From: Sasha Levin Date: Wed, 18 Mar 2020 20:57:36 +0000 (-0400) Subject: Fixes for 4.9 X-Git-Tag: v4.4.217~27 X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=cfbaaceb7ab5ff8939e9be8736e553760291b6b7;p=thirdparty%2Fkernel%2Fstable-queue.git Fixes for 4.9 Signed-off-by: Sasha Levin --- diff --git a/queue-4.9/acpi-watchdog-allow-disabling-wdat-at-boot.patch b/queue-4.9/acpi-watchdog-allow-disabling-wdat-at-boot.patch new file mode 100644 index 00000000000..b69eb885ce7 --- /dev/null +++ b/queue-4.9/acpi-watchdog-allow-disabling-wdat-at-boot.patch @@ -0,0 +1,74 @@ +From 20aaa9be90b06ba600a6dafa19340497bb8ae179 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Thu, 6 Feb 2020 16:58:45 +0100 +Subject: ACPI: watchdog: Allow disabling WDAT at boot + +From: Jean Delvare + +[ Upstream commit 3f9e12e0df012c4a9a7fd7eb0d3ae69b459d6b2c ] + +In case the WDAT interface is broken, give the user an option to +ignore it to let a native driver bind to the watchdog device instead. + +Signed-off-by: Jean Delvare +Acked-by: Mika Westerberg +Signed-off-by: Rafael J. Wysocki +Signed-off-by: Sasha Levin +--- + Documentation/kernel-parameters.txt | 4 ++++ + drivers/acpi/acpi_watchdog.c | 12 +++++++++++- + 2 files changed, 15 insertions(+), 1 deletion(-) + +diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt +index b2d2f4539a3fe..e05d65d6fcb68 100644 +--- a/Documentation/kernel-parameters.txt ++++ b/Documentation/kernel-parameters.txt +@@ -335,6 +335,10 @@ bytes respectively. Such letter suffixes can also be entirely omitted. + dynamic table installation which will install SSDT + tables to /sys/firmware/acpi/tables/dynamic. + ++ acpi_no_watchdog [HW,ACPI,WDT] ++ Ignore the ACPI-based watchdog interface (WDAT) and let ++ a native driver control the watchdog device instead. ++ + acpi_rsdp= [ACPI,EFI,KEXEC] + Pass the RSDP address to the kernel, mostly used + on machines running EFI runtime service to boot the +diff --git a/drivers/acpi/acpi_watchdog.c b/drivers/acpi/acpi_watchdog.c +index 7ef0a0e105e18..4296f4932294f 100644 +--- a/drivers/acpi/acpi_watchdog.c ++++ b/drivers/acpi/acpi_watchdog.c +@@ -58,12 +58,14 @@ static bool acpi_watchdog_uses_rtc(const struct acpi_table_wdat *wdat) + } + #endif + ++static bool acpi_no_watchdog; ++ + static const struct acpi_table_wdat *acpi_watchdog_get_wdat(void) + { + const struct acpi_table_wdat *wdat = NULL; + acpi_status status; + +- if (acpi_disabled) ++ if (acpi_disabled || acpi_no_watchdog) + return NULL; + + status = acpi_get_table(ACPI_SIG_WDAT, 0, +@@ -91,6 +93,14 @@ bool acpi_has_watchdog(void) + } + EXPORT_SYMBOL_GPL(acpi_has_watchdog); + ++/* ACPI watchdog can be disabled on boot command line */ ++static int __init disable_acpi_watchdog(char *str) ++{ ++ acpi_no_watchdog = true; ++ return 1; ++} ++__setup("acpi_no_watchdog", disable_acpi_watchdog); ++ + void __init acpi_watchdog_init(void) + { + const struct acpi_wdat_entry *entries; +-- +2.20.1 + diff --git a/queue-4.9/cfg80211-check-reg_rule-for-null-in-handle_channel_c.patch b/queue-4.9/cfg80211-check-reg_rule-for-null-in-handle_channel_c.patch new file mode 100644 index 00000000000..6c3c88da502 --- /dev/null +++ b/queue-4.9/cfg80211-check-reg_rule-for-null-in-handle_channel_c.patch @@ -0,0 +1,37 @@ +From 126a5ae6e3d67663f34cc4c3f4d8e08bbaf1d570 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 21 Feb 2020 10:44:50 +0100 +Subject: cfg80211: check reg_rule for NULL in handle_channel_custom() + +From: Johannes Berg + +[ Upstream commit a7ee7d44b57c9ae174088e53a668852b7f4f452d ] + +We may end up with a NULL reg_rule after the loop in +handle_channel_custom() if the bandwidth didn't fit, +check if this is the case and bail out if so. + +Signed-off-by: Johannes Berg +Link: https://lore.kernel.org/r/20200221104449.3b558a50201c.I4ad3725c4dacaefd2d18d3cc65ba6d18acd5dbfe@changeid +Signed-off-by: Johannes Berg +Signed-off-by: Sasha Levin +--- + net/wireless/reg.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/net/wireless/reg.c b/net/wireless/reg.c +index 0e66768427ba7..6d5f3f737207d 100644 +--- a/net/wireless/reg.c ++++ b/net/wireless/reg.c +@@ -1730,7 +1730,7 @@ static void handle_channel_custom(struct wiphy *wiphy, + break; + } + +- if (IS_ERR(reg_rule)) { ++ if (IS_ERR_OR_NULL(reg_rule)) { + pr_debug("Disabling freq %d MHz as custom regd has no rule that fits it\n", + chan->center_freq); + if (wiphy->regulatory_flags & REGULATORY_WIPHY_SELF_MANAGED) { +-- +2.20.1 + diff --git a/queue-4.9/hid-apple-add-support-for-recent-firmware-on-magic-k.patch b/queue-4.9/hid-apple-add-support-for-recent-firmware-on-magic-k.patch new file mode 100644 index 00000000000..933942e7f35 --- /dev/null +++ b/queue-4.9/hid-apple-add-support-for-recent-firmware-on-magic-k.patch @@ -0,0 +1,37 @@ +From 6bf9f2e1f477717e1bc28b59a7ea88cdc6d66a52 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Wed, 29 Jan 2020 17:26:31 +1100 +Subject: HID: apple: Add support for recent firmware on Magic Keyboards + +From: Mansour Behabadi + +[ Upstream commit e433be929e63265b7412478eb7ff271467aee2d7 ] + +Magic Keyboards with more recent firmware (0x0100) report Fn key differently. +Without this patch, Fn key may not behave as expected and may not be +configurable via hid_apple fnmode module parameter. + +Signed-off-by: Mansour Behabadi +Signed-off-by: Jiri Kosina +Signed-off-by: Sasha Levin +--- + drivers/hid/hid-apple.c | 3 ++- + 1 file changed, 2 insertions(+), 1 deletion(-) + +diff --git a/drivers/hid/hid-apple.c b/drivers/hid/hid-apple.c +index 31c087e1746d6..197eb75d10ef2 100644 +--- a/drivers/hid/hid-apple.c ++++ b/drivers/hid/hid-apple.c +@@ -341,7 +341,8 @@ static int apple_input_mapping(struct hid_device *hdev, struct hid_input *hi, + unsigned long **bit, int *max) + { + if (usage->hid == (HID_UP_CUSTOM | 0x0003) || +- usage->hid == (HID_UP_MSVENDOR | 0x0003)) { ++ usage->hid == (HID_UP_MSVENDOR | 0x0003) || ++ usage->hid == (HID_UP_HPVENDOR2 | 0x0003)) { + /* The fn key on Apple USB keyboards */ + set_bit(EV_REP, hi->input->evbit); + hid_map_usage_clear(hi, usage, bit, max, EV_KEY, KEY_FN); +-- +2.20.1 + diff --git a/queue-4.9/hid-i2c-hid-add-trekstor-surfbook-e11b-to-descriptor.patch b/queue-4.9/hid-i2c-hid-add-trekstor-surfbook-e11b-to-descriptor.patch new file mode 100644 index 00000000000..2795601f2be --- /dev/null +++ b/queue-4.9/hid-i2c-hid-add-trekstor-surfbook-e11b-to-descriptor.patch @@ -0,0 +1,43 @@ +From d4de12f7a1b2d637d9d96576f6dd59012d17330a Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 14 Feb 2020 14:53:07 +0800 +Subject: HID: i2c-hid: add Trekstor Surfbook E11B to descriptor override + +From: Kai-Heng Feng + +[ Upstream commit be0aba826c4a6ba5929def1962a90d6127871969 ] + +The Surfbook E11B uses the SIPODEV SP1064 touchpad, which does not supply +descriptors, so it has to be added to the override list. + +BugLink: https://bugs.launchpad.net/bugs/1858299 +Signed-off-by: Kai-Heng Feng +Reviewed-by: Hans de Goede +Signed-off-by: Benjamin Tissoires +Signed-off-by: Sasha Levin +--- + drivers/hid/i2c-hid/i2c-hid-dmi-quirks.c | 8 ++++++++ + 1 file changed, 8 insertions(+) + +diff --git a/drivers/hid/i2c-hid/i2c-hid-dmi-quirks.c b/drivers/hid/i2c-hid/i2c-hid-dmi-quirks.c +index 10af8585c820d..95052373a8282 100644 +--- a/drivers/hid/i2c-hid/i2c-hid-dmi-quirks.c ++++ b/drivers/hid/i2c-hid/i2c-hid-dmi-quirks.c +@@ -341,6 +341,14 @@ static const struct dmi_system_id i2c_hid_dmi_desc_override_table[] = { + }, + .driver_data = (void *)&sipodev_desc + }, ++ { ++ .ident = "Trekstor SURFBOOK E11B", ++ .matches = { ++ DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TREKSTOR"), ++ DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "SURFBOOK E11B"), ++ }, ++ .driver_data = (void *)&sipodev_desc ++ }, + { + .ident = "Direkt-Tek DTLAPY116-2", + .matches = { +-- +2.20.1 + diff --git a/queue-4.9/jbd2-fix-data-races-at-struct-journal_head.patch b/queue-4.9/jbd2-fix-data-races-at-struct-journal_head.patch new file mode 100644 index 00000000000..864a03fe0bf --- /dev/null +++ b/queue-4.9/jbd2-fix-data-races-at-struct-journal_head.patch @@ -0,0 +1,110 @@ +From 5c4cb92b3851495c304e904f1c10163e3bb7087f Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 21 Feb 2020 23:31:11 -0500 +Subject: jbd2: fix data races at struct journal_head + +From: Qian Cai + +[ Upstream commit 6c5d911249290f41f7b50b43344a7520605b1acb ] + +journal_head::b_transaction and journal_head::b_next_transaction could +be accessed concurrently as noticed by KCSAN, + + LTP: starting fsync04 + /dev/zero: Can't open blockdev + EXT4-fs (loop0): mounting ext3 file system using the ext4 subsystem + EXT4-fs (loop0): mounted filesystem with ordered data mode. Opts: (null) + ================================================================== + BUG: KCSAN: data-race in __jbd2_journal_refile_buffer [jbd2] / jbd2_write_access_granted [jbd2] + + write to 0xffff99f9b1bd0e30 of 8 bytes by task 25721 on cpu 70: + __jbd2_journal_refile_buffer+0xdd/0x210 [jbd2] + __jbd2_journal_refile_buffer at fs/jbd2/transaction.c:2569 + jbd2_journal_commit_transaction+0x2d15/0x3f20 [jbd2] + (inlined by) jbd2_journal_commit_transaction at fs/jbd2/commit.c:1034 + kjournald2+0x13b/0x450 [jbd2] + kthread+0x1cd/0x1f0 + ret_from_fork+0x27/0x50 + + read to 0xffff99f9b1bd0e30 of 8 bytes by task 25724 on cpu 68: + jbd2_write_access_granted+0x1b2/0x250 [jbd2] + jbd2_write_access_granted at fs/jbd2/transaction.c:1155 + jbd2_journal_get_write_access+0x2c/0x60 [jbd2] + __ext4_journal_get_write_access+0x50/0x90 [ext4] + ext4_mb_mark_diskspace_used+0x158/0x620 [ext4] + ext4_mb_new_blocks+0x54f/0xca0 [ext4] + ext4_ind_map_blocks+0xc79/0x1b40 [ext4] + ext4_map_blocks+0x3b4/0x950 [ext4] + _ext4_get_block+0xfc/0x270 [ext4] + ext4_get_block+0x3b/0x50 [ext4] + __block_write_begin_int+0x22e/0xae0 + __block_write_begin+0x39/0x50 + ext4_write_begin+0x388/0xb50 [ext4] + generic_perform_write+0x15d/0x290 + ext4_buffered_write_iter+0x11f/0x210 [ext4] + ext4_file_write_iter+0xce/0x9e0 [ext4] + new_sync_write+0x29c/0x3b0 + __vfs_write+0x92/0xa0 + vfs_write+0x103/0x260 + ksys_write+0x9d/0x130 + __x64_sys_write+0x4c/0x60 + do_syscall_64+0x91/0xb05 + entry_SYSCALL_64_after_hwframe+0x49/0xbe + + 5 locks held by fsync04/25724: + #0: ffff99f9911093f8 (sb_writers#13){.+.+}, at: vfs_write+0x21c/0x260 + #1: ffff99f9db4c0348 (&sb->s_type->i_mutex_key#15){+.+.}, at: ext4_buffered_write_iter+0x65/0x210 [ext4] + #2: ffff99f5e7dfcf58 (jbd2_handle){++++}, at: start_this_handle+0x1c1/0x9d0 [jbd2] + #3: ffff99f9db4c0168 (&ei->i_data_sem){++++}, at: ext4_map_blocks+0x176/0x950 [ext4] + #4: ffffffff99086b40 (rcu_read_lock){....}, at: jbd2_write_access_granted+0x4e/0x250 [jbd2] + irq event stamp: 1407125 + hardirqs last enabled at (1407125): [] __find_get_block+0x107/0x790 + hardirqs last disabled at (1407124): [] __find_get_block+0x49/0x790 + softirqs last enabled at (1405528): [] __do_softirq+0x34c/0x57c + softirqs last disabled at (1405521): [] irq_exit+0xa2/0xc0 + + Reported by Kernel Concurrency Sanitizer on: + CPU: 68 PID: 25724 Comm: fsync04 Tainted: G L 5.6.0-rc2-next-20200221+ #7 + Hardware name: HPE ProLiant DL385 Gen10/ProLiant DL385 Gen10, BIOS A40 07/10/2019 + +The plain reads are outside of jh->b_state_lock critical section which result +in data races. Fix them by adding pairs of READ|WRITE_ONCE(). + +Reviewed-by: Jan Kara +Signed-off-by: Qian Cai +Link: https://lore.kernel.org/r/20200222043111.2227-1-cai@lca.pw +Signed-off-by: Theodore Ts'o +Signed-off-by: Sasha Levin +--- + fs/jbd2/transaction.c | 8 ++++---- + 1 file changed, 4 insertions(+), 4 deletions(-) + +diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c +index 04dd0652bb5ca..8de458d64134a 100644 +--- a/fs/jbd2/transaction.c ++++ b/fs/jbd2/transaction.c +@@ -1037,8 +1037,8 @@ static bool jbd2_write_access_granted(handle_t *handle, struct buffer_head *bh, + /* For undo access buffer must have data copied */ + if (undo && !jh->b_committed_data) + goto out; +- if (jh->b_transaction != handle->h_transaction && +- jh->b_next_transaction != handle->h_transaction) ++ if (READ_ONCE(jh->b_transaction) != handle->h_transaction && ++ READ_ONCE(jh->b_next_transaction) != handle->h_transaction) + goto out; + /* + * There are two reasons for the barrier here: +@@ -2448,8 +2448,8 @@ void __jbd2_journal_refile_buffer(struct journal_head *jh) + * our jh reference and thus __jbd2_journal_file_buffer() must not + * take a new one. + */ +- jh->b_transaction = jh->b_next_transaction; +- jh->b_next_transaction = NULL; ++ WRITE_ONCE(jh->b_transaction, jh->b_next_transaction); ++ WRITE_ONCE(jh->b_next_transaction, NULL); + if (buffer_freed(bh)) + jlist = BJ_Forget; + else if (jh->b_modified) +-- +2.20.1 + diff --git a/queue-4.9/mac80211-rx-avoid-rcu-list-traversal-under-mutex.patch b/queue-4.9/mac80211-rx-avoid-rcu-list-traversal-under-mutex.patch new file mode 100644 index 00000000000..b6c1d841fd5 --- /dev/null +++ b/queue-4.9/mac80211-rx-avoid-rcu-list-traversal-under-mutex.patch @@ -0,0 +1,39 @@ +From 82c652c6b3a10cb942e05f8c685a1c972bf4b205 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Sun, 23 Feb 2020 20:03:02 +0530 +Subject: mac80211: rx: avoid RCU list traversal under mutex + +From: Madhuparna Bhowmik + +[ Upstream commit 253216ffb2a002a682c6f68bd3adff5b98b71de8 ] + +local->sta_mtx is held in __ieee80211_check_fast_rx_iface(). +No need to use list_for_each_entry_rcu() as it also requires +a cond argument to avoid false lockdep warnings when not used in +RCU read-side section (with CONFIG_PROVE_RCU_LIST). +Therefore use list_for_each_entry(); + +Signed-off-by: Madhuparna Bhowmik +Link: https://lore.kernel.org/r/20200223143302.15390-1-madhuparnabhowmik10@gmail.com +Signed-off-by: Johannes Berg +Signed-off-by: Sasha Levin +--- + net/mac80211/rx.c | 2 +- + 1 file changed, 1 insertion(+), 1 deletion(-) + +diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c +index 74652eb2f90fd..a6f265262f151 100644 +--- a/net/mac80211/rx.c ++++ b/net/mac80211/rx.c +@@ -3841,7 +3841,7 @@ void __ieee80211_check_fast_rx_iface(struct ieee80211_sub_if_data *sdata) + + lockdep_assert_held(&local->sta_mtx); + +- list_for_each_entry_rcu(sta, &local->sta_list, list) { ++ list_for_each_entry(sta, &local->sta_list, list) { + if (sdata != sta->sdata && + (!sta->sdata->bss || sta->sdata->bss != sdata->bss)) + continue; +-- +2.20.1 + diff --git a/queue-4.9/net-ks8851-ml-fix-irq-handling-and-locking.patch b/queue-4.9/net-ks8851-ml-fix-irq-handling-and-locking.patch new file mode 100644 index 00000000000..82be0ccd190 --- /dev/null +++ b/queue-4.9/net-ks8851-ml-fix-irq-handling-and-locking.patch @@ -0,0 +1,102 @@ +From a9a9a55f2513e1586879ca901c507aaf06345024 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Sun, 23 Feb 2020 14:38:40 +0100 +Subject: net: ks8851-ml: Fix IRQ handling and locking + +From: Marek Vasut + +[ Upstream commit 44343418d0f2f623cb9da6f5000df793131cbe3b ] + +The KS8851 requires that packet RX and TX are mutually exclusive. +Currently, the driver hopes to achieve this by disabling interrupt +from the card by writing the card registers and by disabling the +interrupt on the interrupt controller. This however is racy on SMP. + +Replace this approach by expanding the spinlock used around the +ks_start_xmit() TX path to ks_irq() RX path to assure true mutual +exclusion and remove the interrupt enabling/disabling, which is +now not needed anymore. Furthermore, disable interrupts also in +ks_net_stop(), which was missing before. + +Note that a massive improvement here would be to re-use the KS8851 +driver approach, which is to move the TX path into a worker thread, +interrupt handling to threaded interrupt, and synchronize everything +with mutexes, but that would be a much bigger rework, for a separate +patch. + +Signed-off-by: Marek Vasut +Cc: David S. Miller +Cc: Lukas Wunner +Cc: Petr Stetiar +Cc: YueHaibing +Signed-off-by: David S. Miller +Signed-off-by: Sasha Levin +--- + drivers/net/ethernet/micrel/ks8851_mll.c | 14 ++++++++------ + 1 file changed, 8 insertions(+), 6 deletions(-) + +diff --git a/drivers/net/ethernet/micrel/ks8851_mll.c b/drivers/net/ethernet/micrel/ks8851_mll.c +index d94e151cff12b..d4747caf1e7cc 100644 +--- a/drivers/net/ethernet/micrel/ks8851_mll.c ++++ b/drivers/net/ethernet/micrel/ks8851_mll.c +@@ -831,14 +831,17 @@ static irqreturn_t ks_irq(int irq, void *pw) + { + struct net_device *netdev = pw; + struct ks_net *ks = netdev_priv(netdev); ++ unsigned long flags; + u16 status; + ++ spin_lock_irqsave(&ks->statelock, flags); + /*this should be the first in IRQ handler */ + ks_save_cmd_reg(ks); + + status = ks_rdreg16(ks, KS_ISR); + if (unlikely(!status)) { + ks_restore_cmd_reg(ks); ++ spin_unlock_irqrestore(&ks->statelock, flags); + return IRQ_NONE; + } + +@@ -864,6 +867,7 @@ static irqreturn_t ks_irq(int irq, void *pw) + ks->netdev->stats.rx_over_errors++; + /* this should be the last in IRQ handler*/ + ks_restore_cmd_reg(ks); ++ spin_unlock_irqrestore(&ks->statelock, flags); + return IRQ_HANDLED; + } + +@@ -933,6 +937,7 @@ static int ks_net_stop(struct net_device *netdev) + + /* shutdown RX/TX QMU */ + ks_disable_qmu(ks); ++ ks_disable_int(ks); + + /* set powermode to soft power down to save power */ + ks_set_powermode(ks, PMECR_PM_SOFTDOWN); +@@ -989,10 +994,9 @@ static netdev_tx_t ks_start_xmit(struct sk_buff *skb, struct net_device *netdev) + { + netdev_tx_t retv = NETDEV_TX_OK; + struct ks_net *ks = netdev_priv(netdev); ++ unsigned long flags; + +- disable_irq(netdev->irq); +- ks_disable_int(ks); +- spin_lock(&ks->statelock); ++ spin_lock_irqsave(&ks->statelock, flags); + + /* Extra space are required: + * 4 byte for alignment, 4 for status/length, 4 for CRC +@@ -1006,9 +1010,7 @@ static netdev_tx_t ks_start_xmit(struct sk_buff *skb, struct net_device *netdev) + dev_kfree_skb(skb); + } else + retv = NETDEV_TX_BUSY; +- spin_unlock(&ks->statelock); +- ks_enable_int(ks); +- enable_irq(netdev->irq); ++ spin_unlock_irqrestore(&ks->statelock, flags); + return retv; + } + +-- +2.20.1 + diff --git a/queue-4.9/series b/queue-4.9/series index 5778eefdf5d..6f53cb2acc9 100644 --- a/queue-4.9/series +++ b/queue-4.9/series @@ -76,3 +76,11 @@ batman-adv-update-data-pointers-after-skb_cow.patch batman-adv-avoid-probe-elp-information-leak.patch batman-adv-use-explicit-tvlv-padding-for-elp-packets.patch perf-amd-uncore-replace-manual-sampling-check-with-c.patch +acpi-watchdog-allow-disabling-wdat-at-boot.patch +hid-apple-add-support-for-recent-firmware-on-magic-k.patch +hid-i2c-hid-add-trekstor-surfbook-e11b-to-descriptor.patch +cfg80211-check-reg_rule-for-null-in-handle_channel_c.patch +net-ks8851-ml-fix-irq-handling-and-locking.patch +mac80211-rx-avoid-rcu-list-traversal-under-mutex.patch +signal-avoid-double-atomic-counter-increments-for-us.patch +jbd2-fix-data-races-at-struct-journal_head.patch diff --git a/queue-4.9/signal-avoid-double-atomic-counter-increments-for-us.patch b/queue-4.9/signal-avoid-double-atomic-counter-increments-for-us.patch new file mode 100644 index 00000000000..7ea81f6c7cd --- /dev/null +++ b/queue-4.9/signal-avoid-double-atomic-counter-increments-for-us.patch @@ -0,0 +1,125 @@ +From c17bb196d4ddeaa73e0f23872f9eeebaec9ba20b Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Mon, 24 Feb 2020 12:47:14 -0800 +Subject: signal: avoid double atomic counter increments for user accounting + +From: Linus Torvalds + +[ Upstream commit fda31c50292a5062332fa0343c084bd9f46604d9 ] + +When queueing a signal, we increment both the users count of pending +signals (for RLIMIT_SIGPENDING tracking) and we increment the refcount +of the user struct itself (because we keep a reference to the user in +the signal structure in order to correctly account for it when freeing). + +That turns out to be fairly expensive, because both of them are atomic +updates, and particularly under extreme signal handling pressure on big +machines, you can get a lot of cache contention on the user struct. +That can then cause horrid cacheline ping-pong when you do these +multiple accesses. + +So change the reference counting to only pin the user for the _first_ +pending signal, and to unpin it when the last pending signal is +dequeued. That means that when a user sees a lot of concurrent signal +queuing - which is the only situation when this matters - the only +atomic access needed is generally the 'sigpending' count update. + +This was noticed because of a particularly odd timing artifact on a +dual-socket 96C/192T Cascade Lake platform: when you get into bad +contention, on that machine for some reason seems to be much worse when +the contention happens in the upper 32-byte half of the cacheline. + +As a result, the kernel test robot will-it-scale 'signal1' benchmark had +an odd performance regression simply due to random alignment of the +'struct user_struct' (and pointed to a completely unrelated and +apparently nonsensical commit for the regression). + +Avoiding the double increments (and decrements on the dequeueing side, +of course) makes for much less contention and hugely improved +performance on that will-it-scale microbenchmark. + +Quoting Feng Tang: + + "It makes a big difference, that the performance score is tripled! bump + from original 17000 to 54000. Also the gap between 5.0-rc6 and + 5.0-rc6+Jiri's patch is reduced to around 2%" + +[ The "2% gap" is the odd cacheline placement difference on that + platform: under the extreme contention case, the effect of which half + of the cacheline was hot was 5%, so with the reduced contention the + odd timing artifact is reduced too ] + +It does help in the non-contended case too, but is not nearly as +noticeable. + +Reported-and-tested-by: Feng Tang +Cc: Eric W. Biederman +Cc: Huang, Ying +Cc: Philip Li +Cc: Andi Kleen +Cc: Jiri Olsa +Cc: Peter Zijlstra +Signed-off-by: Linus Torvalds +Signed-off-by: Sasha Levin +--- + kernel/signal.c | 23 ++++++++++++++--------- + 1 file changed, 14 insertions(+), 9 deletions(-) + +diff --git a/kernel/signal.c b/kernel/signal.c +index 57fadbe69c2e6..d90ccbeb909d2 100644 +--- a/kernel/signal.c ++++ b/kernel/signal.c +@@ -373,27 +373,32 @@ __sigqueue_alloc(int sig, struct task_struct *t, gfp_t flags, int override_rlimi + { + struct sigqueue *q = NULL; + struct user_struct *user; ++ int sigpending; + + /* + * Protect access to @t credentials. This can go away when all + * callers hold rcu read lock. ++ * ++ * NOTE! A pending signal will hold on to the user refcount, ++ * and we get/put the refcount only when the sigpending count ++ * changes from/to zero. + */ + rcu_read_lock(); +- user = get_uid(__task_cred(t)->user); +- atomic_inc(&user->sigpending); ++ user = __task_cred(t)->user; ++ sigpending = atomic_inc_return(&user->sigpending); ++ if (sigpending == 1) ++ get_uid(user); + rcu_read_unlock(); + +- if (override_rlimit || +- atomic_read(&user->sigpending) <= +- task_rlimit(t, RLIMIT_SIGPENDING)) { ++ if (override_rlimit || likely(sigpending <= task_rlimit(t, RLIMIT_SIGPENDING))) { + q = kmem_cache_alloc(sigqueue_cachep, flags); + } else { + print_dropped_signal(sig); + } + + if (unlikely(q == NULL)) { +- atomic_dec(&user->sigpending); +- free_uid(user); ++ if (atomic_dec_and_test(&user->sigpending)) ++ free_uid(user); + } else { + INIT_LIST_HEAD(&q->list); + q->flags = 0; +@@ -407,8 +412,8 @@ static void __sigqueue_free(struct sigqueue *q) + { + if (q->flags & SIGQUEUE_PREALLOC) + return; +- atomic_dec(&q->user->sigpending); +- free_uid(q->user); ++ if (atomic_dec_and_test(&q->user->sigpending)) ++ free_uid(q->user); + kmem_cache_free(sigqueue_cachep, q); + } + +-- +2.20.1 +