June 2020 - Kernel - mailweb.openeuler.org

[PATCH 01/12] jbd2: clean __jbd2_journal_abort_hard() and __journal_abort_soft()
by Yang Yingliang 12 Jun '20

12 Jun '20

From: "zhangyi (F)" <yi.zhang(a)huawei.com> mainline inclusion from mainline-5.6-rc1 commit 7f6225e446cc8dfa4c3c7959a4de3dd03ec277bf category: bugfix bugzilla: 34619 CVE: NA --------------------------- __jbd2_journal_abort_hard() is no longer used, so now we can merge __jbd2_journal_abort_hard() and __journal_abort_soft() these two functions into jbd2_journal_abort() and remove them. Signed-off-by: zhangyi (F) <yi.zhang(a)huawei.com> Reviewed-by: Jan Kara <jack(a)suse.cz> Link: https://lore.kernel.org/r/20191204124614.45424-5-yi.zhang@huawei.com Signed-off-by: Theodore Ts'o <tytso(a)mit.edu> Reviewed-by: Zhang Xiaoxu <zhangxiaoxu5(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- fs/jbd2/journal.c | 103 ++++++++++++++++++------------------------- include/linux/jbd2.h | 1 - 2 files changed, 42 insertions(+), 62 deletions(-) diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index 1f085f0bf908..7301bb766172 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -101,7 +101,6 @@ EXPORT_SYMBOL(jbd2_journal_release_jbd_inode); EXPORT_SYMBOL(jbd2_journal_begin_ordered_truncate); EXPORT_SYMBOL(jbd2_inode_cache); -static void __journal_abort_soft (journal_t *journal, int errno); static int jbd2_journal_create_slab(size_t slab_size); #ifdef CONFIG_JBD2_DEBUG @@ -810,7 +809,7 @@ int jbd2_journal_bmap(journal_t *journal, unsigned long blocknr, "at offset %lu on %s\n", __func__, blocknr, journal->j_devname); err = -EIO; - __journal_abort_soft(journal, err); + jbd2_journal_abort(journal, err); } } else { *retp = blocknr; /* +journal->j_blk_offset */ @@ -2108,64 +2107,6 @@ int jbd2_journal_wipe(journal_t *journal, int write) return err; } -/* - * Journal abort has very specific semantics, which we describe - * for journal abort. - * - * Two internal functions, which provide abort to the jbd layer - * itself are here. - */ - -/* - * Quick version for internal journal use (doesn't lock the journal). - * Aborts hard --- we mark the abort as occurred, but do _nothing_ else, - * and don't attempt to make any other journal updates. - */ -void __jbd2_journal_abort_hard(journal_t *journal) -{ - transaction_t *transaction; - - if (journal->j_flags & JBD2_ABORT) - return; - - printk(KERN_ERR "Aborting journal on device %s.\n", - journal->j_devname); - - write_lock(&journal->j_state_lock); - journal->j_flags |= JBD2_ABORT; - transaction = journal->j_running_transaction; - if (transaction) - __jbd2_log_start_commit(journal, transaction->t_tid); - write_unlock(&journal->j_state_lock); -} - -/* Soft abort: record the abort error status in the journal superblock, - * but don't do any other IO. */ -static void __journal_abort_soft (journal_t *journal, int errno) -{ - int old_errno; - - write_lock(&journal->j_state_lock); - old_errno = journal->j_errno; - if (!journal->j_errno || errno == -ESHUTDOWN) - journal->j_errno = errno; - - if (journal->j_flags & JBD2_ABORT) { - write_unlock(&journal->j_state_lock); - if (old_errno != -ESHUTDOWN && errno == -ESHUTDOWN) - jbd2_journal_update_sb_errno(journal); - return; - } - write_unlock(&journal->j_state_lock); - - __jbd2_journal_abort_hard(journal); - - jbd2_journal_update_sb_errno(journal); - write_lock(&journal->j_state_lock); - journal->j_flags |= JBD2_REC_ERR; - write_unlock(&journal->j_state_lock); -} - /** * void jbd2_journal_abort () - Shutdown the journal immediately. * @journal: the journal to shutdown. @@ -2209,7 +2150,47 @@ static void __journal_abort_soft (journal_t *journal, int errno) void jbd2_journal_abort(journal_t *journal, int errno) { - __journal_abort_soft(journal, errno); + transaction_t *transaction; + + /* + * ESHUTDOWN always takes precedence because a file system check + * caused by any other journal abort error is not required after + * a shutdown triggered. + */ + write_lock(&journal->j_state_lock); + if (journal->j_flags & JBD2_ABORT) { + int old_errno = journal->j_errno; + + write_unlock(&journal->j_state_lock); + if (old_errno != -ESHUTDOWN && errno == -ESHUTDOWN) { + journal->j_errno = errno; + jbd2_journal_update_sb_errno(journal); + } + return; + } + + /* + * Mark the abort as occurred and start current running transaction + * to release all journaled buffer. + */ + pr_err("Aborting journal on device %s.\n", journal->j_devname); + + journal->j_flags |= JBD2_ABORT; + journal->j_errno = errno; + transaction = journal->j_running_transaction; + if (transaction) + __jbd2_log_start_commit(journal, transaction->t_tid); + write_unlock(&journal->j_state_lock); + + /* + * Record errno to the journal super block, so that fsck and jbd2 + * layer could realise that a filesystem check is needed. + */ + jbd2_journal_update_sb_errno(journal); + + write_lock(&journal->j_state_lock); + journal->j_flags |= JBD2_REC_ERR; + write_unlock(&journal->j_state_lock); } /** diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h index 7dd9d25a9cbb..12b935c9ec1e 100644 --- a/include/linux/jbd2.h +++ b/include/linux/jbd2.h @@ -1427,7 +1427,6 @@ extern int jbd2_journal_skip_recovery (journal_t *); extern void jbd2_journal_update_sb_errno(journal_t *); extern int jbd2_journal_update_sb_log_tail (journal_t *, tid_t, unsigned long, int); -extern void __jbd2_journal_abort_hard (journal_t *); extern void jbd2_journal_abort (journal_t *, int); extern int jbd2_journal_errno (journal_t *); extern void jbd2_journal_ack_err (journal_t *); -- 2.25.1

1 11

Minutes of the meeting //: 【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30
by Guohanjun (Hanjun Guo) 11 Jun '20

11 Jun '20

Attendees (I can see more people join later but can’t remember now): [cid:image001.png@01D63F6C.270B2CF0] Following action items: * Vincent to share the JIRA link for thermal pressure * Vincent to share the JIRA link for scheduler improvements * Anmar to work on the opening of Power and Performance Project * Hanjun/Jonathan to share some plan for openEuler kernel in the coming 6 or 12 months * Anmar to create the open mailing list for LKQ project, so that Fengguang and the team can participate And see attached for Linaro slide, thanks! 发件人: Meeting Book via Kernel [mailto:kernel@openeuler.org] 发送时间: 2020年6月3日 16:01 收件人: kernel(a)openeuler.org 主题: 【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30 Topic Kernel & testing sync between Linaro and openEuler Time 2020-06-10 16:00-17:30((UTC+08:00)Beijing) Join Conference Join (External) >><https://welink-meeting.zoom.us/j/834726983> Meeting ID 834 726 983 Convener 郭寒军 Agenda * Linux Kernel Update by Linaro - 15 mins * Linux Kernel Functional Testing (LKFT) introduction - 15mins * Brief introduction for openEuler and the kernel - 15mins * Brief introduction for Crystal testing system - 15 mins * Open discussions - about 30 mins Tips: Dial the access number to join conference<http://app.huawei.com/eshare/voiceMeetings> 注：仅支持部分国家和地区。( Only some countries and regions are supported.) 请点击会议链接按照提示下载Zoom客户端入会，并升级Zoom最新版本。 Please click the conference link and download the Zoom to join the conference, and upgrade to the latest version of Zoom.

2 1

[PATCH] irqchip/gic-v3-its: its supoport s3/s4
by Hongbo Yao 09 Jun '20

09 Jun '20

commit dba0bc7b76dcf(irqchip/gic-v3-its: add ability to save/restore ITS state) add support to save/restore its suspend state. However, it needs to set ITS_FLAGS_SAVE_SUSPEND_STATE to 1 when we want to resend MAPC. In GIC500, GITS_typer.HCC is not zero, and it can be used to judge if we can resend MAPC if the firmware restores the state, but when GITS_typer.HCC is zero, and it will never have chance to set the flag, which will lead to a failuer to power off GIC logic in suspend. Signed-off-by: Hongbo Yao <yaohongbo(a)huawei.com> --- drivers/irqchip/irq-gic-v3-its.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index 881466163c99..6ea2809b8680 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -3733,8 +3733,7 @@ static int __init its_probe_one(struct resource *res, ctlr |= GITS_CTLR_ImDe; writel_relaxed(ctlr, its->base + GITS_CTLR); - if (GITS_TYPER_HCC(typer)) - its->flags |= ITS_FLAGS_SAVE_SUSPEND_STATE; + its->flags |= ITS_FLAGS_SAVE_SUSPEND_STATE; err = its_init_domain(handle, its); if (err) -- 2.20.1

1 0

(no subject)
by liaichun 09 Jun '20

09 Jun '20

From: LiAichun <liaichun(a)huawei.com> euleros inclusion commit NA category:bugfix bugzilla:NA CVE:NA issue:I1JZHT ------------------------------------------------- From:YueHaibing<> Subject: [PATCH] netdevsim: Fix error handling in nsim_fib_init and neim_fib_exit Date: Mon, 8 Jun 2020 15:13:40 +0800 In nsim_fib_init(),if register_fib_notifier failed,nsim_fib_net_ops should be unregistered before return; In nsim_fib_exit(),unregister_fib_notifier should be called before nsim_fib_net_ops be unregistered, otherwise may cause use-after-free More detailed information can refer to: https://lkml.org/lkml/2019/10/11/216 Reported-by:Hulk Robot <hulkci(a)huawei.com> Fixes: 59c84b9fcf42("netdevsim: Restore per-network namespace accounting for fib entries") Signed-off-by: YueHaibing <yuehaibing(a)huawei.com> Signed-off-by: LiAichun <liaichun(a)huawei.com> --- drivers/net/netdevsim/fib.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

(no subject)
by liaichun 09 Jun '20

09 Jun '20

From: LiAichun <liaichun(a)huawei.com> euleros inclusion commit NA category:bugfix bugzilla:NA CVE:NA Fix issue:gitee.com/openeuler/kernel/issues/I1JZHT ------------------------------------------------- From:YueHaibing<> Subject: [PATCH] netdevsim: Fix error handling in nsim_fib_init and neim_fib_exit Date: Mon, 8 Jun 2020 15:13:40 +0800 In nsim_fib_init(),if register_fib_notifier failed,nsim_fib_net_ops should be unregistered before return; In nsim_fib_exit(),unregister_fib_notifier should be called before nsim_fib_net_ops be unregistered, otherwise may cause use-after-free More detailed information can refer to: https://lkml.org/lkml/2019/10/11/216 Reported-by:Hulk Robot <hulkci(a)huawei.com> Fixes: 59c84b9fcf42("netdevsim: Restore per-network namespace accounting for fib entries") Signed-off-by: YueHaibing <yuehaibing(a)huawei.com> Signed-off-by: LiAichun <liaichun(a)huawei.com> --- drivers/net/netdevsim/fib.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

(no subject)
by liaichun 09 Jun '20

09 Jun '20

From: LiAichun <liaichun(a)huawei.com> mainline inclusion from mainline-4.19.90-2003.4.0 commit a91b7a28e76a702bd81357bf41a45dab6486716b category:bugfix bugzilla:NA CVE:NA ------------------------------------------------- From:YueHaibing<> Subject: [PATCH] netdevsim: Fix error handling in nsim_fib_init and neim_fib_exit Date: Mon, 8 Jun 2020 15:13:40 +0800 In nsim_fib_init(),if register_fib_notifier failed,nsim_fib_net_ops should be unregistered before return; In nsim_fib_exit(),unregister_fib_notifier should be called before nsim_fib_net_ops be unregistered, otherwise may cause use-after-free More detailed information can refer to: https://lkml.org/lkml/2019/10/11/216 Reported-by:Hulk Robot <hulkci(a)huawei.com> Fixes: 59c84b9fcf42("netdevsim: Restore per-network namespace accounting for fib entries") Signed-off-by: YueHaibing <yuehaibing(a)huawei.com> Signed-off-by: LiAichun <liaichun(a)huawei.com> --- drivers/net/netdevsim/fib.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

2 1

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 09 Jun '20

09 Jun '20

In nsim_fib_init(),if register_fib_notifier failed,nsim_fib_net_ops should be unregistered before return; In nsim_fib_exit(),unregister_fib_notifier should be called before nsim_fib_net_ops be unregistered, otherwise may cause use-after-free More detailed information can refer to: https://lkml.org/lkml/2019/10/11/216 kernel inclusion from:kernel-4.19 commit:2dca76fa95bf03c15bdcacd7cfa8f974c1a799f6 category:bugfix bugzilla:NA CVE:NA Signed-off-by: liaichun <513565428(a)qq.com> DESC:fix netdevsim resource leak --- drivers/net/netdevsim/fib.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

2 1

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

In nsim_fib_init(),if register_fib_notifier failed,nsim_fib_net_ops should be unregistered before return; In nsim_fib_exit(),unregister_fib_notifier should be called before nsim_fib_net_ops be unregistered, otherwise may cause use-after-free More detailed information can refer to: https://lkml.org/lkml/2019/10/11/216 Signed-off-by: liaichun <513565428(a)qq.com> Tested-by: liaichun <513565428(a)qq.com> Type:bugfix ID:NA SUG:restart DESC:fix netdevsim resource leak --- drivers/net/netdevsim/fib.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

2 1

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

From: liaichun <liaichun(a)huawei.com> diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

From: liaichun <liaichun(a)huawei.com> diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH] net-netdevsim-fix-resource-leak
by liaichun 08 Jun '20

08 Jun '20

diff --git a/drivers/net/netdevsim/fib.c b/drivers/net/netdevsim/fib.c index f61d094..d5c01b8 100644 --- a/drivers/net/netdevsim/fib.c +++ b/drivers/net/netdevsim/fib.c @@ -241,8 +241,8 @@ static int __net_init nsim_fib_netns_init(struct net *net) void nsim_fib_exit(void) { - unregister_pernet_subsys(&nsim_fib_net_ops); unregister_fib_notifier(&nsim_fib_nb); + unregister_pernet_subsys(&nsim_fib_net_ops); } int nsim_fib_init(void) @@ -257,6 +257,7 @@ int nsim_fib_init(void) err = register_fib_notifier(&nsim_fib_nb, nsim_fib_dump_inconsistent); if (err < 0) { + unregister_pernet_subsys(&nsim_fib_net_ops); pr_err("Failed to register fib notifier\n"); goto err_out; } -- 1.8.3.1

1 0

[PATCH v2] scsi: hisi_sas: directly trigger error handle of scsi instead of waiting for timeout
by Luo Jiaxing 05 Jun '20

05 Jun '20

We used timeout mechanism of scsi layer to trigger some IO's error handle, this type of abnormal IO require driver to enter error handle to clear the residue in the hardware. But timeout mechanism caught error handle time to be longer, some threads need to wait for tens of seconds when IO failed and enter error handle, so we try to trigger error handling directly. Signed-off-by: Luo Jiaxing <luojiaxing(a)huawei.com> --- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 4 +++- 3 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c index 031091e..49ea3b7 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c @@ -1263,8 +1263,10 @@ static int slot_complete_v1_hw(struct hisi_hba *hisi_hba, !(cmplt_hdr_data & CMPLT_HDR_RSPNS_XFRD_MSK)) { slot_err_v1_hw(hisi_hba, task, slot); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c index 6568e20..8ebddf5 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c @@ -2406,8 +2406,10 @@ slot_complete_v2_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) error_info[0], error_info[1], error_info[2], error_info[3]); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c index 5b80856..7fd067c 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c @@ -2408,8 +2408,10 @@ slot_complete_v3_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) dev_info(dev, "data underflow without sense, rsp_code:0x%x, rc:%d.\n", iu->resp_data[0], rc); } - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } -- 2.7.4

1 0

[PATCH v1 hulk-4.19-next] scsi: hisi_sas: directly trigger error handle of scsi instead of waiting for timeout
by Luo Jiaxing 05 Jun '20

05 Jun '20

driver inclusion category: bugfix bugzilla: NA DTS: NA ----------------------------------------------------------------------- We used timeout mechanism of scsi layer to trigger some IO's error handle, this type of abnormal IO require driver to enter error handle to clear the residue in the hardware. But timeout mechanism caught error handle time to be longer, some threads need to wait for tens of seconds when IO failed and enter error handle, so we try to trigger error handling directly. Signed-off-by: Luo Jiaxing <luojiaxing(a)huawei.com> --- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 4 +++- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 4 +++- 3 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c index 031091e..49ea3b7 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c @@ -1263,8 +1263,10 @@ static int slot_complete_v1_hw(struct hisi_hba *hisi_hba, !(cmplt_hdr_data & CMPLT_HDR_RSPNS_XFRD_MSK)) { slot_err_v1_hw(hisi_hba, task, slot); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c index 6568e20..8ebddf5 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c @@ -2406,8 +2406,10 @@ slot_complete_v2_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) error_info[0], error_info[1], error_info[2], error_info[3]); - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c index 5b80856..7fd067c 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c @@ -2408,8 +2408,10 @@ slot_complete_v3_hw(struct hisi_hba *hisi_hba, struct hisi_sas_slot *slot) dev_info(dev, "data underflow without sense, rsp_code:0x%x, rc:%d.\n", iu->resp_data[0], rc); } - if (unlikely(slot->abort)) + if (unlikely(slot->abort)) { + sas_task_abort(task); return ts->stat; + } goto out; } -- 2.7.4

1 0

【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30
by Meeting Book 03 Jun '20

03 Jun '20

1 0

【Meeting Notice】Kernel & testing sync between Linaro and openEuler Time: 2020-06-10 16:00-17:30
by Meeting Book 03 Jun '20

03 Jun '20

1 0

[PATCH 01/76] net: dsa: Do not make user port errors fatal
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Florian Fainelli <f.fainelli(a)gmail.com> commit 86f8b1c01a0a537a73d2996615133be63cdf75db upstream. Prior to 1d27732f411d ("net: dsa: setup and teardown ports"), we would not treat failures to set-up an user port as fatal, but after this commit we would, which is a regression for some systems where interfaces may be declared in the Device Tree, but the underlying hardware may not be present (pluggable daughter cards for instance). Fixes: 1d27732f411d ("net: dsa: setup and teardown ports") Signed-off-by: Florian Fainelli <f.fainelli(a)gmail.com> Reviewed-by: Andrew Lunn <andrew(a)lunn.ch> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/dsa/dsa2.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/dsa/dsa2.c b/net/dsa/dsa2.c index 46ae4de..b036c55 100644 --- a/net/dsa/dsa2.c +++ b/net/dsa/dsa2.c @@ -412,7 +412,7 @@ static int dsa_tree_setup_switches(struct dsa_switch_tree *dst) err = dsa_switch_setup(ds); if (err) - return err; + continue; for (port = 0; port < ds->num_ports; port++) { dp = &ds->ports[port]; -- 1.8.3

1 75

[PATCH 1/9] net/hinic: Fix the firmware compatibility bug in the MAC reuse scenario
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Chiqijun <chiqijun(a)huawei.com> driver inclusion category: bugfix bugzilla: 4472 ----------------------------------------------------------------------- 2.3.0.1 and later drivers and firmware add MAC reuse feature, when the driver is compatible with firmware version 2.3.0.0 and before, if the ip link command is used to configure the MAC and VLAN for the VF, it will cause the VF network anomaly. To solve this problem, the driver obtains the function capability from the firmware when loading, and determines whether to support MAC reuse for different processing. Signed-off-by: Chiqijun <chiqijun(a)huawei.com> Reviewed-by: Zengweiliang <zengweiliang.zengweiliang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/net/ethernet/huawei/hinic/hinic_hwdev.h | 1 + .../ethernet/huawei/hinic/hinic_mgmt_interface.h | 11 ++++ drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c | 45 +++++++++++++++-- drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h | 4 ++ drivers/net/ethernet/huawei/hinic/hinic_nic_io.c | 10 +++- drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h | 2 + drivers/net/ethernet/huawei/hinic/hinic_sriov.c | 59 ++++++++++++++-------- 7 files changed, 106 insertions(+), 26 deletions(-) diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h index a096f91..2a81867 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.h @@ -319,6 +319,7 @@ struct hinic_hwdev { struct hinic_board_info board_info; #define MGMT_VERSION_MAX_LEN 32 u8 mgmt_ver[MGMT_VERSION_MAX_LEN]; + u64 fw_support_func_flag; }; int hinic_init_comm_ch(struct hinic_hwdev *hwdev); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h index 3c3be6e..8673cf5 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h @@ -690,6 +690,15 @@ struct hinic_port_rt_cmd { u8 rsvd1[6]; }; +struct fw_support_func { + u8 status; + u8 version; + u8 rsvd0[6]; + + u64 flag; + u64 rsvd; +}; + struct hinic_vf_dcb_state { u8 status; u8 version; @@ -741,6 +750,8 @@ struct hinic_set_link_follow { int hinic_get_base_qpn(void *hwdev, u16 *global_qpn); +int hinic_get_fw_support_func(void *hwdev); + int hinic_vf_func_init(struct hinic_hwdev *hwdev); void hinic_vf_func_free(struct hinic_hwdev *hwdev); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c index f9e16ec..7dfedbd 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c @@ -201,6 +201,34 @@ int hinic_get_base_qpn(void *hwdev, u16 *global_qpn) return 0; } +int hinic_get_fw_support_func(void *hwdev) +{ + struct fw_support_func support_flag = {0}; + struct hinic_hwdev *dev = hwdev; + u16 out_size = sizeof(support_flag); + int err; + + if (!hwdev) + return -EINVAL; + + err = hinic_msg_to_mgmt_sync(hwdev, HINIC_MOD_L2NIC, + HINIC_PORT_CMD_GET_FW_SUPPORT_FLAG, + &support_flag, sizeof(support_flag), + &support_flag, &out_size, 0); + if (support_flag.status == HINIC_MGMT_CMD_UNSUPPORTED) { + nic_info(dev->dev_hdl, "Current firmware doesn't support to get function capability\n"); + support_flag.flag = 0; + } else if (support_flag.status || err || !out_size) { + nic_err(dev->dev_hdl, "Failed to get function capability, err: %d, status: 0x%x, out size: 0x%x\n", + err, support_flag.status, out_size); + return -EFAULT; + } + + dev->fw_support_func_flag = support_flag.flag; + + return 0; +} + #define HINIC_ADD_VLAN_IN_MAC 0x8000 #define HINIC_VLAN_ID_MASK 0x7FFF @@ -352,6 +380,11 @@ int hinic_update_mac_vlan(void *hwdev, u16 old_vlan, u16 new_vlan, int vf_id) if (!vf_info->pf_set_mac) return 0; + if (!FW_SUPPORT_MAC_REUSE_FUNC(dev)) { + nic_info(dev->dev_hdl, "Current firmware doesn't support mac reuse\n"); + return 0; + } + func_id = hinic_glb_pf_vf_offset(dev) + (u16)vf_id; vlan_id = old_vlan; if (vlan_id) @@ -2562,9 +2595,14 @@ static int hinic_init_vf_config(struct hinic_hwdev *hwdev, u16 vf_id) vf_info = hwdev->nic_io->vf_infos + HW_VF_ID_TO_OS(vf_id); if (vf_info->pf_set_mac) { func_id = hinic_glb_pf_vf_offset(hwdev) + vf_id; - vlan_id = vf_info->pf_vlan; - if (vlan_id) - vlan_id |= HINIC_ADD_VLAN_IN_MAC; + if (FW_SUPPORT_MAC_REUSE_FUNC(hwdev)) { + vlan_id = vf_info->pf_vlan; + if (vlan_id) + vlan_id |= HINIC_ADD_VLAN_IN_MAC; + } else { + vlan_id = 0; + } + err = hinic_set_mac(hwdev, vf_info->vf_mac_addr, vlan_id, func_id); if (err) { @@ -4065,3 +4103,4 @@ int hinic_get_sfp_type(void *hwdev, u8 *data0, u8 *data1) return 0; } + diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h index 427d469..e0979df 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h @@ -19,6 +19,10 @@ #define OS_VF_ID_TO_HW(os_vf_id) ((os_vf_id) + 1) #define HW_VF_ID_TO_OS(hw_vf_id) ((hw_vf_id) - 1) +#define FW_SUPPORT_MAC_REUSE 0x1 +#define FW_SUPPORT_MAC_REUSE_FUNC(hwdev) \ + ((hwdev)->fw_support_func_flag & FW_SUPPORT_MAC_REUSE) + #define HINIC_VLAN_PRIORITY_SHIFT 13 #define HINIC_RSS_INDIR_SIZE 256 diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c index 1cb4e26..f2fb0bc 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c @@ -831,9 +831,17 @@ int hinic_init_nic_hwdev(void *hwdev, u16 rx_buff_len) } /* VFs don't set port routine command report */ - if (hinic_func_type(dev) != TYPE_VF) + if (hinic_func_type(dev) != TYPE_VF) { + /* Get the fw support mac reuse flag */ + err = hinic_get_fw_support_func(hwdev); + if (err) { + nic_err(dev->dev_hdl, "Failed to get function capability\n"); + return err; + } + /* Inform mgmt to send sfp's information to driver */ err = hinic_set_port_routine_cmd_report(hwdev, true); + } return err; } diff --git a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h index 46464b1f..d428f48 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h @@ -146,6 +146,8 @@ enum hinic_port_cmd { HINIC_PORT_CMD_SET_PORT_LINK_STATUS = 0x76, HINIC_PORT_CMD_SET_CGE_PAUSE_TIME_CFG = 0x77, + HINIC_PORT_CMD_GET_FW_SUPPORT_FLAG = 0x79, + HINIC_PORT_CMD_SET_PORT_REPORT = 0x7B, HINIC_PORT_CMD_LINK_STATUS_REPORT = 0xa0, diff --git a/drivers/net/ethernet/huawei/hinic/hinic_sriov.c b/drivers/net/ethernet/huawei/hinic/hinic_sriov.c index 37d0116..73ab2d7 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_sriov.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_sriov.c @@ -266,6 +266,42 @@ int hinic_ndo_set_vf_mac(struct net_device *netdev, int vf, u8 *mac) /*lint -save -e574 -e734*/ #ifdef IFLA_VF_MAX +static int set_hw_vf_vlan(struct hinic_sriov_info *sriov_info, + u16 cur_vlanprio, int vf, u16 vlan, u8 qos) +{ + int err = 0; + u16 old_vlan = cur_vlanprio & VLAN_VID_MASK; + + if (vlan || qos) { + if (cur_vlanprio) { + err = hinic_kill_vf_vlan(sriov_info->hwdev, + OS_VF_ID_TO_HW(vf)); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to delete vf %d old vlan %d\n", + vf, old_vlan); + return err; + } + } + err = hinic_add_vf_vlan(sriov_info->hwdev, + OS_VF_ID_TO_HW(vf), vlan, qos); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to add vf %d new vlan %d\n", + vf, vlan); + return err; + } + } else { + err = hinic_kill_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf)); + if (err) { + nic_err(&sriov_info->pdev->dev, "Failed to delete vf %d vlan %d\n", + vf, old_vlan); + return err; + } + } + + return hinic_update_mac_vlan(sriov_info->hwdev, old_vlan, vlan, + OS_VF_ID_TO_HW(vf)); +} + #ifdef IFLA_VF_VLAN_INFO_MAX int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos, __be16 vlan_proto) @@ -276,7 +312,6 @@ int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos) struct hinic_nic_dev *adapter = netdev_priv(netdev); struct hinic_sriov_info *sriov_info; u16 vlanprio, cur_vlanprio; - int err = 0; if (!FUNC_SUPPORT_SET_VF_MAC_VLAN(adapter->hwdev)) { nicif_err(adapter, drv, netdev, @@ -298,27 +333,7 @@ int hinic_ndo_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos) if (vlanprio == cur_vlanprio) return 0; - if (vlan || qos) { - if (cur_vlanprio) - err = hinic_kill_vf_vlan(sriov_info->hwdev, - OS_VF_ID_TO_HW(vf)); - if (err) - goto out; - err = hinic_add_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf), - vlan, qos); - } else { - err = hinic_kill_vf_vlan(sriov_info->hwdev, OS_VF_ID_TO_HW(vf)); - } - - if (err) - return err; - - err = hinic_update_mac_vlan(sriov_info->hwdev, - cur_vlanprio & VLAN_VID_MASK, vlan, - OS_VF_ID_TO_HW(vf)); - -out: - return err; + return set_hw_vf_vlan(sriov_info, cur_vlanprio, vf, vlan, qos); } #endif -- 1.8.3

1 8

[PATCH] irqchip/gic-v3-its: Probe ITS page size for all GITS_BASERn registers
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Marc Zyngier <maz(a)kernel.org> mainline inclusion from mainline-v5.6-rc4 commit d5df9dc96eb7423d3f742b13d5e1e479ff795eaa category: bugfix bugzilla: NA CVE: NA --------------------------- The GICv3 ITS driver assumes that once it has latched on a page size for a given BASER register, it can use the same page size as the maximum page size for all subsequent BASER registers. Although it worked so far, nothing in the architecture guarantees this, and Nianyao Tang hit this problem on some undisclosed implementation. Let's bite the bullet and probe the the supported page size on all BASER registers before starting to populate the tables. This simplifies the setup a bit, at the expense of a few additional MMIO accesses. Signed-off-by: Marc Zyngier <maz(a)kernel.org> Reported-by: Nianyao Tang <tangnianyao(a)huawei.com> Tested-by: Nianyao Tang <tangnianyao(a)huawei.com> Link: https://lore.kernel.org/r/1584089195-63897-1-git-send-email-zhangshaokun@hi… Signed-off-by: Hongbo Yao <yaohongbo(a)huawei.com> conflicts: drivers/irqchip/irq-gic-v3-its.c Reviewed-by: Hanjun Guo <guohanjun(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/irqchip/irq-gic-v3-its.c | 101 ++++++++++++++++++++++++++------------- 1 file changed, 67 insertions(+), 34 deletions(-) diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index 59c6c0d..bfedab2 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -17,6 +17,7 @@ #include <linux/acpi.h> #include <linux/acpi_iort.h> +#include <linux/bitfield.h> #include <linux/bitmap.h> #include <linux/cpu.h> #include <linux/crash_dump.h> @@ -1768,17 +1769,16 @@ static void its_write_baser(struct its_node *its, struct its_baser *baser, } static int its_setup_baser(struct its_node *its, struct its_baser *baser, - u64 cache, u64 shr, u32 psz, u32 order, - bool indirect) + u64 cache, u64 shr, u32 order, bool indirect) { u64 val = its_read_baser(its, baser); u64 esz = GITS_BASER_ENTRY_SIZE(val); u64 type = GITS_BASER_TYPE(val); u64 baser_phys, tmp; - u32 alloc_pages; + u32 alloc_pages, psz; void *base; -retry_alloc_baser: + psz = baser->psz; alloc_pages = (PAGE_ORDER_TO_SIZE(order) / psz); if (alloc_pages > GITS_BASER_PAGES_MAX) { pr_warn("ITS@%pa: %s too large, reduce ITS pages %u->%u\n", @@ -1851,25 +1851,6 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser, goto retry_baser; } - if ((val ^ tmp) & GITS_BASER_PAGE_SIZE_MASK) { - /* - * Page size didn't stick. Let's try a smaller - * size and retry. If we reach 4K, then - * something is horribly wrong... - */ - free_pages((unsigned long)base, order); - baser->base = NULL; - - switch (psz) { - case SZ_16K: - psz = SZ_4K; - goto retry_alloc_baser; - case SZ_64K: - psz = SZ_16K; - goto retry_alloc_baser; - } - } - if (val != tmp) { pr_err("ITS@%pa: %s doesn't stick: %llx %llx\n", &its->phys_base, its_base_type_string[type], @@ -1895,13 +1876,14 @@ static int its_setup_baser(struct its_node *its, struct its_baser *baser, static bool its_parse_indirect_baser(struct its_node *its, struct its_baser *baser, - u32 psz, u32 *order, u32 ids) + u32 *order, u32 ids) { u64 tmp = its_read_baser(its, baser); u64 type = GITS_BASER_TYPE(tmp); u64 esz = GITS_BASER_ENTRY_SIZE(tmp); u64 val = GITS_BASER_InnerShareable | GITS_BASER_RaWaWb; u32 new_order = *order; + u32 psz = baser->psz; bool indirect = false; /* No need to enable Indirection if memory requirement < (psz*2)bytes */ @@ -1960,11 +1942,58 @@ static void its_free_tables(struct its_node *its) } } +static int its_probe_baser_psz(struct its_node *its, struct its_baser *baser) +{ + u64 psz = SZ_64K; + + while (psz) { + u64 val, gpsz; + + val = its_read_baser(its, baser); + val &= ~GITS_BASER_PAGE_SIZE_MASK; + + switch (psz) { + case SZ_64K: + gpsz = GITS_BASER_PAGE_SIZE_64K; + break; + case SZ_16K: + gpsz = GITS_BASER_PAGE_SIZE_16K; + break; + case SZ_4K: + default: + gpsz = GITS_BASER_PAGE_SIZE_4K; + break; + } + + gpsz >>= GITS_BASER_PAGE_SIZE_SHIFT; + + val |= FIELD_PREP(GITS_BASER_PAGE_SIZE_MASK, gpsz); + its_write_baser(its, baser, val); + + if (FIELD_GET(GITS_BASER_PAGE_SIZE_MASK, baser->val) == gpsz) + break; + + switch (psz) { + case SZ_64K: + psz = SZ_16K; + break; + case SZ_16K: + psz = SZ_4K; + break; + case SZ_4K: + default: + return -1; + } + } + + baser->psz = psz; + return 0; +} + static int its_alloc_tables(struct its_node *its) { u64 shr = GITS_BASER_InnerShareable; u64 cache = GITS_BASER_RaWaWb; - u32 psz = SZ_64K; int err, i; if (its->flags & ITS_FLAGS_WORKAROUND_CAVIUM_22375) @@ -1975,34 +2004,38 @@ static int its_alloc_tables(struct its_node *its) struct its_baser *baser = its->tables + i; u64 val = its_read_baser(its, baser); u64 type = GITS_BASER_TYPE(val); - u32 order = get_order(psz); bool indirect = false; + u32 order; - switch (type) { - case GITS_BASER_TYPE_NONE: + if (type == GITS_BASER_TYPE_NONE) continue; + if (its_probe_baser_psz(its, baser)) { + its_free_tables(its); + return -ENXIO; + } + + order = get_order(baser->psz); + + switch (type) { case GITS_BASER_TYPE_DEVICE: - indirect = its_parse_indirect_baser(its, baser, - psz, &order, + indirect = its_parse_indirect_baser(its, baser, &order, its->device_ids); break; case GITS_BASER_TYPE_VCPU: - indirect = its_parse_indirect_baser(its, baser, - psz, &order, + indirect = its_parse_indirect_baser(its, baser, &order, ITS_MAX_VPEID_BITS); break; } - err = its_setup_baser(its, baser, cache, shr, psz, order, indirect); + err = its_setup_baser(its, baser, cache, shr, order, indirect); if (err < 0) { its_free_tables(its); return err; } /* Update settings which will be used for next BASERn */ - psz = baser->psz; cache = baser->val & GITS_BASER_CACHEABILITY_MASK; shr = baser->val & GITS_BASER_SHAREABILITY_MASK; } -- 1.8.3

1 0

[PATCH 1/4] vfio/mdev: Improve the create/remove sequence
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Parav Pandit <parav(a)mellanox.com> mainline inclusion from mainline-5.2 commit 522ecce08ab20b57342d65b05601818e0f95fb2c category: bugfix bugzilla: NA CVE: NA ------------------------------------------------- This patch addresses below two issues and prepares the code to address 3rd issue listed below. 1. mdev device is placed on the mdev bus before it is created in the vendor driver. Once a device is placed on the mdev bus without creating its supporting underlying vendor device, mdev driver's probe() gets triggered. However there isn't a stable mdev available to work on. create_store() mdev_create_device() device_register() ... vfio_mdev_probe() [...] parent->ops->create() vfio_ap_mdev_create() mdev_set_drvdata(mdev, matrix_mdev); /* Valid pointer set above */ Due to this way of initialization, mdev driver who wants to use the mdev, doesn't have a valid mdev to work on. 2. Current creation sequence is, parent->ops_create() groups_register() Remove sequence is, parent->ops->remove() groups_unregister() However, remove sequence should be exact mirror of creation sequence. Once this is achieved, all users of the mdev will be terminated first before removing underlying vendor device. (Follow standard linux driver model). At that point vendor's remove() ops shouldn't fail because taking the device off the bus should terminate any usage. 3. When remove operation fails, mdev sysfs removal attempts to add the file back on already removed device. Following call trace [1] is observed. [1] call trace: kernel: WARNING: CPU: 2 PID: 9348 at fs/sysfs/file.c:327 sysfs_create_file_ns+0x7f/0x90 kernel: CPU: 2 PID: 9348 Comm: bash Kdump: loaded Not tainted 5.1.0-rc6-vdevbus+ #6 kernel: Hardware name: Supermicro SYS-6028U-TR4+/X10DRU-i+, BIOS 2.0b 08/09/2016 kernel: RIP: 0010:sysfs_create_file_ns+0x7f/0x90 kernel: Call Trace: kernel: remove_store+0xdc/0x100 [mdev] kernel: kernfs_fop_write+0x113/0x1a0 kernel: vfs_write+0xad/0x1b0 kernel: ksys_write+0x5a/0xe0 kernel: do_syscall_64+0x5a/0x210 kernel: entry_SYSCALL_64_after_hwframe+0x49/0xbe Therefore, mdev core is improved in following ways. 1. Split the device registration/deregistration sequence so that some things can be done between initialization of the device and hooking it up to the bus respectively after deregistering it from the bus but before giving up our final reference. In particular, this means invoking the ->create() and ->remove() callbacks in those new windows. This gives the vendor driver an initialized mdev device to work with during creation. At the same time, a bus driver who wish to bind to mdev driver also gets initialized mdev device. This follows standard Linux kernel bus and device model. 2. During remove flow, first remove the device from the bus. This ensures that any bus specific devices are removed. Once device is taken off the mdev bus, invoke remove() of mdev from the vendor driver. 3. The driver core device model provides way to register and auto unregister the device sysfs attribute groups at dev->groups. Make use of dev->groups to let core create the groups and eliminate code to avoid explicit groups creation and removal. To ensure, that new sequence is solid, a below stack dump of a process is taken who attempts to remove the device while device is in use by vfio driver and user application. This stack dump validates that vfio driver guards against such device removal when device is in use. cat /proc/21962/stack [<0>] vfio_del_group_dev+0x216/0x3c0 [vfio] [<0>] mdev_remove+0x21/0x40 [mdev] [<0>] device_release_driver_internal+0xe8/0x1b0 [<0>] bus_remove_device+0xf9/0x170 [<0>] device_del+0x168/0x350 [<0>] mdev_device_remove_common+0x1d/0x50 [mdev] [<0>] mdev_device_remove+0x8c/0xd0 [mdev] [<0>] remove_store+0x71/0x90 [mdev] [<0>] kernfs_fop_write+0x113/0x1a0 [<0>] vfs_write+0xad/0x1b0 [<0>] ksys_write+0x5a/0xe0 [<0>] do_syscall_64+0x5a/0x210 [<0>] entry_SYSCALL_64_after_hwframe+0x49/0xbe [<0>] 0xffffffffffffffff This prepares the code to eliminate calling device_create_file() in subsequent patch. Reviewed-by: Cornelia Huck <cohuck(a)redhat.com> Signed-off-by: Parav Pandit <parav(a)mellanox.com> Signed-off-by: Alex Williamson <alex.williamson(a)redhat.com> Conflicts: drivers/vfio/mdev/mdev_core.c -mdev_device_remove_ops() and -mdev_device_remove_cb() are removed anyway, although they are not identical compared within the original patch. The type of uuid is still uuid_le, and it has not been changed to guid_t. drivers/vfio/mdev/mdev_private.h The type of uuid is still uuid_le, and it has not been changed to guid_t. Signed-off-by: Xiaoyang Xu <xuxiaoyang2(a)huawei.com> Reviewed-by: Yingtai Xie <xieyingtai(a)huawei.com> Reviewed-by: Hanjun Guo <guohanjun(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/vfio/mdev/mdev_core.c | 94 +++++++++++----------------------------- drivers/vfio/mdev/mdev_private.h | 2 +- drivers/vfio/mdev/mdev_sysfs.c | 2 +- 3 files changed, 27 insertions(+), 71 deletions(-) diff --git a/drivers/vfio/mdev/mdev_core.c b/drivers/vfio/mdev/mdev_core.c index e052f62..9705a4a 100644 --- a/drivers/vfio/mdev/mdev_core.c +++ b/drivers/vfio/mdev/mdev_core.c @@ -103,55 +103,10 @@ static inline void mdev_put_parent(struct mdev_parent *parent) kref_put(&parent->ref, mdev_release_parent); } -static int mdev_device_create_ops(struct kobject *kobj, - struct mdev_device *mdev) -{ - struct mdev_parent *parent = mdev->parent; - int ret; - - ret = parent->ops->create(kobj, mdev); - if (ret) - return ret; - - ret = sysfs_create_groups(&mdev->dev.kobj, - parent->ops->mdev_attr_groups); - if (ret) - parent->ops->remove(mdev); - - return ret; -} - -/* - * mdev_device_remove_ops gets called from sysfs's 'remove' and when parent - * device is being unregistered from mdev device framework. - * - 'force_remove' is set to 'false' when called from sysfs's 'remove' which - * indicates that if the mdev device is active, used by VMM or userspace - * application, vendor driver could return error then don't remove the device. - * - 'force_remove' is set to 'true' when called from mdev_unregister_device() - * which indicate that parent device is being removed from mdev device - * framework so remove mdev device forcefully. - */ -static int mdev_device_remove_ops(struct mdev_device *mdev, bool force_remove) -{ - struct mdev_parent *parent = mdev->parent; - int ret; - - /* - * Vendor driver can return error if VMM or userspace application is - * using this mdev device. - */ - ret = parent->ops->remove(mdev); - if (ret && !force_remove) - return -EBUSY; - - sysfs_remove_groups(&mdev->dev.kobj, parent->ops->mdev_attr_groups); - return 0; -} - static int mdev_device_remove_cb(struct device *dev, void *data) { if (dev_is_mdev(dev)) - mdev_device_remove(dev, true); + mdev_device_remove(dev); return 0; } @@ -311,41 +266,43 @@ int mdev_device_create(struct kobject *kobj, struct device *dev, uuid_le uuid) mdev->parent = parent; kref_init(&mdev->ref); + device_initialize(&mdev->dev); mdev->dev.parent = dev; mdev->dev.bus = &mdev_bus_type; mdev->dev.release = mdev_device_release; dev_set_name(&mdev->dev, "%pUl", uuid.b); + mdev->dev.groups = parent->ops->mdev_attr_groups; + mdev->type_kobj = kobj; - ret = device_register(&mdev->dev); - if (ret) { - put_device(&mdev->dev); - goto mdev_fail; - } + ret = parent->ops->create(kobj, mdev); + if (ret) + goto ops_create_fail; - ret = mdev_device_create_ops(kobj, mdev); + ret = device_add(&mdev->dev); if (ret) - goto create_fail; + goto add_fail; ret = mdev_create_sysfs_files(&mdev->dev, type); - if (ret) { - mdev_device_remove_ops(mdev, true); - goto create_fail; - } + if (ret) + goto sysfs_fail; - mdev->type_kobj = kobj; mdev->active = true; dev_dbg(&mdev->dev, "MDEV: created\n"); return 0; -create_fail: - device_unregister(&mdev->dev); +sysfs_fail: + device_del(&mdev->dev); +add_fail: + parent->ops->remove(mdev); +ops_create_fail: + put_device(&mdev->dev); mdev_fail: mdev_put_parent(parent); return ret; } -int mdev_device_remove(struct device *dev, bool force_remove) +int mdev_device_remove(struct device *dev) { struct mdev_device *mdev, *tmp; struct mdev_parent *parent; @@ -374,16 +331,15 @@ int mdev_device_remove(struct device *dev, bool force_remove) mutex_unlock(&mdev_list_lock); type = to_mdev_type(mdev->type_kobj); + mdev_remove_sysfs_files(dev, type); + device_del(&mdev->dev); parent = mdev->parent; + ret = parent->ops->remove(mdev); + if (ret) + dev_err(&mdev->dev, "Remove failed: err=%d\n", ret); - ret = mdev_device_remove_ops(mdev, force_remove); - if (ret) { - mdev->active = true; - return ret; - } - - mdev_remove_sysfs_files(dev, type); - device_unregister(dev); + /* Balances with device_initialize() */ + put_device(&mdev->dev); mdev_put_parent(parent); return 0; diff --git a/drivers/vfio/mdev/mdev_private.h b/drivers/vfio/mdev/mdev_private.h index b5819b7..9c65d8e 100644 --- a/drivers/vfio/mdev/mdev_private.h +++ b/drivers/vfio/mdev/mdev_private.h @@ -59,6 +59,6 @@ struct mdev_type { void mdev_remove_sysfs_files(struct device *dev, struct mdev_type *type); int mdev_device_create(struct kobject *kobj, struct device *dev, uuid_le uuid); -int mdev_device_remove(struct device *dev, bool force_remove); +int mdev_device_remove(struct device *dev); #endif /* MDEV_PRIVATE_H */ diff --git a/drivers/vfio/mdev/mdev_sysfs.c b/drivers/vfio/mdev/mdev_sysfs.c index e7770b5..1286366 100644 --- a/drivers/vfio/mdev/mdev_sysfs.c +++ b/drivers/vfio/mdev/mdev_sysfs.c @@ -236,7 +236,7 @@ static ssize_t remove_store(struct device *dev, struct device_attribute *attr, if (val && device_remove_file_self(dev, attr)) { int ret; - ret = mdev_device_remove(dev, false); + ret = mdev_device_remove(dev); if (ret) { device_create_file(dev, attr); return ret; -- 1.8.3

1 3

[PATCH 1/2] arm64/mpam: Fix unreset resources when mkdir ctrl group or umount resctrl
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Wang ShaoBo <bobo.shaobowang(a)huawei.com> hulk inclusion category: bugfix bugzilla: 34588 CVE: NA ----------------------------------------------- There are two problems related to schemata: 1) When rmdir a group and then mkdir a new group under resctrl root directory, the new group still inherits the schemata configuration from old. e.g. > mount -t resctrl resctrl /sys/fs/resctrl > cd /sys/fs/resctrl > mkdir p1 && cd p1 > echo 'L3:0=7f' > schemata > cd .. && rmdir p1 && mkdir p1 && cd p1 > cat schemata L3:0=7f;1=7fff;2=7fff;3=7fff MB:0=100;1=100;2=100;3=100 2) It still exists when umount /sys/fs/resctrl and remount. e.g. > mount -t resctrl resctrl /sys/fs/resctrl > cd /sys/fs/resctrl > echo 'L3:0=7f' > schemata > umount /sys/fs/resctrl > mount -t resctrl resctrl /sys/fs/resctrl > cat schemata L3:0=7f;1=7fff;2=7fff;3=7fff MB:0=100;1=100;2=100;3=100 Firstly we make each resctrl resource obtains their corresponding default configuration. NOTE we use zero to initialize L3 default value instead of max cpbm bits, as zero configurarion equals to maximum configuration for L3 MSCs. And we use max-percentage masks of max bandwidth to generate maximum configuration for MB. Then we reset resources' configuration settings to default value and back MSCs to default state, when mkdir or umount happended. Fixes: caf75b6b2540 ("resctrlfs: mpam: init struct for mpam") Fixes: 916dd9321e3c ("resctrlfs: init support resctrlfs") Signed-off-by: Wang ShaoBo <bobo.shaobowang(a)huawei.com> Reviewed-by: Xie XiuQi <xiexiuqi(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- arch/arm64/include/asm/mpam_resource.h | 4 ++++ arch/arm64/include/asm/resctrl.h | 2 ++ arch/arm64/kernel/mpam.c | 17 +++++++++++++++++ arch/arm64/kernel/mpam_ctrlmon.c | 28 ++++++++++++++++++++++++++++ fs/resctrlfs.c | 5 +++++ 5 files changed, 56 insertions(+) diff --git a/arch/arm64/include/asm/mpam_resource.h b/arch/arm64/include/asm/mpam_resource.h index ab90596..1a6904c 100644 --- a/arch/arm64/include/asm/mpam_resource.h +++ b/arch/arm64/include/asm/mpam_resource.h @@ -89,6 +89,10 @@ /* [FIXME] hard code for hardlim */ #define MBW_MAX_SET(v) (MBW_MAX_HARDLIM|((v) << (16 - BWA_WD))) #define MBW_MAX_GET(v) (((v) & MBW_MAX_MASK) >> (16 - BWA_WD)) + +/* hard code for mbw_max max-percentage's cresponding masks */ +#define MBA_MAX_WD 63u + /* * emulate the mpam nodes * These should be reported by ACPI MPAM Table. diff --git a/arch/arm64/include/asm/resctrl.h b/arch/arm64/include/asm/resctrl.h index 0a0a12b..fb5fa6c 100644 --- a/arch/arm64/include/asm/resctrl.h +++ b/arch/arm64/include/asm/resctrl.h @@ -65,4 +65,6 @@ int mkdir_mondata_all(struct kernfs_node *parent_kn, mongroup_create_dir(struct kernfs_node *parent_kn, struct resctrl_group *prgrp, char *name, struct kernfs_node **dest_kn); +int rdtgroup_init_alloc(struct rdtgroup *rdtgrp); + #endif /* _ASM_ARM64_RESCTRL_H */ diff --git a/arch/arm64/kernel/mpam.c b/arch/arm64/kernel/mpam.c index 202e49a..120795e 100644 --- a/arch/arm64/kernel/mpam.c +++ b/arch/arm64/kernel/mpam.c @@ -86,6 +86,9 @@ static inline void mpam_node_assign_val(struct mpam_node *n, n->addr = hwpage_address; n->component_id = component_id; n->cpus_list = "0"; + + if (n->type == MPAM_RESOURCE_MC) + n->default_ctrl = MBA_MAX_WD; } #define MPAM_NODE_NAME_SIZE (10) @@ -544,6 +547,20 @@ void post_resctrl_mount(void) static int reset_all_ctrls(struct resctrl_resource *r) { + struct raw_resctrl_resource *rr; + struct rdt_domain *d; + int partid; + + rr = (struct raw_resctrl_resource *)r->res; + for (partid = 0; partid < rr->num_partid; partid++) { + list_for_each_entry(d, &r->domains, list) { + d->new_ctrl = rr->default_ctrl; + d->ctrl_val[partid] = rr->default_ctrl; + d->have_new_ctrl = true; + rr->msr_update(d, partid); + } + } + return 0; } diff --git a/arch/arm64/kernel/mpam_ctrlmon.c b/arch/arm64/kernel/mpam_ctrlmon.c index b9f9495..2b94efc 100644 --- a/arch/arm64/kernel/mpam_ctrlmon.c +++ b/arch/arm64/kernel/mpam_ctrlmon.c @@ -585,3 +585,31 @@ int resctrl_mkdir_ctrlmon_mondata(struct kernfs_node *parent_kn, } return ret; } + +/* Initialize the RDT group's allocations. */ +int rdtgroup_init_alloc(struct rdtgroup *rdtgrp) +{ + struct resctrl_resource *r; + struct raw_resctrl_resource *rr; + struct rdt_domain *d; + int ret; + + for_each_resctrl_resource(r) { + if (!r->alloc_enabled) + continue; + + rr = (struct raw_resctrl_resource *)r->res; + list_for_each_entry(d, &r->domains, list) { + d->new_ctrl = rr->default_ctrl; + d->have_new_ctrl = true; + } + + ret = update_domains(r, rdtgrp); + if (ret < 0) { + rdt_last_cmd_puts("Failed to initialize allocations\n"); + return ret; + } + } + + return 0; +} diff --git a/fs/resctrlfs.c b/fs/resctrlfs.c index 09e56384..8d463b8 100644 --- a/fs/resctrlfs.c +++ b/fs/resctrlfs.c @@ -670,6 +670,11 @@ static int resctrl_group_mkdir_ctrl_mon(struct kernfs_node *parent_kn, ret = 0; rdtgrp->closid = closid; + + ret = rdtgroup_init_alloc(rdtgrp); + if (ret < 0) + goto out_id_free; + list_add(&rdtgrp->resctrl_group_list, &resctrl_all_groups); if (resctrl_mon_capable) { -- 1.8.3

1 1

[PATCH] ext4: report error to userspace by netlink
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Zhao Minmin <zhaominmin1(a)huawei.com> hulk inclusion category: feature bugzilla: 34592 CVE: NA ------------------------------------------------- Implement the ext3/ext4 file system error report. This patch is used to implement abnormal alarm of ext3/ext4 filesystem. You can archieve this by setting "FILESYSTEM_MONITOR" or "FILESYSTEM_ALARM" on in configuration file. With this setting, alarm will be raised when ext3/ext4 file system expection occurs. Signed-off-by: Zhao Minmin <zhaominmin1(a)huawei.com> Reviewed-by: Yi Zhang <yi.zhang(a)huawei.com> Link: http://hulk.huawei.com/pipermail/kernel.openeuler/2016-March/009711.html Signed-off-by: Wang Hui <john.wanghui(a)huawei.com> Signed-off-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> [yebin: cherry-pick this patch from redhat-7.5, commit 13848b1856c1] Signed-off-by: Ye Bin <yebin10(a)huawei.com> Reviewed-by: Wenan Mao <maowenan(a)huawei.com> Reviewed-by: Yi Zhang <yi.zhang(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- fs/ext4/ext4.h | 9 ++++++++ fs/ext4/super.c | 51 ++++++++++++++++++++++++++++++++++++++++++++ include/uapi/linux/netlink.h | 1 + 3 files changed, 61 insertions(+) diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index b1bcc6e..83bec07 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -54,6 +54,15 @@ #endif #endif +#define NL_EXT4_ERROR_GROUP 1 +#define EXT4_ERROR_MAGIC 0xAE32014U +struct ext4_err_msg { + int magic; + char s_id[32]; + unsigned long s_flags; + int ext4_errno; +}; + /* * The fourth extended filesystem constants/structures */ diff --git a/fs/ext4/super.c b/fs/ext4/super.c index bd6c1cf..33f1e7c 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -54,6 +54,10 @@ #include "mballoc.h" #include "fsmap.h" +#include <uapi/linux/netlink.h> +#include <net/sock.h> +#include <net/net_namespace.h> + #define CREATE_TRACE_POINTS #include <trace/events/ext4.h> @@ -84,6 +88,8 @@ static struct dentry *ext4_mount(struct file_system_type *fs_type, int flags, static void ext4_clear_request_list(void); static struct inode *ext4_get_journal_inode(struct super_block *sb, unsigned int journal_inum); +static void ext4_netlink_send_info(struct super_block *sb, int ext4_errno); +static struct sock *ext4nl; /* * Lock ordering @@ -439,6 +445,42 @@ static bool system_going_down(void) || system_state == SYSTEM_RESTART; } +static void ext4_netlink_send_info(struct super_block *sb, int ext4_errno) +{ + int size; + sk_buff_data_t old_tail; + struct sk_buff *skb; + struct nlmsghdr *nlh; + struct ext4_err_msg *msg; + + if (ext4nl) { + size = NLMSG_SPACE(sizeof(struct ext4_err_msg)); + skb = alloc_skb(size, GFP_ATOMIC); + if (!skb) { + printk(KERN_ERR "Cannot alloc skb!"); + return; + } + old_tail = skb->tail; + nlh = nlmsg_put(skb, 0, 0, NLMSG_ERROR, size - sizeof(*nlh), 0); + if (!nlh) + goto nlmsg_failure; + msg = (struct ext4_err_msg *)NLMSG_DATA(nlh); + msg->magic = EXT4_ERROR_MAGIC; + memcpy(msg->s_id, sb->s_id, sizeof(sb->s_id)); + msg->s_flags = sb->s_flags; + msg->ext4_errno = ext4_errno; + nlh->nlmsg_len = skb->tail - old_tail; + NETLINK_CB(skb).portid = 0; + NETLINK_CB(skb).dst_group = NL_EXT4_ERROR_GROUP; + netlink_broadcast(ext4nl, skb, 0, NL_EXT4_ERROR_GROUP, + GFP_ATOMIC); + return; +nlmsg_failure: + if (skb) + kfree_skb(skb); + } +} + /* Deal with the reporting of failure conditions on a filesystem such as * inconsistencies detected or read IO failures. * @@ -469,6 +511,9 @@ static void ext4_handle_error(struct super_block *sb) if (journal) jbd2_journal_abort(journal, -EIO); } + + ext4_netlink_send_info(sb, 1); + /* * We force ERRORS_RO behavior when system is rebooting. Otherwise we * could panic during 'reboot -f' as the underlying device got already @@ -700,6 +745,7 @@ void __ext4_abort(struct super_block *sb, const char *function, if (EXT4_SB(sb)->s_journal) jbd2_journal_abort(EXT4_SB(sb)->s_journal, -EIO); save_error_info(sb, function, line); + ext4_netlink_send_info(sb, 2); } if (test_opt(sb, ERRORS_PANIC) && !system_going_down()) { if (EXT4_SB(sb)->s_journal && @@ -6063,6 +6109,7 @@ static inline int ext3_feature_set_ok(struct super_block *sb) static int __init ext4_init_fs(void) { int i, err; + struct netlink_kernel_cfg cfg = {.groups = NL_EXT4_ERROR_GROUP,}; ratelimit_state_init(&ext4_mount_msg_ratelimit, 30 * HZ, 64); ext4_li_info = NULL; @@ -6102,6 +6149,9 @@ static int __init ext4_init_fs(void) if (err) goto out; + ext4nl = netlink_kernel_create(&init_net, NETLINK_FILESYSTEM, &cfg); + if (!ext4nl) + printk(KERN_ERR "EXT4-fs: Cannot create netlink socket.\n"); return 0; out: unregister_as_ext2(); @@ -6133,6 +6183,7 @@ static void __exit ext4_exit_fs(void) ext4_exit_system_zone(); ext4_exit_pageio(); ext4_exit_es(); + netlink_kernel_release(ext4nl); } MODULE_AUTHOR("Remy Card, Stephen Tweedie, Andrew Morton, Andreas Dilger, Theodore Ts'o and others"); diff --git a/include/uapi/linux/netlink.h b/include/uapi/linux/netlink.h index 776bc92..ecaefee 100644 --- a/include/uapi/linux/netlink.h +++ b/include/uapi/linux/netlink.h @@ -29,6 +29,7 @@ #define NETLINK_RDMA 20 #define NETLINK_CRYPTO 21 /* Crypto layer */ #define NETLINK_SMC 22 /* SMC monitoring */ +#define NETLINK_FILESYSTEM 28 /* filesystem alarm*/ #define NETLINK_INET_DIAG NETLINK_SOCK_DIAG -- 1.8.3

1 0

[PATCH 1/3] Added ethtool_ops interface to query optical module information
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Chiqijun <chiqijun(a)huawei.com> driver inclusion category: feature bugzilla: 4472 ----------------------------------------------------------------------- Obtaining optical module information is time-consuming. If the user frequently acquires, it will affect the firmware to process other commands. To prevent such events, the firmware will actively report the optical module information to the driver when it is idle. Obtained from the driver cache information. When the user obtains through the command, directly obtain from the information saved by the driver. Signed-off-by: Chiqijun <chiqijun(a)huawei.com> Reviewed-by: Luoshaokai <luoshaokai(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/net/ethernet/huawei/hinic/hinic_ethtool.c | 77 ++++++++++ drivers/net/ethernet/huawei/hinic/hinic_hw.h | 54 +++++++ drivers/net/ethernet/huawei/hinic/hinic_hwdev.c | 168 +++++++++++++++++++++ drivers/net/ethernet/huawei/hinic/hinic_lld.c | 2 + .../ethernet/huawei/hinic/hinic_mgmt_interface.h | 12 ++ drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c | 139 ++++++++++++++++- drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h | 2 + drivers/net/ethernet/huawei/hinic/hinic_nic_io.c | 11 +- drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h | 3 + 9 files changed, 465 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c b/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c index f321365..78d32ea 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_ethtool.c @@ -2039,6 +2039,74 @@ static void hinic_diag_test(struct net_device *netdev, hinic_lp_test(netdev, eth_test, data, 0); } +#ifdef ETHTOOL_GMODULEEEPROM +static int hinic_get_module_info(struct net_device *netdev, + struct ethtool_modinfo *modinfo) +{ + struct hinic_nic_dev *nic_dev = netdev_priv(netdev); + u8 sfp_type; + u8 sfp_type_ext; + int err; + + err = hinic_get_sfp_type(nic_dev->hwdev, &sfp_type, &sfp_type_ext); + if (err) + return err; + + switch (sfp_type) { + case MODULE_TYPE_SFP: + modinfo->type = ETH_MODULE_SFF_8472; + modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN; + break; + case MODULE_TYPE_QSFP: + modinfo->type = ETH_MODULE_SFF_8436; + modinfo->eeprom_len = STD_SFP_INFO_MAX_SIZE; + break; + case MODULE_TYPE_QSFP_PLUS: + if (sfp_type_ext >= 0x3) { + modinfo->type = ETH_MODULE_SFF_8636; + modinfo->eeprom_len = STD_SFP_INFO_MAX_SIZE; + + } else { + modinfo->type = ETH_MODULE_SFF_8436; + modinfo->eeprom_len = STD_SFP_INFO_MAX_SIZE; + } + break; + case MODULE_TYPE_QSFP28: + modinfo->type = ETH_MODULE_SFF_8636; + modinfo->eeprom_len = STD_SFP_INFO_MAX_SIZE; + break; + default: + nicif_warn(nic_dev, drv, netdev, + "Optical module unknown: 0x%x\n", sfp_type); + return -EINVAL; + } + + return 0; +} + +static int hinic_get_module_eeprom(struct net_device *netdev, + struct ethtool_eeprom *ee, u8 *data) +{ + struct hinic_nic_dev *nic_dev = netdev_priv(netdev); + u8 sfp_data[STD_SFP_INFO_MAX_SIZE]; + u16 len; + int err; + + if (!ee->len || ((ee->len + ee->offset) > STD_SFP_INFO_MAX_SIZE)) + return -EINVAL; + + memset(data, 0, ee->len); + + err = hinic_get_sfp_eeprom(nic_dev->hwdev, sfp_data, &len); + if (err) + return err; + + memcpy(data, sfp_data + ee->offset, ee->len); + + return 0; +} +#endif /* ETHTOOL_GMODULEEEPROM */ + static int set_l4_rss_hash_ops(struct ethtool_rxnfc *cmd, struct nic_rss_type *rss_type) { @@ -2532,6 +2600,10 @@ static int hinic_set_rxfh_indir(struct net_device *netdev, const u32 *indir) #ifndef HAVE_RHEL6_ETHTOOL_OPS_EXT_STRUCT .get_channels = hinic_get_channels, .set_channels = hinic_set_channels, +#ifdef ETHTOOL_GMODULEEEPROM + .get_module_info = hinic_get_module_info, + .get_module_eeprom = hinic_get_module_eeprom, +#endif #ifndef NOT_HAVE_GET_RXFH_INDIR_SIZE .get_rxfh_indir_size = hinic_get_rxfh_indir_size, #endif @@ -2552,6 +2624,11 @@ static int hinic_set_rxfh_indir(struct net_device *netdev, const u32 *indir) .set_phys_id = hinic_set_phys_id, .get_channels = hinic_get_channels, .set_channels = hinic_set_channels, +#ifdef ETHTOOL_GMODULEEEPROM + .get_module_info = hinic_get_module_info, + .get_module_eeprom = hinic_get_module_eeprom, +#endif + #ifndef NOT_HAVE_GET_RXFH_INDIR_SIZE .get_rxfh_indir_size = hinic_get_rxfh_indir_size, #endif diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hw.h b/drivers/net/ethernet/huawei/hinic/hinic_hw.h index e489a00..d4becb6 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hw.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_hw.h @@ -264,6 +264,56 @@ struct hinic_init_para { #define INIT_SUCCESS 1 #define MAX_DRV_BUF_SIZE 4096 +struct hinic_cmd_get_light_module_abs { + u8 status; + u8 version; + u8 rsvd0[6]; + + u8 port_id; + u8 abs_status; /* 0:present, 1:absent */ + u8 rsv[2]; +}; + +#define MODULE_TYPE_SFP 0x3 +#define MODULE_TYPE_QSFP28 0x11 +#define MODULE_TYPE_QSFP 0x0C +#define MODULE_TYPE_QSFP_PLUS 0x0D + +#define SFP_INFO_MAX_SIZE 512 +struct hinic_cmd_get_sfp_qsfp_info { + u8 status; + u8 version; + u8 rsvd0[6]; + + u8 port_id; + u8 wire_type; + u16 out_len; + u8 sfp_qsfp_info[SFP_INFO_MAX_SIZE]; +}; + +#define STD_SFP_INFO_MAX_SIZE 640 +struct hinic_cmd_get_std_sfp_info { + u8 status; + u8 version; + u8 rsvd0[6]; + + u8 port_id; + u8 wire_type; + u16 eeprom_len; + u32 rsvd; + u8 sfp_info[STD_SFP_INFO_MAX_SIZE]; +}; + +#define HINIC_MAX_PORT_ID 4 + +struct hinic_port_routine_cmd { + int up_send_sfp_info; + int up_send_sfp_abs; + + struct hinic_cmd_get_sfp_qsfp_info sfp_info; + struct hinic_cmd_get_light_module_abs abs; +}; + struct card_node { struct list_head node; struct list_head func_list; @@ -282,6 +332,10 @@ struct card_node { bool disable_vf_load[HINIC_MAX_PF_NUM]; u32 vf_mbx_old_rand_id[MAX_FUNCTION_NUM]; u32 vf_mbx_rand_id[MAX_FUNCTION_NUM]; + struct hinic_port_routine_cmd rt_cmd[HINIC_MAX_PORT_ID]; + + /* mutex used for copy sfp info */ + struct mutex sfp_mutex; }; enum hinic_hwdev_init_state { diff --git a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c index 101b671..172be6c 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_hwdev.c @@ -59,6 +59,8 @@ #define HINIC_OK_FLAG_FAILED 1 +#define HINIC_GET_SFP_INFO_REAL_TIME 0x1 + #define HINIC_GLB_SO_RO_CFG_SHIFT 0x0 #define HINIC_GLB_SO_RO_CFG_MASK 0x1 #define HINIC_DISABLE_ORDER 0 @@ -913,6 +915,80 @@ int hinic_pf_msg_to_mgmt_sync(void *hwdev, enum hinic_mod_type mod, u8 cmd, return err; } +static bool is_sfp_info_cmd_cached(struct hinic_hwdev *hwdev, + enum hinic_mod_type mod, u8 cmd, + void *buf_in, u16 in_size, + void *buf_out, u16 *out_size) +{ + struct hinic_cmd_get_sfp_qsfp_info *sfp_info = NULL; + struct hinic_port_routine_cmd *rt_cmd = NULL; + struct card_node *chip_node = hwdev->chip_node; + + sfp_info = buf_in; + if (sfp_info->port_id >= HINIC_MAX_PORT_ID || + *out_size < sizeof(*sfp_info)) + return false; + + if (sfp_info->version == HINIC_GET_SFP_INFO_REAL_TIME) + return false; + + rt_cmd = &chip_node->rt_cmd[sfp_info->port_id]; + mutex_lock(&chip_node->sfp_mutex); + memcpy(buf_out, &rt_cmd->sfp_info, sizeof(*sfp_info)); + mutex_unlock(&chip_node->sfp_mutex); + + return true; +} + +static bool is_sfp_abs_cmd_cached(struct hinic_hwdev *hwdev, + enum hinic_mod_type mod, u8 cmd, + void *buf_in, u16 in_size, + void *buf_out, u16 *out_size) +{ + struct hinic_cmd_get_light_module_abs *abs = NULL; + struct hinic_port_routine_cmd *rt_cmd = NULL; + struct card_node *chip_node = hwdev->chip_node; + + abs = buf_in; + if (abs->port_id >= HINIC_MAX_PORT_ID || + *out_size < sizeof(*abs)) + return false; + + if (abs->version == HINIC_GET_SFP_INFO_REAL_TIME) + return false; + + rt_cmd = &chip_node->rt_cmd[abs->port_id]; + mutex_lock(&chip_node->sfp_mutex); + memcpy(buf_out, &rt_cmd->abs, sizeof(*abs)); + mutex_unlock(&chip_node->sfp_mutex); + + return true; +} + +static bool driver_processed_cmd(struct hinic_hwdev *hwdev, + enum hinic_mod_type mod, u8 cmd, + void *buf_in, u16 in_size, + void *buf_out, u16 *out_size) +{ + struct card_node *chip_node = hwdev->chip_node; + + if (mod == HINIC_MOD_L2NIC) { + if (cmd == HINIC_PORT_CMD_GET_SFP_INFO && + chip_node->rt_cmd->up_send_sfp_info) { + return is_sfp_info_cmd_cached(hwdev, mod, cmd, buf_in, + in_size, buf_out, + out_size); + } else if (cmd == HINIC_PORT_CMD_GET_SFP_ABS && + chip_node->rt_cmd->up_send_sfp_abs) { + return is_sfp_abs_cmd_cached(hwdev, mod, cmd, buf_in, + in_size, buf_out, + out_size); + } + } + + return false; +} + int hinic_msg_to_mgmt_sync(void *hwdev, enum hinic_mod_type mod, u8 cmd, void *buf_in, u16 in_size, void *buf_out, u16 *out_size, u32 timeout) @@ -949,6 +1025,10 @@ int hinic_msg_to_mgmt_sync(void *hwdev, enum hinic_mod_type mod, u8 cmd, err = __func_send_mbox(hwdev, mod, cmd, buf_in, in_size, buf_out, out_size, timeout); } else { + if (driver_processed_cmd(hwdev, mod, cmd, buf_in, in_size, + buf_out, out_size)) + return 0; + do { if (!hinic_get_mgmt_channel_status(hwdev) || !hinic_get_chip_present_flag(hwdev)) @@ -3202,6 +3282,8 @@ enum hinic_event_cmd { HINIC_EVENT_MGMT_PCIE_DFX, HINIC_EVENT_MCTP_HOST_INFO, HINIC_EVENT_MGMT_HEARTBEAT_EHD, + HINIC_EVENT_SFP_INFO_REPORT, + HINIC_EVENT_SFP_ABS_REPORT, HINIC_EVENT_MAX_TYPE, }; @@ -3284,6 +3366,16 @@ struct hinic_event_convert { .cmd = HINIC_MGMT_CMD_HEARTBEAT_EVENT, .event = HINIC_EVENT_MGMT_HEARTBEAT_EHD, }, + { + .mod = HINIC_MOD_L2NIC, + .cmd = HINIC_PORT_CMD_GET_SFP_INFO, + .event = HINIC_EVENT_SFP_INFO_REPORT, + }, + { + .mod = HINIC_MOD_L2NIC, + .cmd = HINIC_PORT_CMD_GET_SFP_ABS, + .event = HINIC_EVENT_SFP_ABS_REPORT, + }, }; static enum hinic_event_cmd __get_event_type(u8 mod, u8 cmd) @@ -3587,12 +3679,22 @@ static void module_status_event(struct hinic_hwdev *hwdev, struct hinic_cable_plug_event *plug_event; struct hinic_link_err_event *link_err; struct hinic_event_info event_info = {0}; + struct hinic_port_routine_cmd *rt_cmd; + struct card_node *chip_node = hwdev->chip_node; event_info.type = HINIC_EVENT_PORT_MODULE_EVENT; if (cmd == HINIC_EVENT_CABLE_PLUG) { plug_event = buf_in; + if (plug_event->port_id < HINIC_MAX_PORT_ID) { + rt_cmd = &chip_node->rt_cmd[plug_event->port_id]; + mutex_lock(&chip_node->sfp_mutex); + rt_cmd->up_send_sfp_abs = false; + rt_cmd->up_send_sfp_info = false; + mutex_unlock(&chip_node->sfp_mutex); + } + event_info.module_event.type = plug_event->plugged ? HINIC_PORT_MODULE_CABLE_PLUGGED : HINIC_PORT_MODULE_CABLE_UNPLUGGED; @@ -3733,6 +3835,64 @@ static void mgmt_watchdog_timeout_event_handler(struct hinic_hwdev *hwdev, queue_work(hwdev->workq, &hwdev->fault_work); } +static void port_sfp_info_event(struct hinic_hwdev *hwdev, void *buf_in, + u16 in_size, void *buf_out, u16 *out_size) +{ + struct hinic_cmd_get_sfp_qsfp_info *sfp_info = buf_in; + struct hinic_port_routine_cmd *rt_cmd; + struct card_node *chip_node = hwdev->chip_node; + + if (in_size != sizeof(*sfp_info)) { + sdk_err(hwdev->dev_hdl, "Invalid sfp info cmd, length: %d, should be %ld\n", + in_size, sizeof(*sfp_info)); + return; + } + + if (sfp_info->port_id >= HINIC_MAX_PORT_ID) { + sdk_err(hwdev->dev_hdl, "Invalid sfp port id: %d, max port is %d\n", + sfp_info->port_id, HINIC_MAX_PORT_ID - 1); + return; + } + + if (!chip_node->rt_cmd) + return; + + rt_cmd = &chip_node->rt_cmd[sfp_info->port_id]; + mutex_lock(&chip_node->sfp_mutex); + memcpy(&rt_cmd->sfp_info, sfp_info, sizeof(rt_cmd->sfp_info)); + rt_cmd->up_send_sfp_info = true; + mutex_unlock(&chip_node->sfp_mutex); +} + +static void port_sfp_abs_event(struct hinic_hwdev *hwdev, void *buf_in, + u16 in_size, void *buf_out, u16 *out_size) +{ + struct hinic_cmd_get_light_module_abs *sfp_abs = buf_in; + struct hinic_port_routine_cmd *rt_cmd; + struct card_node *chip_node = hwdev->chip_node; + + if (in_size != sizeof(*sfp_abs)) { + sdk_err(hwdev->dev_hdl, "Invalid sfp absent cmd, length: %d, should be %ld\n", + in_size, sizeof(*sfp_abs)); + return; + } + + if (sfp_abs->port_id >= HINIC_MAX_PORT_ID) { + sdk_err(hwdev->dev_hdl, "Invalid sfp port id: %d, max port is %d\n", + sfp_abs->port_id, HINIC_MAX_PORT_ID - 1); + return; + } + + if (!chip_node->rt_cmd) + return; + + rt_cmd = &chip_node->rt_cmd[sfp_abs->port_id]; + mutex_lock(&chip_node->sfp_mutex); + memcpy(&rt_cmd->abs, sfp_abs, sizeof(rt_cmd->abs)); + rt_cmd->up_send_sfp_abs = true; + mutex_unlock(&chip_node->sfp_mutex); +} + static void mgmt_reset_event_handler(struct hinic_hwdev *hwdev) { sdk_info(hwdev->dev_hdl, "Mgmt is reset\n"); @@ -4241,6 +4401,14 @@ static void _event_handler(struct hinic_hwdev *hwdev, enum hinic_event_cmd cmd, buf_out, out_size); break; + case HINIC_EVENT_SFP_INFO_REPORT: + port_sfp_info_event(hwdev, buf_in, in_size, buf_out, out_size); + break; + + case HINIC_EVENT_SFP_ABS_REPORT: + port_sfp_abs_event(hwdev, buf_in, in_size, buf_out, out_size); + break; + default: sdk_warn(hwdev->dev_hdl, "Unsupported event %d to process\n", cmd); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_lld.c b/drivers/net/ethernet/huawei/hinic/hinic_lld.c index e8e446a..3bf47ac 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_lld.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_lld.c @@ -1990,6 +1990,8 @@ static int alloc_chip_node(struct hinic_pcidev *pci_adapter) INIT_LIST_HEAD(&chip_node->func_list); pci_adapter->chip_node = chip_node; + mutex_init(&chip_node->sfp_mutex); + return 0; alloc_dbgtool_attr_file_err: diff --git a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h index 13bc351..3c3be6e 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_mgmt_interface.h @@ -680,6 +680,16 @@ struct hinic_capture_info { u32 data_vlan; }; +struct hinic_port_rt_cmd { + u8 status; + u8 version; + u8 rsvd0[6]; + + u8 pf_id; + u8 enable; + u8 rsvd1[6]; +}; + struct hinic_vf_dcb_state { u8 status; u8 version; @@ -737,6 +747,8 @@ struct hinic_set_link_follow { void hinic_unregister_vf_msg_handler(void *hwdev, u16 vf_id); +int hinic_set_port_routine_cmd_report(void *hwdev, bool enable); + int hinic_refresh_nic_cfg(void *hwdev, struct nic_port_info *port_info); int hinic_save_dcb_state(struct hinic_hwdev *hwdev, diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c index 07858be..6ddae36 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.c @@ -35,7 +35,6 @@ #include "hinic_nic.h" #include "hinic_mgmt_interface.h" #include "hinic_hwif.h" -#include "hinic_eqs.h" static unsigned char set_vf_link_state; module_param(set_vf_link_state, byte, 0444); @@ -3571,6 +3570,35 @@ int hinic_set_super_cqe_state(void *hwdev, bool enable) return 0; } +int hinic_set_port_routine_cmd_report(void *hwdev, bool enable) +{ + struct hinic_port_rt_cmd rt_cmd = { 0 }; + struct hinic_hwdev *dev = hwdev; + u16 out_size = sizeof(rt_cmd); + int err; + + if (!hwdev) + return -EINVAL; + + rt_cmd.pf_id = (u8)hinic_global_func_id(hwdev); + rt_cmd.enable = enable; + + err = l2nic_msg_to_mgmt_sync(hwdev, + HINIC_PORT_CMD_SET_PORT_REPORT, + &rt_cmd, sizeof(rt_cmd), &rt_cmd, + &out_size); + if (rt_cmd.status == HINIC_MGMT_CMD_UNSUPPORTED) { + nic_info(dev->dev_hdl, "Current firmware doesn't support to set port routine command report\n"); + } else if (rt_cmd.status || err || !out_size) { + nic_err(dev->dev_hdl, + "Failed to set port routine command report, err: %d, status: 0x%x, out size: 0x%x\n", + err, rt_cmd.status, out_size); + return -EFAULT; + } + + return 0; +} + int hinic_set_func_capture_en(void *hwdev, u16 func_id, bool cap_en) { struct hinic_hwdev *dev = hwdev; @@ -3889,3 +3917,112 @@ int hinic_disable_tx_promisc(void *hwdev) } return 0; } + +static bool hinic_if_sfp_absent(void *hwdev) +{ + struct card_node *chip_node = ((struct hinic_hwdev *)hwdev)->chip_node; + struct hinic_port_routine_cmd *rt_cmd; + struct hinic_cmd_get_light_module_abs sfp_abs = {0}; + u8 port_id = hinic_physical_port_id(hwdev); + u16 out_size = sizeof(sfp_abs); + int err; + bool sfp_abs_valid; + bool sfp_abs_status; + + rt_cmd = &chip_node->rt_cmd[port_id]; + mutex_lock(&chip_node->sfp_mutex); + sfp_abs_valid = rt_cmd->up_send_sfp_abs; + sfp_abs_status = (bool)rt_cmd->abs.abs_status; + if (sfp_abs_valid) { + mutex_unlock(&chip_node->sfp_mutex); + return sfp_abs_status; + } + mutex_unlock(&chip_node->sfp_mutex); + + sfp_abs.port_id = port_id; + err = l2nic_msg_to_mgmt_sync(hwdev, HINIC_PORT_CMD_GET_SFP_ABS, + &sfp_abs, sizeof(sfp_abs), &sfp_abs, + &out_size); + if (sfp_abs.status || err || !out_size) { + nic_err(((struct hinic_hwdev *)hwdev)->dev_hdl, + "Failed to get port%d sfp absent status, err: %d, status: 0x%x, out size: 0x%x\n", + port_id, err, sfp_abs.status, out_size); + return true; + } + + return ((sfp_abs.abs_status == 0) ? false : true); +} + +int hinic_get_sfp_eeprom(void *hwdev, u8 *data, u16 *len) +{ + struct hinic_cmd_get_std_sfp_info sfp_info = {0}; + u8 port_id; + u16 out_size = sizeof(sfp_info); + int err; + + if (!hwdev || !data || !len) + return -EINVAL; + + port_id = hinic_physical_port_id(hwdev); + if (port_id >= HINIC_MAX_PORT_ID) + return -EINVAL; + + if (hinic_if_sfp_absent(hwdev)) + return -ENXIO; + + sfp_info.port_id = port_id; + err = l2nic_msg_to_mgmt_sync(hwdev, HINIC_PORT_CMD_GET_STD_SFP_INFO, + &sfp_info, sizeof(sfp_info), &sfp_info, + &out_size); + if (sfp_info.status || err || !out_size) { + nic_err(((struct hinic_hwdev *)hwdev)->dev_hdl, + "Failed to get port%d sfp eeprom information, err: %d, status: 0x%x, out size: 0x%x\n", + port_id, err, sfp_info.status, out_size); + return -EIO; + } + + *len = min_t(u16, sfp_info.eeprom_len, STD_SFP_INFO_MAX_SIZE); + memcpy(data, sfp_info.sfp_info, STD_SFP_INFO_MAX_SIZE); + + return 0; +} + +int hinic_get_sfp_type(void *hwdev, u8 *data0, u8 *data1) +{ + struct card_node *chip_node = NULL; + struct hinic_port_routine_cmd *rt_cmd; + u8 sfp_data[STD_SFP_INFO_MAX_SIZE]; + u16 len; + u8 port_id; + int err; + + if (!hwdev || !data0 || !data1) + return -EINVAL; + + port_id = hinic_physical_port_id(hwdev); + if (port_id >= HINIC_MAX_PORT_ID) + return -EINVAL; + + if (hinic_if_sfp_absent(hwdev)) + return -ENXIO; + + chip_node = ((struct hinic_hwdev *)hwdev)->chip_node; + rt_cmd = &chip_node->rt_cmd[port_id]; + mutex_lock(&chip_node->sfp_mutex); + if (rt_cmd->up_send_sfp_info) { + *data0 = rt_cmd->sfp_info.sfp_qsfp_info[0]; + *data1 = rt_cmd->sfp_info.sfp_qsfp_info[1]; + mutex_unlock(&chip_node->sfp_mutex); + return 0; + } + mutex_unlock(&chip_node->sfp_mutex); + + err = hinic_get_sfp_eeprom(hwdev, sfp_data, &len); + if (err) + return err; + + *data0 = sfp_data[0]; + *data1 = sfp_data[1]; + + return 0; +} diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h index 395f6d8..eb060fb 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_cfg.h @@ -620,5 +620,7 @@ int hinic_add_hw_rqfilter(void *hwdev, struct hinic_rq_filter_info *filter_info); int hinic_del_hw_rqfilter(void *hwdev, struct hinic_rq_filter_info *filter_info); +int hinic_get_sfp_eeprom(void *hwdev, u8 *data, u16 *len); +int hinic_get_sfp_type(void *hwdev, u8 *data0, u8 *data1); #endif diff --git a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c index 8d13105..1cb4e26 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_nic_io.c @@ -829,13 +829,20 @@ int hinic_init_nic_hwdev(void *hwdev, u16 rx_buff_len) } } } - return 0; + + /* VFs don't set port routine command report */ + if (hinic_func_type(dev) != TYPE_VF) + /* Inform mgmt to send sfp's information to driver */ + err = hinic_set_port_routine_cmd_report(hwdev, true); + + return err; } EXPORT_SYMBOL(hinic_init_nic_hwdev); void hinic_free_nic_hwdev(void *hwdev) { - /* nothing to do for now */ + if (hinic_func_type(hwdev) != TYPE_VF) + hinic_set_port_routine_cmd_report(hwdev, false); } EXPORT_SYMBOL(hinic_free_nic_hwdev); diff --git a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h index 8730565..46464b1f 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h +++ b/drivers/net/ethernet/huawei/hinic/hinic_port_cmd.h @@ -146,6 +146,8 @@ enum hinic_port_cmd { HINIC_PORT_CMD_SET_PORT_LINK_STATUS = 0x76, HINIC_PORT_CMD_SET_CGE_PAUSE_TIME_CFG = 0x77, + HINIC_PORT_CMD_SET_PORT_REPORT = 0x7B, + HINIC_PORT_CMD_LINK_STATUS_REPORT = 0xa0, HINIC_PORT_CMD_SET_LOSSLESS_ETH = 0xa3, @@ -200,6 +202,7 @@ enum hinic_port_cmd { HINIC_PORT_CMD_SET_LINK_FOLLOW = 0xF8, HINIC_PORT_CMD_SET_VF_MAX_MIN_RATE = 0xF9, HINIC_PORT_CMD_SET_RXQ_LRO_ADPT = 0xFA, + HINIC_PORT_CMD_GET_SFP_ABS = 0xFB, HINIC_PORT_CMD_Q_FILTER = 0xFC, HINIC_PORT_CMD_TCAM_FILTER = 0xFE, HINIC_PORT_CMD_SET_VLAN_FILTER = 0xFF, -- 1.8.3

1 2

[PATCH] Revert "consolemap: Fix a memory leaking bug in drivers/tty/vt/consolemap.c"
by Yang Yingliang 01 Jun '20

01 Jun '20

From: Ben Hutchings <ben(a)decadent.org.uk> mainline inclusion from mainline-v5.3-rc1 commit 15b3cd8ef46ad1b100e0d3c7e38774f330726820 category: bugfix bugzilla: 34617 CVE: CVE-2019-12379 --------------------------- This reverts commit 84ecc2f6eb1cb12e6d44818f94fa49b50f06e6ac. con_insert_unipair() is working with a sparse 3-dimensional array: - p->uni_pgdir[] is the top layer - p1 points to a middle layer - p2 points to a bottom layer If it needs to allocate a new middle layer, and then fails to allocate a new bottom layer, it would previously free only p2, and now it frees both p1 and p2. But since the new middle layer was already registered in the top layer, it was not leaked. However, if it looks up an *existing* middle layer and then fails to allocate a bottom layer, it now frees both p1 and p2 but does *not* free any other bottom layers under p1. So it *introduces* a memory leak. The error path also cleared the wrong index in p->uni_pgdir[], introducing a use-after-free. Signed-off-by: Ben Hutchings <ben(a)decadent.org.uk> Fixes: 84ecc2f6eb1c ("consolemap: Fix a memory leaking bug in drivers/tty/vt/consolemap.c") Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Zheng Bin <zhengbin13(a)huawei.com> Reviewed-by: Jason Yan <yanaijie(a)huawei.com> Signed-off-by: Yang Yingliang <yangyingliang(a)huawei.com> --- drivers/tty/vt/consolemap.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/tty/vt/consolemap.c b/drivers/tty/vt/consolemap.c index 814d1b7..7c7ada0 100644 --- a/drivers/tty/vt/consolemap.c +++ b/drivers/tty/vt/consolemap.c @@ -489,11 +489,7 @@ static int con_unify_unimap(struct vc_data *conp, struct uni_pagedir *p) p2 = p1[n = (unicode >> 6) & 0x1f]; if (!p2) { p2 = p1[n] = kmalloc_array(64, sizeof(u16), GFP_KERNEL); - if (!p2) { - kfree(p1); - p->uni_pgdir[n] = NULL; - return -ENOMEM; - } + if (!p2) return -ENOMEM; memset(p2, 0xff, 64*sizeof(u16)); /* No glyphs for the characters (yet) */ } -- 1.8.3

1 0