hardware reports a RAS to the driver,
the driver record this kind of error. Then a IMP reset will happen,
the driver checks the reason and takes the corresponding action
when doing IMP reset.

So this patch adds imp_err_state field to the struct hclge_dev
to record the error type, and handle_imp_error ops to handle it.

Feature or Bugfix: Bugfix
Signed-off-by: NWeihang Li <liweihang@hisilicon.com>
Signed-off-by: NHuazhong Tan <tanhuazhong@huawei.com>
Reviewed-by: Nlipeng <lipeng321@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category:bugfix
bugzilla:4472
CVE:NA

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

This patch adds card_id protection.
Reviewed-by: NWu Like <wulike1@huawei.com>
Signed-off-by: NXue <xuechaojing@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

The format specifier "%p" "%llx" can leak kernel addresses.
Use "%pK" instead.
Signed-off-by: Ntanghui20 <tanghui20@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

This patch adds state machine for uacce and QM. Related state machine and
lock design documents can be found:

https://github.com/hisilicon/dev-docs/blob/master/warpdrive/state_model.rst
https://github.com/hisilicon/dev-docs/blob/master/warpdrive/uacce_lock.rst

This patch also solves below problems:

- Remove uacce_queue pool in uacce, let's low level driver to maintain
queue pool. So locks(uacce_mutex, uacce->q_lock) for uacce are also
removed.

- Provide an ioctl to put hardware queue, as hardware queue which is put in
uacce .release callback will be delayed to put in kernel.

- Modify reset logic of uacce. UACCE_ST_RST state of uacce has been
deleted. Current logic is: before doing hardware reset, a SIGIO has been
sent to process which has been binded with uacce_queue, low level driver
waits user to close all fds before doing uacce_unregister; after hardware
reset done, low level driver can register to uacce subsystem again.

- Modify the way to send reset signal: use send_sig_info to send signal.

Known issue as comments of function uacce_send_sig_to_client:

"This function can be called in low level driver, which may bring a race
with uacce_fops_release. The problem is this function may be called
when q is NULL. Low level driver should avoid this by locking hardware
queue pool and check if there is related hardware queue before calling
this function.

Modify the sec/rde module codes to adapt the uacce/qm changes.
Signed-off-by: Ntanshukun (A) <tanshukun1@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: NMingqiang Ling <lingmingqiang@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

As SQE/CQE dfx regs high 16bits is q_index, low 6bits is qm_index.
need to distinguish this.

q_index and qm_index need different dfx interface.
fix qm_index in zip and hpre.
and fix hpre nfe reg enable value.

user write decimal val by design.
test sample: echo 1 > current_q
not like: echo 0x10000 > current_q
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

RAS err print level from dev_warn to dev_err.
RAS CE/NFE/FE err enable need reconfig after control reset.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

As mentioned in this DTS, insmod zip module fails after triggering zip
2ECC error when zip module is not loading.

The reason is when zip 2ECC irq source is not cleared and zip PCI bus
master enable bit is setting which will trigger FLR in hardware level
happen at the same time, hardware(FLR flow) will hange.

And when zip 2ECC irq source is not cleared, zip master interface will
be blocked.

So in this patch, we configure qm register to disable FLR when zip PCI
bus master enable bit is setting, and also clear zip error source in
driver probe flow which will make zip master interface is not blocked
(otherwise even FLR is OK, but master interface is block, mailbox
operation will still fail).
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Nfanghao <fanghao11@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

uacce_mode only support 0/1/2, other option is invalid.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

clear the cnt regs of hpre, when probe and rmmod hisi_hpre.ko
Signed-off-by: Nyumeng18 <yumeng18@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

1. add copyright state.
2. modify some nonstandard, as
        (2018-2023) -> (2018-2019)
        Hisilicon -> HiSilicon
        GPL -> GPL v2
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nhuangdaode <huangdaode@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

1.Devil digital used
2.function arguments exceed
3.Macro usage irregular
Signed-off-by: Ntanghui20 <tanghui20@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

crypto/hisilicon/zip: fix bug of reset when enable VF.
Signed-off-by: Ntanshukun (A) <tanshukun1@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

"QM all queues are busy!" will print more if multiple threads
request queues and the number of requests more than queues's.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

for "PKCS#1 RSA" before the kernel of 4.17, when it is N-crt mode,
parameters of p,q,dp,dq,qinv are 0x30, which are different from 4.17
or newer version(they are 0x00), so we should judge it is CRT or
NCRT, may u have better idea.
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nhuangdaode <huangdaode@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Current logic will cause error operations as the parameter's
high byte(big number's low byte) is zero, which will be cut by
'hpre_bn_format' function. Now, we just cut the leading zeros, which
are the big number's high bytes zero. Corresponding updating should be
carried out in user space warpdrive.
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

rde newest code merge.
Signed-off-by: NWangyuan (ange, IT) <wangyuan46@huawei.com>
Reviewed-by: Nhucheng.hu <hucheng.hu@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Hpre FLR bug fixed.
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

fix for qp full status check by qp->used.
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

sq_head pointer need udapte after get sqe buffer
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nlingmingqiang <lingmingqiang@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

update hisi_sec driver
Signed-off-by: NZhangwei <zhangwei375@huawei.com>
Reviewed-by: Nhucheng.hu <hucheng.hu@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

crypto/hisilicon/qm: mask eqe irq for uacce_q.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Force to set QP_STOP when there is BD in ACC.
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

limit the values of uacce_mode when insmod hisi_hpre.ko
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

These problems are found by QM UT by TanXiaofei, thanks xiaofei!
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Nhuangdaode <huangdaode@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

SRIOV problem, add "#ifdef CONFIG_PCI_IOV" in hpre_remove().
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

This patch adds uacce_reset_prepare before doing zip and hpre FLR and
uacce_reset_done after zip and hpre FLR.
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Nfanghao <fanghao11@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

fix eq_depth and cq_head flush logic.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

fix hpre's BD_AWUSER register addr.
Signed-off-by: Nyumeng <yumeng18@huawei.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

[  686.684022] Call trace:
[  686.684025]  __vunmap+0x44/0xf8
[  686.684026]  vunmap+0x24/0x30
[  686.684030]  arch_dma_free+0x84/0xa8
[  686.684032]  dma_direct_free+0x30/0x38
[  686.684033]  dma_free_attrs+0xa0/0xe8
[  686.684037]  uacce_queue_drain+0x12c/0x180 [uacce]
[  686.684039]  uacce_fops_flush+0x70/0x80 [uacce]
[  686.684041]  filp_close+0x30/0x90
[  686.684045]  put_files_struct+0xf8/0x100
[  686.684048]  exit_files+0x40/0x58
[  686.684052]  do_exit+0x294/0xa18
[  686.684054]  do_group_exit+0x38/0xa0
[  686.684055]  __arm64_sys_exit_group+0x14/0x18
[  686.684059]  el0_svc_common+0x5c/0x100
[  686.684060]  el0_svc_handler+0x2c/0x80
[  686.684062]  el0_svc+0x8/0xc
[  686.684063] ---[ end trace 82054841c5907aec ]---
[  686.684070] Unable to handle kernel paging request at virtual
	       address ffff00001aee0204
As user tries to disable SRIOV on the PCI device, we try to check
whether there is any user including kernel crypto or user space. If
any, disabling returns failure. For devices who don't use UACCE, there
is no such check.
Signed-off-by: Nxuzaibo <xuzaibo@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Bugfix on "QM wirteback sqc cache fail"

Changes to be committed:
	modified:   drivers/crypto/hisilicon/qm.c
Reviewed-by: Nhucheng.hu <hucheng.hu@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

crypto/hisilicon/qm: bugfix for clear the qp full status for async
interface.
1.qp_full flags should clear if sqe/cqe deal with.
2.cqe head point should update otherwise cq overflow.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nlingmingqiang <lingmingqiang@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Add sec asyn interface

	modified:   drivers/crypto/hisilicon/sec2/sec.h
	modified:   drivers/crypto/hisilicon/sec2/sec_crypto.c
	modified:   drivers/crypto/hisilicon/sec2/sec_crypto.h
	modified:   drivers/crypto/hisilicon/sec2/sec_main.c
	modified:   drivers/crypto/hisilicon/sec2/sec_usr_if.h
Signed-off-by: NZhangwei <zhangwei375@huawei.com>
Reviewed-by: Nhucheng.hu <hucheng.hu@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

crypto/hisilicon/qm: bugfix for clear the qp full status for async interface
1.qp_full flags should clear if sqe/cqe deal with.
2.cqe head point should update otherwise cq overflow.
Signed-off-by: NHao Fang <fanghao11@huawei.com>
Reviewed-by: Nlingmingqiang <lingmingqiang@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Add the sgl code for sec/zip module.
Reviewed-by: Nfanghao <fanghao11@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Reviewed-by: Nhucheng.hu <hucheng.hu@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NMingqiang Ling <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Flush file operation is an OS default calling, in some scenarioes, which
will cause dangerous behavior. So, as user space ioctl for putting queue
is brought in, to make sure all the kernel resources are freed immediately
as user space calls API of wd_release_queue.
Signed-off-by: Nxuzaibo <xuzaibo@huawei.com>
Reviewed-by: Nlingmingqiang <lingmingqiang@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NMingqiang Ling <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

Description:uacce: Fix vunmap in uacce_start_queue
Signed-off-by: NZhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Nxuzaibo <xuzaibo@huawei.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NMingqiang Ling <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

[  867.662129] Call trace:
[  867.664565]  dump_backtrace+0x0/0x1c0
[  867.668213]  show_stack+0x24/0x30
[  867.671516]  dump_stack+0xa8/0xcc
[  867.674818]  oom_show_debug_info+0x20/0xe0
[  867.678900]  out_of_memory+0x1f0/0x520
[  867.682634]  __alloc_pages_nodemask+0xca8/0xd38
[  867.687151]  iommu_dma_alloc+0x178/0x428
[  867.691060]  __iommu_alloc_attrs+0x280/0x3f8
[  867.695318]  uacce_create_region+0x42c/0x5e8 [uacce]
[  867.700269]  uacce_fops_mmap+0x298/0x338 [uacce]
[  867.704871]  mmap_region+0x3e8/0x5b8
[  867.708432]  do_mmap+0x304/0x470
[  867.711646]  vm_mmap_pgoff+0xf4/0x128
[  867.715294]  ksys_mmap_pgoff+0xb4/0x258
[  867.719115]  __arm64_sys_mmap+0x34/0x48
[  867.722937]  el0_svc_common+0xa0/0x180
[  867.726672]  el0_svc_handler+0x38/0x78
[  867.730407]  el0_svc+0x8/0xc
[  867.742023] error, MEM_PrintAllMemory is NULL!
[  867.746455]
[  867.746455] slab info:
[  867.750287] slabinfo - version: 2.1

Add munmap as wd_release_queue in UACCE
Signed-off-by: Nxuzaibo <xuzaibo@huawei.com>
Reviewed-by: Nwangzhou <wangzhou1@hisilicon.com>
Signed-off-by: Nlingmingqiang <lingmingqiang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

This patch modifies some comments on debugfs centralized review.

Feature or Bugfix:Bugfix
Signed-off-by: NYufeng Mo <moyufeng@huawei.com>
Reviewed-by: Nlipeng <lipeng321@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

In some scenarios, upper application can ensure that there is
no concurrency when processing io, thus lock free can be used
to improve performance.

Feature or Bugfix:Bugfix
Signed-off-by: NYixian Liu <liuyixian@huawei.com>
Reviewed-by: Nwangxi <wangxi11@huawei.com>
Reviewed-by: Noulijun <oulijun@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

If msg start sge idx in sq wqe is not initialized, data inconsistency
might occur during retransmission.

Feature or Bugfix: Bugfix
Signed-off-by: NWeihang Li <liweihang@hisilicon.com>
Reviewed-by: Nliuyixian <liuyixian@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>