From a249cc8bc2e2fed680047d326eb9a50756724198 Mon Sep 17 00:00:00 2001 From: Aurelien Aptel Date: Thu, 4 Mar 2021 17:42:21 +0000 Subject: [PATCH 01/22] cifs: fix credit accounting for extra channel With multichannel, operations like the queries from "ls -lR" can cause all credits to be used and errors to be returned since max_credits was not being set correctly on the secondary channels and thus the client was requesting 0 credits incorrectly in some cases (which can lead to not having enough credits to perform any operation on that channel). Signed-off-by: Aurelien Aptel CC: # v5.8+ Reviewed-by: Shyam Prasad N Signed-off-by: Steve French --- fs/cifs/connect.c | 10 +++++----- fs/cifs/sess.c | 1 + 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 112692300fb6..68642e3d4270 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -1429,6 +1429,11 @@ cifs_get_tcp_session(struct smb3_fs_context *ctx) tcp_ses->min_offload = ctx->min_offload; tcp_ses->tcpStatus = CifsNeedNegotiate; + if ((ctx->max_credits < 20) || (ctx->max_credits > 60000)) + tcp_ses->max_credits = SMB2_MAX_CREDITS_AVAILABLE; + else + tcp_ses->max_credits = ctx->max_credits; + tcp_ses->nr_targets = 1; tcp_ses->ignore_signature = ctx->ignore_signature; /* thread spawned, put it on the list */ @@ -2832,11 +2837,6 @@ static int mount_get_conns(struct smb3_fs_context *ctx, struct cifs_sb_info *cif *nserver = server; - if ((ctx->max_credits < 20) || (ctx->max_credits > 60000)) - server->max_credits = SMB2_MAX_CREDITS_AVAILABLE; - else - server->max_credits = ctx->max_credits; - /* get a reference to a SMB session */ ses = cifs_get_smb_ses(server, ctx); if (IS_ERR(ses)) { diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 183a3a868d7b..63d517b9f2ff 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -230,6 +230,7 @@ cifs_ses_add_channel(struct cifs_sb_info *cifs_sb, struct cifs_ses *ses, ctx.noautotune = ses->server->noautotune; ctx.sockopt_tcp_nodelay = ses->server->tcp_nodelay; ctx.echo_interval = ses->server->echo_interval / HZ; + ctx.max_credits = ses->server->max_credits; /* * This will be used for encoding/decoding user/domain/pw From 88fd98a2306755b965e4f4567f84e73db3b6738c Mon Sep 17 00:00:00 2001 From: Aurelien Aptel Date: Thu, 4 Mar 2021 17:51:48 +0000 Subject: [PATCH 02/22] cifs: ask for more credit on async read/write code paths When doing a large read or write workload we only very gradually increase the number of credits which can cause problems with parallelizing large i/o (I/O ramps up more slowly than it should for large read/write workloads) especially with multichannel when the number of credits on the secondary channels starts out low (e.g. less than about 130) or when recovering after server throttled back the number of credit. Signed-off-by: Aurelien Aptel Reviewed-by: Shyam Prasad N Signed-off-by: Steve French --- fs/cifs/smb2pdu.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/cifs/smb2pdu.c b/fs/cifs/smb2pdu.c index 4bbb6126b14d..2199a9bfae8f 100644 --- a/fs/cifs/smb2pdu.c +++ b/fs/cifs/smb2pdu.c @@ -4041,8 +4041,7 @@ smb2_async_readv(struct cifs_readdata *rdata) if (rdata->credits.value > 0) { shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(rdata->bytes, SMB2_MAX_BUFFER_SIZE)); - shdr->CreditRequest = - cpu_to_le16(le16_to_cpu(shdr->CreditCharge) + 1); + shdr->CreditRequest = cpu_to_le16(le16_to_cpu(shdr->CreditCharge) + 8); rc = adjust_credits(server, &rdata->credits, rdata->bytes); if (rc) @@ -4348,8 +4347,7 @@ smb2_async_writev(struct cifs_writedata *wdata, if (wdata->credits.value > 0) { shdr->CreditCharge = cpu_to_le16(DIV_ROUND_UP(wdata->bytes, SMB2_MAX_BUFFER_SIZE)); - shdr->CreditRequest = - cpu_to_le16(le16_to_cpu(shdr->CreditCharge) + 1); + shdr->CreditRequest = cpu_to_le16(le16_to_cpu(shdr->CreditCharge) + 8); rc = adjust_credits(server, &wdata->credits, wdata->bytes); if (rc) From eba8e1af5a61e61e5d77e1dfe1e8e20735ebc9c6 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Tue, 23 Feb 2021 19:52:20 +0100 Subject: [PATCH 03/22] s390/time,idle: get rid of unsigned long long Get rid of unsigned long long, and use unsigned long instead everywhere. The usage of unsigned long long is a leftover from 31 bit kernel support. Signed-off-by: Heiko Carstens --- arch/s390/include/asm/idle.h | 12 ++++++------ arch/s390/include/asm/timex.h | 36 +++++++++++++++++------------------ arch/s390/kernel/idle.c | 12 ++++++------ arch/s390/kernel/time.c | 28 +++++++++++++-------------- arch/s390/kvm/interrupt.c | 2 +- drivers/s390/cio/device_fsm.c | 2 +- 6 files changed, 46 insertions(+), 46 deletions(-) diff --git a/arch/s390/include/asm/idle.h b/arch/s390/include/asm/idle.h index b04f6a794cdf..5cea629c548e 100644 --- a/arch/s390/include/asm/idle.h +++ b/arch/s390/include/asm/idle.h @@ -14,12 +14,12 @@ struct s390_idle_data { seqcount_t seqcount; - unsigned long long idle_count; - unsigned long long idle_time; - unsigned long long clock_idle_enter; - unsigned long long clock_idle_exit; - unsigned long long timer_idle_enter; - unsigned long long timer_idle_exit; + unsigned long idle_count; + unsigned long idle_time; + unsigned long clock_idle_enter; + unsigned long clock_idle_exit; + unsigned long timer_idle_enter; + unsigned long timer_idle_exit; unsigned long mt_cycles_enter[8]; }; diff --git a/arch/s390/include/asm/timex.h b/arch/s390/include/asm/timex.h index c4e23e925665..f6326c6d2abe 100644 --- a/arch/s390/include/asm/timex.h +++ b/arch/s390/include/asm/timex.h @@ -98,10 +98,10 @@ extern unsigned char ptff_function_mask[16]; /* Query TOD offset result */ struct ptff_qto { - unsigned long long physical_clock; - unsigned long long tod_offset; - unsigned long long logical_tod_offset; - unsigned long long tod_epoch_difference; + unsigned long physical_clock; + unsigned long tod_offset; + unsigned long logical_tod_offset; + unsigned long tod_epoch_difference; } __packed; static inline int ptff_query(unsigned int nr) @@ -151,9 +151,9 @@ struct ptff_qui { rc; \ }) -static inline unsigned long long local_tick_disable(void) +static inline unsigned long local_tick_disable(void) { - unsigned long long old; + unsigned long old; old = S390_lowcore.clock_comparator; S390_lowcore.clock_comparator = clock_comparator_max; @@ -161,7 +161,7 @@ static inline unsigned long long local_tick_disable(void) return old; } -static inline void local_tick_enable(unsigned long long comp) +static inline void local_tick_enable(unsigned long comp) { S390_lowcore.clock_comparator = comp; set_clock_comparator(S390_lowcore.clock_comparator); @@ -169,9 +169,9 @@ static inline void local_tick_enable(unsigned long long comp) #define CLOCK_TICK_RATE 1193180 /* Underlying HZ */ -typedef unsigned long long cycles_t; +typedef unsigned long cycles_t; -static inline unsigned long long get_tod_clock(void) +static inline unsigned long get_tod_clock(void) { union tod_clock clk; @@ -179,10 +179,10 @@ static inline unsigned long long get_tod_clock(void) return clk.tod; } -static inline unsigned long long get_tod_clock_fast(void) +static inline unsigned long get_tod_clock_fast(void) { #ifdef CONFIG_HAVE_MARCH_Z9_109_FEATURES - unsigned long long clk; + unsigned long clk; asm volatile("stckf %0" : "=Q" (clk) : : "cc"); return clk; @@ -208,9 +208,9 @@ extern union tod_clock tod_clock_base; * Therefore preemption must be disabled, otherwise the returned * value is not guaranteed to be monotonic. */ -static inline unsigned long long get_tod_clock_monotonic(void) +static inline unsigned long get_tod_clock_monotonic(void) { - unsigned long long tod; + unsigned long tod; preempt_disable_notrace(); tod = get_tod_clock() - tod_clock_base.tod; @@ -237,7 +237,7 @@ static inline unsigned long long get_tod_clock_monotonic(void) * -> ns = (th * 125) + ((tl * 125) >> 9); * */ -static inline unsigned long long tod_to_ns(unsigned long long todval) +static inline unsigned long tod_to_ns(unsigned long todval) { return ((todval >> 9) * 125) + (((todval & 0x1ff) * 125) >> 9); } @@ -249,10 +249,10 @@ static inline unsigned long long tod_to_ns(unsigned long long todval) * * Returns: true if a is later than b */ -static inline int tod_after(unsigned long long a, unsigned long long b) +static inline int tod_after(unsigned long a, unsigned long b) { if (MACHINE_HAS_SCC) - return (long long) a > (long long) b; + return (long) a > (long) b; return a > b; } @@ -263,10 +263,10 @@ static inline int tod_after(unsigned long long a, unsigned long long b) * * Returns: true if a is later than b */ -static inline int tod_after_eq(unsigned long long a, unsigned long long b) +static inline int tod_after_eq(unsigned long a, unsigned long b) { if (MACHINE_HAS_SCC) - return (long long) a >= (long long) b; + return (long) a >= (long) b; return a >= b; } diff --git a/arch/s390/kernel/idle.c b/arch/s390/kernel/idle.c index 812073ea073e..4bf1ee293f2b 100644 --- a/arch/s390/kernel/idle.c +++ b/arch/s390/kernel/idle.c @@ -47,7 +47,7 @@ void account_idle_time_irq(void) void arch_cpu_idle(void) { struct s390_idle_data *idle = this_cpu_ptr(&s390_idle); - unsigned long long idle_time; + unsigned long idle_time; unsigned long psw_mask; /* Wait for external, I/O or machine check interrupt. */ @@ -73,7 +73,7 @@ static ssize_t show_idle_count(struct device *dev, struct device_attribute *attr, char *buf) { struct s390_idle_data *idle = &per_cpu(s390_idle, dev->id); - unsigned long long idle_count; + unsigned long idle_count; unsigned int seq; do { @@ -82,14 +82,14 @@ static ssize_t show_idle_count(struct device *dev, if (READ_ONCE(idle->clock_idle_enter)) idle_count++; } while (read_seqcount_retry(&idle->seqcount, seq)); - return sprintf(buf, "%llu\n", idle_count); + return sprintf(buf, "%lu\n", idle_count); } DEVICE_ATTR(idle_count, 0444, show_idle_count, NULL); static ssize_t show_idle_time(struct device *dev, struct device_attribute *attr, char *buf) { - unsigned long long now, idle_time, idle_enter, idle_exit, in_idle; + unsigned long now, idle_time, idle_enter, idle_exit, in_idle; struct s390_idle_data *idle = &per_cpu(s390_idle, dev->id); unsigned int seq; @@ -109,14 +109,14 @@ static ssize_t show_idle_time(struct device *dev, } } idle_time += in_idle; - return sprintf(buf, "%llu\n", idle_time >> 12); + return sprintf(buf, "%lu\n", idle_time >> 12); } DEVICE_ATTR(idle_time_us, 0444, show_idle_time, NULL); u64 arch_cpu_idle_time(int cpu) { struct s390_idle_data *idle = &per_cpu(s390_idle, cpu); - unsigned long long now, idle_enter, idle_exit, in_idle; + unsigned long now, idle_enter, idle_exit, in_idle; unsigned int seq; do { diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c index 06bcfa636638..165da961f901 100644 --- a/arch/s390/kernel/time.c +++ b/arch/s390/kernel/time.c @@ -68,10 +68,10 @@ EXPORT_SYMBOL(s390_epoch_delta_notifier); unsigned char ptff_function_mask[16]; -static unsigned long long lpar_offset; -static unsigned long long initial_leap_seconds; -static unsigned long long tod_steering_end; -static long long tod_steering_delta; +static unsigned long lpar_offset; +static unsigned long initial_leap_seconds; +static unsigned long tod_steering_end; +static long tod_steering_delta; /* * Get time offsets with PTFF @@ -96,7 +96,7 @@ void __init time_early_init(void) /* get initial leap seconds */ if (ptff_query(PTFF_QUI) && ptff(&qui, sizeof(qui), PTFF_QUI) == 0) - initial_leap_seconds = (unsigned long long) + initial_leap_seconds = (unsigned long) ((long) qui.old_leap * 4096000000L); } @@ -222,7 +222,7 @@ void __init read_persistent_wall_and_boot_offset(struct timespec64 *wall_time, static u64 read_tod_clock(struct clocksource *cs) { - unsigned long long now, adj; + unsigned long now, adj; preempt_disable(); /* protect from changes to steering parameters */ now = get_tod_clock(); @@ -362,7 +362,7 @@ static inline int check_sync_clock(void) * Apply clock delta to the global data structures. * This is called once on the CPU that performed the clock sync. */ -static void clock_sync_global(unsigned long long delta) +static void clock_sync_global(unsigned long delta) { unsigned long now, adj; struct ptff_qto qto; @@ -378,7 +378,7 @@ static void clock_sync_global(unsigned long long delta) -(adj >> 15) : (adj >> 15); tod_steering_delta += delta; if ((abs(tod_steering_delta) >> 48) != 0) - panic("TOD clock sync offset %lli is too large to drift\n", + panic("TOD clock sync offset %li is too large to drift\n", tod_steering_delta); tod_steering_end = now + (abs(tod_steering_delta) << 15); vdso_data->arch_data.tod_steering_end = tod_steering_end; @@ -394,7 +394,7 @@ static void clock_sync_global(unsigned long long delta) * Apply clock delta to the per-CPU data structures of this CPU. * This is called for each online CPU after the call to clock_sync_global. */ -static void clock_sync_local(unsigned long long delta) +static void clock_sync_local(unsigned long delta) { /* Add the delta to the clock comparator. */ if (S390_lowcore.clock_comparator != clock_comparator_max) { @@ -418,7 +418,7 @@ static void __init time_init_wq(void) struct clock_sync_data { atomic_t cpus; int in_sync; - unsigned long long clock_delta; + unsigned long clock_delta; }; /* @@ -538,7 +538,7 @@ static int stpinfo_valid(void) static int stp_sync_clock(void *data) { struct clock_sync_data *sync = data; - unsigned long long clock_delta, flags; + u64 clock_delta, flags; static int first; int rc; @@ -720,8 +720,8 @@ static ssize_t ctn_id_show(struct device *dev, mutex_lock(&stp_mutex); if (stpinfo_valid()) - ret = sprintf(buf, "%016llx\n", - *(unsigned long long *) stp_info.ctnid); + ret = sprintf(buf, "%016lx\n", + *(unsigned long *) stp_info.ctnid); mutex_unlock(&stp_mutex); return ret; } @@ -794,7 +794,7 @@ static ssize_t leap_seconds_scheduled_show(struct device *dev, if (!stzi.lsoib.p) return sprintf(buf, "0,0\n"); - return sprintf(buf, "%llu,%d\n", + return sprintf(buf, "%lu,%d\n", tod_to_ns(stzi.lsoib.nlsout - TOD_UNIX_EPOCH) / NSEC_PER_SEC, stzi.lsoib.nlso - stzi.lsoib.also); } diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c index e3183bd05910..d548d60caed2 100644 --- a/arch/s390/kvm/interrupt.c +++ b/arch/s390/kvm/interrupt.c @@ -1287,7 +1287,7 @@ static u64 __calculate_sltime(struct kvm_vcpu *vcpu) /* already expired? */ if (cputm >> 63) return 0; - return min(sltime, tod_to_ns(cputm)); + return min_t(u64, sltime, tod_to_ns(cputm)); } } else if (cpu_timer_interrupts_enabled(vcpu)) { sltime = kvm_s390_get_cpu_timer(vcpu); diff --git a/drivers/s390/cio/device_fsm.c b/drivers/s390/cio/device_fsm.c index 6420b197bb05..05e136cfb8be 100644 --- a/drivers/s390/cio/device_fsm.c +++ b/drivers/s390/cio/device_fsm.c @@ -47,7 +47,7 @@ static void ccw_timeout_log(struct ccw_device *cdev) orb = &private->orb; cc = stsch(sch->schid, &schib); - printk(KERN_WARNING "cio: ccw device timeout occurred at %llx, " + printk(KERN_WARNING "cio: ccw device timeout occurred at %lx, " "device information:\n", get_tod_clock()); printk(KERN_WARNING "cio: orb:\n"); print_hex_dump(KERN_WARNING, "cio: ", DUMP_PREFIX_NONE, 16, 1, From f9d8cbf33e9fceee671a49760cdcfa4be6a55102 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Wed, 17 Feb 2021 12:47:41 +0100 Subject: [PATCH 04/22] s390/topology: remove always false if check The cpumask being checked in cpu_group_map() must have at least one cpu set; therefore remove the check. Reviewed-by: Alexander Gordeev Signed-off-by: Heiko Carstens --- arch/s390/kernel/topology.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/arch/s390/kernel/topology.c b/arch/s390/kernel/topology.c index e7ce447651b9..bfcc327acc6b 100644 --- a/arch/s390/kernel/topology.c +++ b/arch/s390/kernel/topology.c @@ -76,8 +76,6 @@ static void cpu_group_map(cpumask_t *dst, struct mask_info *info, unsigned int c } info = info->next; } - if (cpumask_empty(&mask)) - cpumask_copy(&mask, cpumask_of(cpu)); break; case TOPOLOGY_MODE_PACKAGE: cpumask_copy(&mask, cpu_present_mask); From c41b20de1a7c9a41ceab293a6f08927312ada679 Mon Sep 17 00:00:00 2001 From: Thomas Richter Date: Thu, 25 Feb 2021 14:28:52 +0100 Subject: [PATCH 05/22] s390/cpumf: remove 60 seconds read limit Remove the 60 seconds read interval limit. Do not impose any limit at all and allow read of complete counter sets. Signed-off-by: Thomas Richter Acked-by: Heiko Carstens Signed-off-by: Heiko Carstens --- arch/s390/kernel/perf_cpum_cf_diag.c | 18 ++---------------- 1 file changed, 2 insertions(+), 16 deletions(-) diff --git a/arch/s390/kernel/perf_cpum_cf_diag.c b/arch/s390/kernel/perf_cpum_cf_diag.c index db4877bbb9aa..5eebc912df5a 100644 --- a/arch/s390/kernel/perf_cpum_cf_diag.c +++ b/arch/s390/kernel/perf_cpum_cf_diag.c @@ -29,9 +29,7 @@ #include #define CF_DIAG_CTRSET_DEF 0xfeef /* Counter set header mark */ -#define CF_DIAG_MIN_INTERVAL 60 /* Minimum counter set read */ /* interval in seconds */ -static unsigned long cf_diag_interval = CF_DIAG_MIN_INTERVAL; static unsigned int cf_diag_cpu_speed; static debug_info_t *cf_diag_dbg; @@ -729,7 +727,6 @@ static DEFINE_MUTEX(cf_diag_ctrset_mutex); static struct cf_diag_ctrset { unsigned long ctrset; /* Bit mask of counter set to read */ cpumask_t mask; /* CPU mask to read from */ - time64_t lastread; /* Epoch counter set last read */ } cf_diag_ctrset; static void cf_diag_ctrset_clear(void) @@ -866,27 +863,16 @@ static int cf_diag_all_read(unsigned long arg) { struct cf_diag_call_on_cpu_parm p; cpumask_var_t mask; - time64_t now; - int rc = 0; + int rc; debug_sprintf_event(cf_diag_dbg, 5, "%s\n", __func__); if (!alloc_cpumask_var(&mask, GFP_KERNEL)) return -ENOMEM; - now = ktime_get_seconds(); - if (cf_diag_ctrset.lastread + cf_diag_interval > now) { - debug_sprintf_event(cf_diag_dbg, 5, "%s now %lld " - " lastread %lld\n", __func__, now, - cf_diag_ctrset.lastread); - rc = -EAGAIN; - goto out; - } else { - cf_diag_ctrset.lastread = now; - } + p.sets = cf_diag_ctrset.ctrset; cpumask_and(mask, &cf_diag_ctrset.mask, cpu_online_mask); on_each_cpu_mask(mask, cf_diag_cpu_read, &p, 1); rc = cf_diag_all_copy(arg, mask); -out: free_cpumask_var(mask); debug_sprintf_event(cf_diag_dbg, 5, "%s rc %d\n", __func__, rc); return rc; From 7449ca87312a5b0390b765be65a126e6e5451026 Mon Sep 17 00:00:00 2001 From: Alexander Egorenkov Date: Fri, 26 Feb 2021 11:21:05 +0100 Subject: [PATCH 06/22] s390/zcore: no need to check return value of debugfs_create functions When calling debugfs functions, there is no need to ever check the return value. The function can work or not, but the code logic should never do something different based on this. See commit 7dd541a3fb34 ("s390: no need to check return value of debugfs_create functions"). Signed-off-by: Alexander Egorenkov Signed-off-by: Heiko Carstens --- drivers/s390/char/zcore.c | 18 +----------------- 1 file changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/s390/char/zcore.c b/drivers/s390/char/zcore.c index 1515fdc3c1ab..1bf507c00eb6 100644 --- a/drivers/s390/char/zcore.c +++ b/drivers/s390/char/zcore.c @@ -293,28 +293,12 @@ static int __init zcore_init(void) goto fail; zcore_dir = debugfs_create_dir("zcore" , NULL); - if (!zcore_dir) { - rc = -ENOMEM; - goto fail; - } zcore_reipl_file = debugfs_create_file("reipl", S_IRUSR, zcore_dir, NULL, &zcore_reipl_fops); - if (!zcore_reipl_file) { - rc = -ENOMEM; - goto fail_dir; - } zcore_hsa_file = debugfs_create_file("hsa", S_IRUSR|S_IWUSR, zcore_dir, NULL, &zcore_hsa_fops); - if (!zcore_hsa_file) { - rc = -ENOMEM; - goto fail_reipl_file; - } - return 0; -fail_reipl_file: - debugfs_remove(zcore_reipl_file); -fail_dir: - debugfs_remove(zcore_dir); + return 0; fail: diag308(DIAG308_REL_HSA, NULL); return rc; From dabdfac0e85c8c1e811b10c08742f49285e78a17 Mon Sep 17 00:00:00 2001 From: Alexander Egorenkov Date: Thu, 25 Feb 2021 14:28:52 +0100 Subject: [PATCH 07/22] s390/zcore: release dump save area on restart or power down The zFCP/NVMe standalone dumper is supposed to release the dump save area resource as soon as possible but might fail to do so, for instance, if it crashes. To avoid this situation, register a reboot notifier and ensure the dump save area resource is released on reboot or power down. Signed-off-by: Alexander Egorenkov Reviewed-by: Heiko Carstens Signed-off-by: Heiko Carstens --- drivers/s390/char/zcore.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) diff --git a/drivers/s390/char/zcore.c b/drivers/s390/char/zcore.c index 1bf507c00eb6..bd3c724bf695 100644 --- a/drivers/s390/char/zcore.c +++ b/drivers/s390/char/zcore.c @@ -15,6 +15,7 @@ #include #include #include +#include #include #include @@ -238,6 +239,28 @@ static int __init zcore_reipl_init(void) return 0; } +static int zcore_reboot_and_on_panic_handler(struct notifier_block *self, + unsigned long event, + void *data) +{ + if (hsa_available) + release_hsa(); + + return NOTIFY_OK; +} + +static struct notifier_block zcore_reboot_notifier = { + .notifier_call = zcore_reboot_and_on_panic_handler, + /* we need to be notified before reipl and kdump */ + .priority = INT_MAX, +}; + +static struct notifier_block zcore_on_panic_notifier = { + .notifier_call = zcore_reboot_and_on_panic_handler, + /* we need to be notified before reipl and kdump */ + .priority = INT_MAX, +}; + static int __init zcore_init(void) { unsigned char arch; @@ -298,6 +321,9 @@ static int __init zcore_init(void) zcore_hsa_file = debugfs_create_file("hsa", S_IRUSR|S_IWUSR, zcore_dir, NULL, &zcore_hsa_fops); + register_reboot_notifier(&zcore_reboot_notifier); + atomic_notifier_chain_register(&panic_notifier_list, &zcore_on_panic_notifier); + return 0; fail: diag308(DIAG308_REL_HSA, NULL); From 46b635b6abcf2ee106d36f2e84e942f56030c8f5 Mon Sep 17 00:00:00 2001 From: Thomas Richter Date: Fri, 26 Feb 2021 15:00:56 +0100 Subject: [PATCH 08/22] s390/cpumf: rename header file to hwctrset.h Signed-off-by: Thomas Richter Suggested-by: Hendrick Brueckner Acked-by: Heiko Carstens Signed-off-by: Heiko Carstens --- arch/s390/include/uapi/asm/{perf_cpum_cf_diag.h => hwctrset.h} | 0 arch/s390/kernel/perf_cpum_cf_diag.c | 2 +- 2 files changed, 1 insertion(+), 1 deletion(-) rename arch/s390/include/uapi/asm/{perf_cpum_cf_diag.h => hwctrset.h} (100%) diff --git a/arch/s390/include/uapi/asm/perf_cpum_cf_diag.h b/arch/s390/include/uapi/asm/hwctrset.h similarity index 100% rename from arch/s390/include/uapi/asm/perf_cpum_cf_diag.h rename to arch/s390/include/uapi/asm/hwctrset.h diff --git a/arch/s390/kernel/perf_cpum_cf_diag.c b/arch/s390/kernel/perf_cpum_cf_diag.c index 5eebc912df5a..bc302b86ce28 100644 --- a/arch/s390/kernel/perf_cpum_cf_diag.c +++ b/arch/s390/kernel/perf_cpum_cf_diag.c @@ -26,7 +26,7 @@ #include #include -#include +#include #define CF_DIAG_CTRSET_DEF 0xfeef /* Counter set header mark */ /* interval in seconds */ From 51c44babdc19aaf882e1213325a0ba291573308f Mon Sep 17 00:00:00 2001 From: Wang Qing Date: Mon, 1 Mar 2021 20:01:33 +0800 Subject: [PATCH 09/22] s390/cio: return -EFAULT if copy_to_user() fails The copy_to_user() function returns the number of bytes remaining to be copied, but we want to return -EFAULT if the copy doesn't complete. Fixes: e01bcdd61320 ("vfio: ccw: realize VFIO_DEVICE_GET_REGION_INFO ioctl") Signed-off-by: Wang Qing Signed-off-by: Heiko Carstens Link: https://lore.kernel.org/r/1614600093-13992-1-git-send-email-wangqing@vivo.com Signed-off-by: Heiko Carstens --- drivers/s390/cio/vfio_ccw_ops.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/s390/cio/vfio_ccw_ops.c b/drivers/s390/cio/vfio_ccw_ops.c index 68106be4ba7a..557d0b888cd3 100644 --- a/drivers/s390/cio/vfio_ccw_ops.c +++ b/drivers/s390/cio/vfio_ccw_ops.c @@ -543,7 +543,7 @@ static ssize_t vfio_ccw_mdev_ioctl(struct mdev_device *mdev, if (ret) return ret; - return copy_to_user((void __user *)arg, &info, minsz); + return copy_to_user((void __user *)arg, &info, minsz) ? -EFAULT : 0; } case VFIO_DEVICE_GET_REGION_INFO: { @@ -561,7 +561,7 @@ static ssize_t vfio_ccw_mdev_ioctl(struct mdev_device *mdev, if (ret) return ret; - return copy_to_user((void __user *)arg, &info, minsz); + return copy_to_user((void __user *)arg, &info, minsz) ? -EFAULT : 0; } case VFIO_DEVICE_GET_IRQ_INFO: { From 942df4be7ab40195e2a839e9de81951a5862bc5b Mon Sep 17 00:00:00 2001 From: Wang Qing Date: Mon, 1 Mar 2021 20:08:21 +0800 Subject: [PATCH 10/22] s390/crypto: return -EFAULT if copy_to_user() fails The copy_to_user() function returns the number of bytes remaining to be copied, but we want to return -EFAULT if the copy doesn't complete. Fixes: e06670c5fe3b ("s390: vfio-ap: implement VFIO_DEVICE_GET_INFO ioctl") Signed-off-by: Wang Qing Reviewed-by: Tony Krowiak Signed-off-by: Heiko Carstens Link: https://lore.kernel.org/r/1614600502-16714-1-git-send-email-wangqing@vivo.com Signed-off-by: Heiko Carstens --- drivers/s390/crypto/vfio_ap_ops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/s390/crypto/vfio_ap_ops.c b/drivers/s390/crypto/vfio_ap_ops.c index 41fc2e4135fe..1ffdd411201c 100644 --- a/drivers/s390/crypto/vfio_ap_ops.c +++ b/drivers/s390/crypto/vfio_ap_ops.c @@ -1286,7 +1286,7 @@ static int vfio_ap_mdev_get_device_info(unsigned long arg) info.num_regions = 0; info.num_irqs = 0; - return copy_to_user((void __user *)arg, &info, minsz); + return copy_to_user((void __user *)arg, &info, minsz) ? -EFAULT : 0; } static ssize_t vfio_ap_mdev_ioctl(struct mdev_device *mdev, From 1c0a9c7997325ef7a8f71fca2e1e6091e757c94b Mon Sep 17 00:00:00 2001 From: Jiapeng Chong Date: Thu, 25 Feb 2021 14:15:36 +0800 Subject: [PATCH 11/22] s390/cpumf: remove unneeded semicolon Fix the following coccicheck warnings: ./arch/s390/kernel/perf_cpum_cf.c:272:2-3: Unneeded semicolon. Reported-by: Abaci Robot Signed-off-by: Jiapeng Chong Signed-off-by: Heiko Carstens Link: https://lore.kernel.org/r/1614233736-87331-1-git-send-email-jiapeng.chong@linux.alibaba.com Signed-off-by: Heiko Carstens --- arch/s390/kernel/perf_cpum_cf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/s390/kernel/perf_cpum_cf.c b/arch/s390/kernel/perf_cpum_cf.c index 0eb1d1cc53a8..b3beef64d3d4 100644 --- a/arch/s390/kernel/perf_cpum_cf.c +++ b/arch/s390/kernel/perf_cpum_cf.c @@ -269,7 +269,7 @@ static int __hw_perf_event_init(struct perf_event *event, unsigned int type) case CPUMF_CTR_SET_MAX: /* The counter could not be associated to a counter set */ return -EINVAL; - }; + } /* Initialize for using the CPU-measurement counter facility */ if (!atomic_inc_not_zero(&num_events)) { From db232eb42c21b7bca8e46c7e74676754b8fe06ea Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Mon, 24 Aug 2020 21:56:16 -0700 Subject: [PATCH 12/22] s390/tty3270: avoid comma separated statements Use semicolons and braces. Signed-off-by: Joe Perches Signed-off-by: Heiko Carstens Link: https://lore.kernel.org/r/9988babd9cca4ac841961d9f0bbf5e49caa87659.1598331149.git.joe@perches.com Signed-off-by: Heiko Carstens --- drivers/s390/char/tty3270.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/s390/char/tty3270.c b/drivers/s390/char/tty3270.c index 15692449a1c3..307a80f85c07 100644 --- a/drivers/s390/char/tty3270.c +++ b/drivers/s390/char/tty3270.c @@ -424,8 +424,10 @@ tty3270_update(struct timer_list *t) * last output position matches the start address * of this line. */ - if (s->string[1] == sba[0] && s->string[2] == sba[1]) - str += 3, len -= 3; + if (s->string[1] == sba[0] && s->string[2] == sba[1]) { + str += 3; + len -= 3; + } if (raw3270_request_add_data(wrq, str, len) != 0) break; list_del_init(&s->update); From d9c48a948d29bcb22f4fe61a81b718ef6de561a0 Mon Sep 17 00:00:00 2001 From: Eric Farman Date: Mon, 1 Mar 2021 19:33:24 +0100 Subject: [PATCH 13/22] s390/cio: return -EFAULT if copy_to_user() fails Fixes: 120e214e504f ("vfio: ccw: realize VFIO_DEVICE_G(S)ET_IRQ_INFO ioctls") Signed-off-by: Eric Farman Signed-off-by: Heiko Carstens --- drivers/s390/cio/vfio_ccw_ops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/s390/cio/vfio_ccw_ops.c b/drivers/s390/cio/vfio_ccw_ops.c index 557d0b888cd3..767ac41686fe 100644 --- a/drivers/s390/cio/vfio_ccw_ops.c +++ b/drivers/s390/cio/vfio_ccw_ops.c @@ -582,7 +582,7 @@ static ssize_t vfio_ccw_mdev_ioctl(struct mdev_device *mdev, if (info.count == -1) return -EINVAL; - return copy_to_user((void __user *)arg, &info, minsz); + return copy_to_user((void __user *)arg, &info, minsz) ? -EFAULT : 0; } case VFIO_DEVICE_SET_IRQS: { From d0ed78e1780eb3738f9c106fbaff6a1181017cd3 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Mon, 1 Mar 2021 21:02:49 +0100 Subject: [PATCH 14/22] s390,alpha: make TMPFS_INODE64 available again Both s390 and alpha have been switched to 64-bit ino_t with commit 96c0a6a72d18 ("s390,alpha: switch to 64-bit ino_t"). Therefore enable TMPFS_INODE64 for both architectures again. Cc: Richard Henderson Cc: Ivan Kokshaysky Link: https://lore.kernel.org/linux-mm/YCV7QiyoweJwvN+m@osiris/ Signed-off-by: Heiko Carstens --- fs/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/Kconfig b/fs/Kconfig index 462253ae483a..a55bda4233bb 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -203,7 +203,7 @@ config TMPFS_XATTR config TMPFS_INODE64 bool "Use 64-bit ino_t by default in tmpfs" - depends on TMPFS && 64BIT && !(S390 || ALPHA) + depends on TMPFS && 64BIT default n help tmpfs has historically used only inode numbers as wide as an unsigned From d50aa69d36be43fa8927fd8ef305c4af88b6b450 Mon Sep 17 00:00:00 2001 From: Heiko Carstens Date: Mon, 1 Mar 2021 21:48:16 +0100 Subject: [PATCH 15/22] s390: update defconfigs Signed-off-by: Heiko Carstens --- arch/s390/configs/debug_defconfig | 16 +++------------- arch/s390/configs/defconfig | 11 ++--------- arch/s390/configs/zfcpdump_defconfig | 1 - 3 files changed, 5 insertions(+), 23 deletions(-) diff --git a/arch/s390/configs/debug_defconfig b/arch/s390/configs/debug_defconfig index 02056b024091..dc0b69058ac4 100644 --- a/arch/s390/configs/debug_defconfig +++ b/arch/s390/configs/debug_defconfig @@ -275,9 +275,9 @@ CONFIG_IP_VS_DH=m CONFIG_IP_VS_SH=m CONFIG_IP_VS_SED=m CONFIG_IP_VS_NQ=m +CONFIG_IP_VS_TWOS=m CONFIG_IP_VS_FTP=m CONFIG_IP_VS_PE_SIP=m -CONFIG_NF_TABLES_IPV4=y CONFIG_NFT_FIB_IPV4=m CONFIG_NF_TABLES_ARP=y CONFIG_IP_NF_IPTABLES=m @@ -298,7 +298,6 @@ CONFIG_IP_NF_SECURITY=m CONFIG_IP_NF_ARPTABLES=m CONFIG_IP_NF_ARPFILTER=m CONFIG_IP_NF_ARP_MANGLE=m -CONFIG_NF_TABLES_IPV6=y CONFIG_NFT_FIB_IPV6=m CONFIG_IP6_NF_IPTABLES=m CONFIG_IP6_NF_MATCH_AH=m @@ -481,7 +480,6 @@ CONFIG_NLMON=m # CONFIG_NET_VENDOR_AQUANTIA is not set # CONFIG_NET_VENDOR_ARC is not set # CONFIG_NET_VENDOR_ATHEROS is not set -# CONFIG_NET_VENDOR_AURORA is not set # CONFIG_NET_VENDOR_BROADCOM is not set # CONFIG_NET_VENDOR_BROCADE is not set # CONFIG_NET_VENDOR_CADENCE is not set @@ -581,7 +579,6 @@ CONFIG_VIRTIO_BALLOON=m CONFIG_VIRTIO_INPUT=y CONFIG_VHOST_NET=m CONFIG_VHOST_VSOCK=m -# CONFIG_SURFACE_PLATFORMS is not set CONFIG_S390_CCW_IOMMU=y CONFIG_S390_AP_IOMMU=y CONFIG_EXT4_FS=y @@ -635,6 +632,7 @@ CONFIG_NTFS_RW=y CONFIG_PROC_KCORE=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y +CONFIG_TMPFS_INODE64=y CONFIG_HUGETLBFS=y CONFIG_CONFIGFS_FS=m CONFIG_ECRYPT_FS=m @@ -714,12 +712,8 @@ CONFIG_CRYPTO_VMAC=m CONFIG_CRYPTO_CRC32=m CONFIG_CRYPTO_BLAKE2S=m CONFIG_CRYPTO_MICHAEL_MIC=m -CONFIG_CRYPTO_RMD128=m CONFIG_CRYPTO_RMD160=m -CONFIG_CRYPTO_RMD256=m -CONFIG_CRYPTO_RMD320=m CONFIG_CRYPTO_SHA3=m -CONFIG_CRYPTO_TGR192=m CONFIG_CRYPTO_WP512=m CONFIG_CRYPTO_AES_TI=m CONFIG_CRYPTO_ANUBIS=m @@ -731,7 +725,6 @@ CONFIG_CRYPTO_CAST6=m CONFIG_CRYPTO_DES=m CONFIG_CRYPTO_FCRYPT=m CONFIG_CRYPTO_KHAZAD=m -CONFIG_CRYPTO_SALSA20=m CONFIG_CRYPTO_SEED=m CONFIG_CRYPTO_SERPENT=m CONFIG_CRYPTO_SM4=m @@ -796,12 +789,9 @@ CONFIG_DEBUG_OBJECTS_RCU_HEAD=y CONFIG_DEBUG_OBJECTS_PERCPU_COUNTER=y CONFIG_SLUB_DEBUG_ON=y CONFIG_SLUB_STATS=y -CONFIG_DEBUG_KMEMLEAK=y -CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y CONFIG_DEBUG_STACK_USAGE=y CONFIG_DEBUG_VM=y CONFIG_DEBUG_VM_VMACACHE=y -CONFIG_DEBUG_VM_RB=y CONFIG_DEBUG_VM_PGFLAGS=y CONFIG_DEBUG_MEMORY_INIT=y CONFIG_MEMORY_NOTIFIER_ERROR_INJECT=m @@ -838,6 +828,7 @@ CONFIG_BPF_KPROBE_OVERRIDE=y CONFIG_HIST_TRIGGERS=y CONFIG_FTRACE_STARTUP_TEST=y # CONFIG_EVENT_TRACE_STARTUP_TEST is not set +CONFIG_DEBUG_ENTRY=y CONFIG_NOTIFIER_ERROR_INJECTION=m CONFIG_NETDEV_NOTIFIER_ERROR_INJECT=m CONFIG_FAULT_INJECTION=y @@ -861,4 +852,3 @@ CONFIG_PERCPU_TEST=m CONFIG_ATOMIC64_SELFTEST=y CONFIG_TEST_BITOPS=m CONFIG_TEST_BPF=m -CONFIG_DEBUG_ENTRY=y diff --git a/arch/s390/configs/defconfig b/arch/s390/configs/defconfig index bac721a501da..320379da96d9 100644 --- a/arch/s390/configs/defconfig +++ b/arch/s390/configs/defconfig @@ -266,9 +266,9 @@ CONFIG_IP_VS_DH=m CONFIG_IP_VS_SH=m CONFIG_IP_VS_SED=m CONFIG_IP_VS_NQ=m +CONFIG_IP_VS_TWOS=m CONFIG_IP_VS_FTP=m CONFIG_IP_VS_PE_SIP=m -CONFIG_NF_TABLES_IPV4=y CONFIG_NFT_FIB_IPV4=m CONFIG_NF_TABLES_ARP=y CONFIG_IP_NF_IPTABLES=m @@ -289,7 +289,6 @@ CONFIG_IP_NF_SECURITY=m CONFIG_IP_NF_ARPTABLES=m CONFIG_IP_NF_ARPFILTER=m CONFIG_IP_NF_ARP_MANGLE=m -CONFIG_NF_TABLES_IPV6=y CONFIG_NFT_FIB_IPV6=m CONFIG_IP6_NF_IPTABLES=m CONFIG_IP6_NF_MATCH_AH=m @@ -473,7 +472,6 @@ CONFIG_NLMON=m # CONFIG_NET_VENDOR_AQUANTIA is not set # CONFIG_NET_VENDOR_ARC is not set # CONFIG_NET_VENDOR_ATHEROS is not set -# CONFIG_NET_VENDOR_AURORA is not set # CONFIG_NET_VENDOR_BROADCOM is not set # CONFIG_NET_VENDOR_BROCADE is not set # CONFIG_NET_VENDOR_CADENCE is not set @@ -573,7 +571,6 @@ CONFIG_VIRTIO_BALLOON=m CONFIG_VIRTIO_INPUT=y CONFIG_VHOST_NET=m CONFIG_VHOST_VSOCK=m -# CONFIG_SURFACE_PLATFORMS is not set CONFIG_S390_CCW_IOMMU=y CONFIG_S390_AP_IOMMU=y CONFIG_EXT4_FS=y @@ -623,6 +620,7 @@ CONFIG_NTFS_RW=y CONFIG_PROC_KCORE=y CONFIG_TMPFS=y CONFIG_TMPFS_POSIX_ACL=y +CONFIG_TMPFS_INODE64=y CONFIG_HUGETLBFS=y CONFIG_CONFIGFS_FS=m CONFIG_ECRYPT_FS=m @@ -703,12 +701,8 @@ CONFIG_CRYPTO_VMAC=m CONFIG_CRYPTO_CRC32=m CONFIG_CRYPTO_BLAKE2S=m CONFIG_CRYPTO_MICHAEL_MIC=m -CONFIG_CRYPTO_RMD128=m CONFIG_CRYPTO_RMD160=m -CONFIG_CRYPTO_RMD256=m -CONFIG_CRYPTO_RMD320=m CONFIG_CRYPTO_SHA3=m -CONFIG_CRYPTO_TGR192=m CONFIG_CRYPTO_WP512=m CONFIG_CRYPTO_AES_TI=m CONFIG_CRYPTO_ANUBIS=m @@ -720,7 +714,6 @@ CONFIG_CRYPTO_CAST6=m CONFIG_CRYPTO_DES=m CONFIG_CRYPTO_FCRYPT=m CONFIG_CRYPTO_KHAZAD=m -CONFIG_CRYPTO_SALSA20=m CONFIG_CRYPTO_SEED=m CONFIG_CRYPTO_SERPENT=m CONFIG_CRYPTO_SM4=m diff --git a/arch/s390/configs/zfcpdump_defconfig b/arch/s390/configs/zfcpdump_defconfig index acf982a2ae4c..039fd499b4a9 100644 --- a/arch/s390/configs/zfcpdump_defconfig +++ b/arch/s390/configs/zfcpdump_defconfig @@ -61,7 +61,6 @@ CONFIG_RAW_DRIVER=y # CONFIG_HID is not set # CONFIG_VIRTIO_MENU is not set # CONFIG_VHOST_MENU is not set -# CONFIG_SURFACE_PLATFORMS is not set # CONFIG_IOMMU_SUPPORT is not set # CONFIG_DNOTIFY is not set # CONFIG_INOTIFY_USER is not set From 78c7cccaab9d5f9ead44579d79dd7d13a05aec7e Mon Sep 17 00:00:00 2001 From: Alexander Egorenkov Date: Fri, 26 Feb 2021 09:40:47 +0100 Subject: [PATCH 16/22] s390: remove IBM_PARTITION and CONFIGFS_FS from zfcpdump defconfig Remove by zfcpdump unused CONFIG_IBM_PARTITION and CONFIG_CONFIGFS_FS. Signed-off-by: Alexander Egorenkov Reviewed-by: Steffen Maier Signed-off-by: Heiko Carstens --- arch/s390/configs/zfcpdump_defconfig | 2 -- 1 file changed, 2 deletions(-) diff --git a/arch/s390/configs/zfcpdump_defconfig b/arch/s390/configs/zfcpdump_defconfig index 039fd499b4a9..76123a4b26ab 100644 --- a/arch/s390/configs/zfcpdump_defconfig +++ b/arch/s390/configs/zfcpdump_defconfig @@ -26,7 +26,6 @@ CONFIG_CRASH_DUMP=y # CONFIG_SECCOMP is not set # CONFIG_GCC_PLUGINS is not set CONFIG_PARTITION_ADVANCED=y -CONFIG_IBM_PARTITION=y # CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS is not set # CONFIG_COMPACTION is not set # CONFIG_MIGRATION is not set @@ -64,7 +63,6 @@ CONFIG_RAW_DRIVER=y # CONFIG_IOMMU_SUPPORT is not set # CONFIG_DNOTIFY is not set # CONFIG_INOTIFY_USER is not set -CONFIG_CONFIGFS_FS=y # CONFIG_MISC_FILESYSTEMS is not set # CONFIG_NETWORK_FILESYSTEMS is not set CONFIG_LSM="yama,loadpin,safesetid,integrity" From ee2e3f50629f17b0752b55b2566c15ce8dafb557 Mon Sep 17 00:00:00 2001 From: Christian Brauner Date: Sat, 6 Mar 2021 11:10:10 +0100 Subject: [PATCH 17/22] mount: fix mounting of detached mounts onto targets that reside on shared mounts Creating a series of detached mounts, attaching them to the filesystem, and unmounting them can be used to trigger an integer overflow in ns->mounts causing the kernel to block any new mounts in count_mounts() and returning ENOSPC because it falsely assumes that the maximum number of mounts in the mount namespace has been reached, i.e. it thinks it can't fit the new mounts into the mount namespace anymore. Depending on the number of mounts in your system, this can be reproduced on any kernel that supportes open_tree() and move_mount() by compiling and running the following program: /* SPDX-License-Identifier: LGPL-2.1+ */ #define _GNU_SOURCE #include #include #include #include #include #include #include #include #include #include #include #include #include /* open_tree() */ #ifndef OPEN_TREE_CLONE #define OPEN_TREE_CLONE 1 #endif #ifndef OPEN_TREE_CLOEXEC #define OPEN_TREE_CLOEXEC O_CLOEXEC #endif #ifndef __NR_open_tree #if defined __alpha__ #define __NR_open_tree 538 #elif defined _MIPS_SIM #if _MIPS_SIM == _MIPS_SIM_ABI32 /* o32 */ #define __NR_open_tree 4428 #endif #if _MIPS_SIM == _MIPS_SIM_NABI32 /* n32 */ #define __NR_open_tree 6428 #endif #if _MIPS_SIM == _MIPS_SIM_ABI64 /* n64 */ #define __NR_open_tree 5428 #endif #elif defined __ia64__ #define __NR_open_tree (428 + 1024) #else #define __NR_open_tree 428 #endif #endif /* move_mount() */ #ifndef MOVE_MOUNT_F_EMPTY_PATH #define MOVE_MOUNT_F_EMPTY_PATH 0x00000004 /* Empty from path permitted */ #endif #ifndef __NR_move_mount #if defined __alpha__ #define __NR_move_mount 539 #elif defined _MIPS_SIM #if _MIPS_SIM == _MIPS_SIM_ABI32 /* o32 */ #define __NR_move_mount 4429 #endif #if _MIPS_SIM == _MIPS_SIM_NABI32 /* n32 */ #define __NR_move_mount 6429 #endif #if _MIPS_SIM == _MIPS_SIM_ABI64 /* n64 */ #define __NR_move_mount 5429 #endif #elif defined __ia64__ #define __NR_move_mount (428 + 1024) #else #define __NR_move_mount 429 #endif #endif static inline int sys_open_tree(int dfd, const char *filename, unsigned int flags) { return syscall(__NR_open_tree, dfd, filename, flags); } static inline int sys_move_mount(int from_dfd, const char *from_pathname, int to_dfd, const char *to_pathname, unsigned int flags) { return syscall(__NR_move_mount, from_dfd, from_pathname, to_dfd, to_pathname, flags); } static bool is_shared_mountpoint(const char *path) { bool shared = false; FILE *f = NULL; char *line = NULL; int i; size_t len = 0; f = fopen("/proc/self/mountinfo", "re"); if (!f) return 0; while (getline(&line, &len, f) > 0) { char *slider1, *slider2; for (slider1 = line, i = 0; slider1 && i < 4; i++) slider1 = strchr(slider1 + 1, ' '); if (!slider1) continue; slider2 = strchr(slider1 + 1, ' '); if (!slider2) continue; *slider2 = '\0'; if (strcmp(slider1 + 1, path) == 0) { /* This is the path. Is it shared? */ slider1 = strchr(slider2 + 1, ' '); if (slider1 && strstr(slider1, "shared:")) { shared = true; break; } } } fclose(f); free(line); return shared; } static void usage(void) { const char *text = "mount-new [--recursive] \n"; fprintf(stderr, "%s", text); _exit(EXIT_SUCCESS); } #define exit_usage(format, ...) \ ({ \ fprintf(stderr, format "\n", ##__VA_ARGS__); \ usage(); \ }) #define exit_log(format, ...) \ ({ \ fprintf(stderr, format "\n", ##__VA_ARGS__); \ exit(EXIT_FAILURE); \ }) static const struct option longopts[] = { {"help", no_argument, 0, 'a'}, { NULL, no_argument, 0, 0 }, }; int main(int argc, char *argv[]) { int exit_code = EXIT_SUCCESS, index = 0; int dfd, fd_tree, new_argc, ret; char *base_dir; char *const *new_argv; char target[PATH_MAX]; while ((ret = getopt_long_only(argc, argv, "", longopts, &index)) != -1) { switch (ret) { case 'a': /* fallthrough */ default: usage(); } } new_argv = &argv[optind]; new_argc = argc - optind; if (new_argc < 1) exit_usage("Missing base directory\n"); base_dir = new_argv[0]; if (*base_dir != '/') exit_log("Please specify an absolute path"); /* Ensure that target is a shared mountpoint. */ if (!is_shared_mountpoint(base_dir)) exit_log("Please ensure that \"%s\" is a shared mountpoint", base_dir); dfd = open(base_dir, O_RDONLY | O_DIRECTORY | O_CLOEXEC); if (dfd < 0) exit_log("%m - Failed to open base directory \"%s\"", base_dir); ret = mkdirat(dfd, "detached-move-mount", 0755); if (ret < 0) exit_log("%m - Failed to create required temporary directories"); ret = snprintf(target, sizeof(target), "%s/detached-move-mount", base_dir); if (ret < 0 || (size_t)ret >= sizeof(target)) exit_log("%m - Failed to assemble target path"); /* * Having a mount table with 10000 mounts is already quite excessive * and shoult account even for weird test systems. */ for (size_t i = 0; i < 10000; i++) { fd_tree = sys_open_tree(dfd, "detached-move-mount", OPEN_TREE_CLONE | OPEN_TREE_CLOEXEC | AT_EMPTY_PATH); if (fd_tree < 0) { fprintf(stderr, "%m - Failed to open %d(detached-move-mount)", dfd); exit_code = EXIT_FAILURE; break; } ret = sys_move_mount(fd_tree, "", dfd, "detached-move-mount", MOVE_MOUNT_F_EMPTY_PATH); if (ret < 0) { if (errno == ENOSPC) fprintf(stderr, "%m - Buggy mount counting"); else fprintf(stderr, "%m - Failed to attach mount to %d(detached-move-mount)", dfd); exit_code = EXIT_FAILURE; break; } close(fd_tree); ret = umount2(target, MNT_DETACH); if (ret < 0) { fprintf(stderr, "%m - Failed to unmount %s", target); exit_code = EXIT_FAILURE; break; } } (void)unlinkat(dfd, "detached-move-mount", AT_REMOVEDIR); close(dfd); exit(exit_code); } and wait for the kernel to refuse any new mounts by returning ENOSPC. How many iterations are needed depends on the number of mounts in your system. Assuming you have something like 50 mounts on a standard system it should be almost instantaneous. The root cause of this is that detached mounts aren't handled correctly when source and target mount are identical and reside on a shared mount causing a broken mount tree where the detached source itself is propagated which propagation prevents for regular bind-mounts and new mounts. This ultimately leads to a miscalculation of the number of mounts in the mount namespace. Detached mounts created via open_tree(fd, path, OPEN_TREE_CLONE) are essentially like an unattached new mount, or an unattached bind-mount. They can then later on be attached to the filesystem via move_mount() which calls into attach_recursive_mount(). Part of attaching it to the filesystem is making sure that mounts get correctly propagated in case the destination mountpoint is MS_SHARED, i.e. is a shared mountpoint. This is done by calling into propagate_mnt() which walks the list of peers calling propagate_one() on each mount in this list making sure it receives the propagation event. The propagate_one() functions thereby skips both new mounts and bind mounts to not propagate them "into themselves". Both are identified by checking whether the mount is already attached to any mount namespace in mnt->mnt_ns. The is what the IS_MNT_NEW() helper is responsible for. However, detached mounts have an anonymous mount namespace attached to them stashed in mnt->mnt_ns which means that IS_MNT_NEW() doesn't realize they need to be skipped causing the mount to propagate "into itself" breaking the mount table and causing a disconnect between the number of mounts recorded as being beneath or reachable from the target mountpoint and the number of mounts actually recorded/counted in ns->mounts ultimately causing an overflow which in turn prevents any new mounts via the ENOSPC issue. So teach propagation to handle detached mounts by making it aware of them. I've been tracking this issue down for the last couple of days and then verifying that the fix is correct by unmounting everything in my current mount table leaving only /proc and /sys mounted and running the reproducer above overnight verifying the number of mounts counted in ns->mounts. With this fix the counts are correct and the ENOSPC issue can't be reproduced. This change will only have an effect on mounts created with the new mount API since detached mounts cannot be created with the old mount API so regressions are extremely unlikely. Link: https://lore.kernel.org/r/20210306101010.243666-1-christian.brauner@ubuntu.com Fixes: 2db154b3ea8e ("vfs: syscall: Add move_mount(2) to move mounts around") Cc: David Howells Cc: Al Viro Cc: linux-fsdevel@vger.kernel.org Cc: Reviewed-by: Christoph Hellwig Signed-off-by: Christian Brauner --- fs/pnode.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/pnode.h b/fs/pnode.h index 26f74e092bd9..988f1aa9b02a 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -12,7 +12,7 @@ #define IS_MNT_SHARED(m) ((m)->mnt.mnt_flags & MNT_SHARED) #define IS_MNT_SLAVE(m) ((m)->mnt_master) -#define IS_MNT_NEW(m) (!(m)->mnt_ns) +#define IS_MNT_NEW(m) (!(m)->mnt_ns || is_anon_ns((m)->mnt_ns)) #define CLEAR_MNT_SHARED(m) ((m)->mnt.mnt_flags &= ~MNT_SHARED) #define IS_MNT_UNBINDABLE(m) ((m)->mnt.mnt_flags & MNT_UNBINDABLE) #define IS_MNT_MARKED(m) ((m)->mnt.mnt_flags & MNT_MARKED) From bf1bc694b6b0cf49756cb06f8f38501b9b2c5527 Mon Sep 17 00:00:00 2001 From: Paulo Alcantara Date: Mon, 8 Mar 2021 12:00:47 -0300 Subject: [PATCH 18/22] cifs: print MIDs in decimal notation The MIDs are mostly printed as decimal, so let's make it consistent. Signed-off-by: Paulo Alcantara (SUSE) Reviewed-by: Aurelien Aptel Reviewed-by: Ronnie Sahlberg Signed-off-by: Steve French --- fs/cifs/cifs_debug.c | 2 +- fs/cifs/connect.c | 4 ++-- fs/cifs/smb2misc.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index 3aedc484e440..88a7958170ee 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -207,7 +207,7 @@ static int cifs_debug_files_proc_show(struct seq_file *m, void *v) from_kuid(&init_user_ns, cfile->uid), cfile->dentry); #ifdef CONFIG_CIFS_DEBUG2 - seq_printf(m, " 0x%llx\n", cfile->fid.mid); + seq_printf(m, " %llu\n", cfile->fid.mid); #else seq_printf(m, "\n"); #endif /* CIFS_DEBUG2 */ diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 68642e3d4270..eec8a2052da2 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -741,7 +741,7 @@ static void clean_demultiplex_info(struct TCP_Server_Info *server) spin_lock(&GlobalMid_Lock); list_for_each_safe(tmp, tmp2, &server->pending_mid_q) { mid_entry = list_entry(tmp, struct mid_q_entry, qhead); - cifs_dbg(FYI, "Clearing mid 0x%llx\n", mid_entry->mid); + cifs_dbg(FYI, "Clearing mid %llu\n", mid_entry->mid); kref_get(&mid_entry->refcount); mid_entry->mid_state = MID_SHUTDOWN; list_move(&mid_entry->qhead, &dispose_list); @@ -752,7 +752,7 @@ static void clean_demultiplex_info(struct TCP_Server_Info *server) /* now walk dispose list and issue callbacks */ list_for_each_safe(tmp, tmp2, &dispose_list) { mid_entry = list_entry(tmp, struct mid_q_entry, qhead); - cifs_dbg(FYI, "Callback mid 0x%llx\n", mid_entry->mid); + cifs_dbg(FYI, "Callback mid %llu\n", mid_entry->mid); list_del_init(&mid_entry->qhead); mid_entry->callback(mid_entry); cifs_mid_q_entry_release(mid_entry); diff --git a/fs/cifs/smb2misc.c b/fs/cifs/smb2misc.c index 60d4bd1eae2b..6e0ea19e710b 100644 --- a/fs/cifs/smb2misc.c +++ b/fs/cifs/smb2misc.c @@ -767,7 +767,7 @@ smb2_cancelled_close_fid(struct work_struct *work) int rc; if (cancelled->mid) - cifs_tcon_dbg(VFS, "Close unmatched open for MID:%llx\n", + cifs_tcon_dbg(VFS, "Close unmatched open for MID:%llu\n", cancelled->mid); else cifs_tcon_dbg(VFS, "Close interrupted close\n"); From e3d100eae44b42f309c1366efb8397368f1cf8ed Mon Sep 17 00:00:00 2001 From: Paulo Alcantara Date: Mon, 8 Mar 2021 12:00:48 -0300 Subject: [PATCH 19/22] cifs: change noisy error message to FYI A customer has reported that their dmesg were being flooded by CIFS: VFS: \\server Cancelling wait for mid xxx cmd: a CIFS: VFS: \\server Cancelling wait for mid yyy cmd: b CIFS: VFS: \\server Cancelling wait for mid zzz cmd: c because some processes that were performing statfs(2) on the share had been interrupted due to their automount setup when certain users logged in and out. Change it to FYI as they should be mostly informative rather than error messages. Signed-off-by: Paulo Alcantara (SUSE) Reviewed-by: Aurelien Aptel Reviewed-by: Ronnie Sahlberg Signed-off-by: Steve French --- fs/cifs/transport.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index e90a1d1380b0..15b72d2e8713 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -1207,7 +1207,7 @@ compound_send_recv(const unsigned int xid, struct cifs_ses *ses, } if (rc != 0) { for (; i < num_rqst; i++) { - cifs_server_dbg(VFS, "Cancelling wait for mid %llu cmd: %d\n", + cifs_server_dbg(FYI, "Cancelling wait for mid %llu cmd: %d\n", midQ[i]->mid, le16_to_cpu(midQ[i]->command)); send_cancel(server, &rqst[i], midQ[i]); spin_lock(&GlobalMid_Lock); From 14302ee3301b3a77b331cc14efb95bf7184c73cc Mon Sep 17 00:00:00 2001 From: Paulo Alcantara Date: Mon, 8 Mar 2021 12:00:49 -0300 Subject: [PATCH 20/22] cifs: return proper error code in statfs(2) In cifs_statfs(), if server->ops->queryfs is not NULL, then we should use its return value rather than always returning 0. Instead, use rc variable as it is properly set to 0 in case there is no server->ops->queryfs. Signed-off-by: Paulo Alcantara (SUSE) Reviewed-by: Aurelien Aptel Reviewed-by: Ronnie Sahlberg CC: Signed-off-by: Steve French --- fs/cifs/cifsfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index d43e935d2df4..099ad9f3660b 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -290,7 +290,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf) rc = server->ops->queryfs(xid, tcon, cifs_sb, buf); free_xid(xid); - return 0; + return rc; } static long cifs_fallocate(struct file *file, int mode, loff_t off, loff_t len) From 04ad69c342fc4de5bd23be9ef15ea7574fb1a87e Mon Sep 17 00:00:00 2001 From: Paulo Alcantara Date: Mon, 8 Mar 2021 12:00:50 -0300 Subject: [PATCH 21/22] cifs: do not send close in compound create+close requests MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In case of interrupted syscalls, prevent sending CLOSE commands for compound CREATE+CLOSE requests by introducing an CIFS_CP_CREATE_CLOSE_OP flag to indicate lower layers that it should not send a CLOSE command to the MIDs corresponding the compound CREATE+CLOSE request. A simple reproducer: #!/bin/bash mount //server/share /mnt -o username=foo,password=*** tc qdisc add dev eth0 root netem delay 450ms stat -f /mnt &>/dev/null & pid=$! sleep 0.01 kill $pid tc qdisc del dev eth0 root umount /mnt Before patch: ... 6 0.256893470 192.168.122.2 → 192.168.122.15 SMB2 402 Create Request File: ;GetInfo Request FS_INFO/FileFsFullSizeInformation;Close Request 7 0.257144491 192.168.122.15 → 192.168.122.2 SMB2 498 Create Response File: ;GetInfo Response;Close Response 9 0.260798209 192.168.122.2 → 192.168.122.15 SMB2 146 Close Request File: 10 0.260841089 192.168.122.15 → 192.168.122.2 SMB2 130 Close Response, Error: STATUS_FILE_CLOSED Signed-off-by: Paulo Alcantara (SUSE) Reviewed-by: Ronnie Sahlberg Reviewed-by: Aurelien Aptel CC: Signed-off-by: Steve French --- fs/cifs/cifsglob.h | 19 ++++++++++--------- fs/cifs/smb2inode.c | 1 + fs/cifs/smb2misc.c | 8 ++++---- fs/cifs/smb2ops.c | 10 +++++----- fs/cifs/smb2proto.h | 3 +-- fs/cifs/transport.c | 2 +- 6 files changed, 22 insertions(+), 21 deletions(-) diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 3de3c5908a72..31fc8695abd6 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -257,7 +257,7 @@ struct smb_version_operations { /* verify the message */ int (*check_message)(char *, unsigned int, struct TCP_Server_Info *); bool (*is_oplock_break)(char *, struct TCP_Server_Info *); - int (*handle_cancelled_mid)(char *, struct TCP_Server_Info *); + int (*handle_cancelled_mid)(struct mid_q_entry *, struct TCP_Server_Info *); void (*downgrade_oplock)(struct TCP_Server_Info *server, struct cifsInodeInfo *cinode, __u32 oplock, unsigned int epoch, bool *purge_cache); @@ -1705,16 +1705,17 @@ static inline bool is_retryable_error(int error) #define CIFS_NO_RSP_BUF 0x040 /* no response buffer required */ /* Type of request operation */ -#define CIFS_ECHO_OP 0x080 /* echo request */ -#define CIFS_OBREAK_OP 0x0100 /* oplock break request */ -#define CIFS_NEG_OP 0x0200 /* negotiate request */ +#define CIFS_ECHO_OP 0x080 /* echo request */ +#define CIFS_OBREAK_OP 0x0100 /* oplock break request */ +#define CIFS_NEG_OP 0x0200 /* negotiate request */ +#define CIFS_CP_CREATE_CLOSE_OP 0x0400 /* compound create+close request */ /* Lower bitmask values are reserved by others below. */ -#define CIFS_SESS_OP 0x2000 /* session setup request */ -#define CIFS_OP_MASK 0x2380 /* mask request type */ +#define CIFS_SESS_OP 0x2000 /* session setup request */ +#define CIFS_OP_MASK 0x2780 /* mask request type */ -#define CIFS_HAS_CREDITS 0x0400 /* already has credits */ -#define CIFS_TRANSFORM_REQ 0x0800 /* transform request before sending */ -#define CIFS_NO_SRV_RSP 0x1000 /* there is no server response */ +#define CIFS_HAS_CREDITS 0x0400 /* already has credits */ +#define CIFS_TRANSFORM_REQ 0x0800 /* transform request before sending */ +#define CIFS_NO_SRV_RSP 0x1000 /* there is no server response */ /* Security Flags: indicate type of session setup needed */ #define CIFSSEC_MAY_SIGN 0x00001 diff --git a/fs/cifs/smb2inode.c b/fs/cifs/smb2inode.c index 1f900b81c34a..a718dc77e604 100644 --- a/fs/cifs/smb2inode.c +++ b/fs/cifs/smb2inode.c @@ -358,6 +358,7 @@ smb2_compound_op(const unsigned int xid, struct cifs_tcon *tcon, if (cfile) goto after_close; /* Close */ + flags |= CIFS_CP_CREATE_CLOSE_OP; rqst[num_rqst].rq_iov = &vars->close_iov[0]; rqst[num_rqst].rq_nvec = 1; rc = SMB2_close_init(tcon, server, diff --git a/fs/cifs/smb2misc.c b/fs/cifs/smb2misc.c index 6e0ea19e710b..b50164e2c88d 100644 --- a/fs/cifs/smb2misc.c +++ b/fs/cifs/smb2misc.c @@ -844,14 +844,14 @@ smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid, } int -smb2_handle_cancelled_mid(char *buffer, struct TCP_Server_Info *server) +smb2_handle_cancelled_mid(struct mid_q_entry *mid, struct TCP_Server_Info *server) { - struct smb2_sync_hdr *sync_hdr = (struct smb2_sync_hdr *)buffer; - struct smb2_create_rsp *rsp = (struct smb2_create_rsp *)buffer; + struct smb2_sync_hdr *sync_hdr = mid->resp_buf; + struct smb2_create_rsp *rsp = mid->resp_buf; struct cifs_tcon *tcon; int rc; - if (sync_hdr->Command != SMB2_CREATE || + if ((mid->optype & CIFS_CP_CREATE_CLOSE_OP) || sync_hdr->Command != SMB2_CREATE || sync_hdr->Status != STATUS_SUCCESS) return 0; diff --git a/fs/cifs/smb2ops.c b/fs/cifs/smb2ops.c index f5087295424c..9bae7e8deb09 100644 --- a/fs/cifs/smb2ops.c +++ b/fs/cifs/smb2ops.c @@ -1195,7 +1195,7 @@ smb2_set_ea(const unsigned int xid, struct cifs_tcon *tcon, struct TCP_Server_Info *server = cifs_pick_channel(ses); __le16 *utf16_path = NULL; int ea_name_len = strlen(ea_name); - int flags = 0; + int flags = CIFS_CP_CREATE_CLOSE_OP; int len; struct smb_rqst rqst[3]; int resp_buftype[3]; @@ -1573,7 +1573,7 @@ smb2_ioctl_query_info(const unsigned int xid, struct smb_query_info qi; struct smb_query_info __user *pqi; int rc = 0; - int flags = 0; + int flags = CIFS_CP_CREATE_CLOSE_OP; struct smb2_query_info_rsp *qi_rsp = NULL; struct smb2_ioctl_rsp *io_rsp = NULL; void *buffer = NULL; @@ -2577,7 +2577,7 @@ smb2_query_info_compound(const unsigned int xid, struct cifs_tcon *tcon, { struct cifs_ses *ses = tcon->ses; struct TCP_Server_Info *server = cifs_pick_channel(ses); - int flags = 0; + int flags = CIFS_CP_CREATE_CLOSE_OP; struct smb_rqst rqst[3]; int resp_buftype[3]; struct kvec rsp_iov[3]; @@ -2975,7 +2975,7 @@ smb2_query_symlink(const unsigned int xid, struct cifs_tcon *tcon, unsigned int sub_offset; unsigned int print_len; unsigned int print_offset; - int flags = 0; + int flags = CIFS_CP_CREATE_CLOSE_OP; struct smb_rqst rqst[3]; int resp_buftype[3]; struct kvec rsp_iov[3]; @@ -3157,7 +3157,7 @@ smb2_query_reparse_tag(const unsigned int xid, struct cifs_tcon *tcon, struct cifs_open_parms oparms; struct cifs_fid fid; struct TCP_Server_Info *server = cifs_pick_channel(tcon->ses); - int flags = 0; + int flags = CIFS_CP_CREATE_CLOSE_OP; struct smb_rqst rqst[3]; int resp_buftype[3]; struct kvec rsp_iov[3]; diff --git a/fs/cifs/smb2proto.h b/fs/cifs/smb2proto.h index 9565e27681a5..a2eb34a8d9c9 100644 --- a/fs/cifs/smb2proto.h +++ b/fs/cifs/smb2proto.h @@ -246,8 +246,7 @@ extern int SMB2_oplock_break(const unsigned int xid, struct cifs_tcon *tcon, extern int smb2_handle_cancelled_close(struct cifs_tcon *tcon, __u64 persistent_fid, __u64 volatile_fid); -extern int smb2_handle_cancelled_mid(char *buffer, - struct TCP_Server_Info *server); +extern int smb2_handle_cancelled_mid(struct mid_q_entry *mid, struct TCP_Server_Info *server); void smb2_cancelled_close_fid(struct work_struct *work); extern int SMB2_QFS_info(const unsigned int xid, struct cifs_tcon *tcon, u64 persistent_file_id, u64 volatile_file_id, diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index 15b72d2e8713..007d99437c77 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -101,7 +101,7 @@ static void _cifs_mid_q_entry_release(struct kref *refcount) if (midEntry->resp_buf && (midEntry->mid_flags & MID_WAIT_CANCELLED) && midEntry->mid_state == MID_RESPONSE_RECEIVED && server->ops->handle_cancelled_mid) - server->ops->handle_cancelled_mid(midEntry->resp_buf, server); + server->ops->handle_cancelled_mid(midEntry, server); midEntry->mid_state = MID_FREE; atomic_dec(&midCount); From 9b1ea29bc0d7b94d420f96a0f4121403efc3dd85 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 10 Mar 2021 10:18:04 -0800 Subject: [PATCH 22/22] Revert "mm, slub: consider rest of partial list if acquire_slab() fails" This reverts commit 8ff60eb052eeba95cfb3efe16b08c9199f8121cf. The kernel test robot reports a huge performance regression due to the commit, and the reason seems fairly straightforward: when there is contention on the page list (which is what causes acquire_slab() to fail), we do _not_ want to just loop and try again, because that will transfer the contention to the 'n->list_lock' spinlock we hold, and just make things even worse. This is admittedly likely a problem only on big machines - the kernel test robot report comes from a 96-thread dual socket Intel Xeon Gold 6252 setup, but the regression there really is quite noticeable: -47.9% regression of stress-ng.rawpkt.ops_per_sec and the commit that was marked as being fixed (7ced37197196: "slub: Acquire_slab() avoid loop") actually did the loop exit early very intentionally (the hint being that "avoid loop" part of that commit message), exactly to avoid this issue. The correct thing to do may be to pick some kind of reasonable middle ground: instead of breaking out of the loop on the very first sign of contention, or trying over and over and over again, the right thing may be to re-try _once_, and then give up on the second failure (or pick your favorite value for "once"..). Reported-by: kernel test robot Link: https://lore.kernel.org/lkml/20210301080404.GF12822@xsang-OptiPlex-9020/ Cc: Jann Horn Cc: David Rientjes Cc: Joonsoo Kim Acked-by: Christoph Lameter Signed-off-by: Linus Torvalds --- mm/slub.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/slub.c b/mm/slub.c index e26c274b4657..3021ce9bf1b3 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -1993,7 +1993,7 @@ static void *get_partial_node(struct kmem_cache *s, struct kmem_cache_node *n, t = acquire_slab(s, n, page, object == NULL, &objects); if (!t) - continue; /* cmpxchg raced */ + break; available += objects; if (!object) {