- Linux-stable-mirror - lists.linaro.org

[PATCH] fpga: zynq-fpga: use sgtable-based scatterlist wrappers

by Marek Szyprowski

Use common wrappers operating directly on the struct sg_table objects to fix incorrect use of statterlists related calls. dma_unmap_sg() function has to be called with the number of elements originally passed to the dma_map_sg() function, not the one returned in sgtable's nents. CC: stable(a)vger.kernel.org Fixes: 425902f5c8e3 ("fpga zynq: Use the scatterlist interface") Signed-off-by: Marek Szyprowski <m.szyprowski(a)samsung.com> --- drivers/fpga/zynq-fpga.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/fpga/zynq-fpga.c b/drivers/fpga/zynq-fpga.c index f7e08f7ea9ef..9bd39d1d4048 100644 --- a/drivers/fpga/zynq-fpga.c +++ b/drivers/fpga/zynq-fpga.c @@ -406,7 +406,7 @@ static int zynq_fpga_ops_write(struct fpga_manager *mgr, struct sg_table *sgt) } priv->dma_nelms = - dma_map_sg(mgr->dev.parent, sgt->sgl, sgt->nents, DMA_TO_DEVICE); + dma_map_sgtable(mgr->dev.parent, sgt, DMA_TO_DEVICE); if (priv->dma_nelms == 0) { dev_err(&mgr->dev, "Unable to DMA map (TO_DEVICE)\n"); return -ENOMEM; @@ -478,7 +478,7 @@ static int zynq_fpga_ops_write(struct fpga_manager *mgr, struct sg_table *sgt) clk_disable(priv->clk); out_free: - dma_unmap_sg(mgr->dev.parent, sgt->sgl, sgt->nents, DMA_TO_DEVICE); + dma_unmap_sgtable(mgr->dev.parent, sgt, DMA_TO_DEVICE); return err; } -- 2.34.1

3 weeks, 1 day

3
3
0 0

[PATCH] wifi: mt76: mt7925: fix incorrect scan probe IE handling for hw_scan

by Mingyen Hsieh

From: Ming Yen Hsieh <mingyen.hsieh(a)mediatek.com> The IEs should be processed and filled into the command tlv separately according to each band. Cc: stable(a)vger.kernel.org Fixes: c948b5da6bbe ("wifi: mt76: mt7925: add Mediatek Wi-Fi7 driver for mt7925 chips") Signed-off-by: Ming Yen Hsieh <mingyen.hsieh(a)mediatek.com> --- .../net/wireless/mediatek/mt76/mt7925/main.c | 2 +- .../net/wireless/mediatek/mt76/mt7925/mcu.c | 75 +++++++++++++++---- .../net/wireless/mediatek/mt76/mt7925/mcu.h | 5 +- 3 files changed, 63 insertions(+), 19 deletions(-) diff --git a/drivers/net/wireless/mediatek/mt76/mt7925/main.c b/drivers/net/wireless/mediatek/mt76/mt7925/main.c index 555ef302ae97..506387b4ce13 100644 --- a/drivers/net/wireless/mediatek/mt76/mt7925/main.c +++ b/drivers/net/wireless/mediatek/mt76/mt7925/main.c @@ -1540,7 +1540,7 @@ mt7925_start_sched_scan(struct ieee80211_hw *hw, struct ieee80211_vif *vif, mt792x_mutex_acquire(dev); - err = mt7925_mcu_sched_scan_req(mphy, vif, req); + err = mt7925_mcu_sched_scan_req(mphy, vif, req, ies); if (err < 0) goto out; diff --git a/drivers/net/wireless/mediatek/mt76/mt7925/mcu.c b/drivers/net/wireless/mediatek/mt76/mt7925/mcu.c index ef1d42af96f7..9993d47e643c 100644 --- a/drivers/net/wireless/mediatek/mt76/mt7925/mcu.c +++ b/drivers/net/wireless/mediatek/mt76/mt7925/mcu.c @@ -192,6 +192,7 @@ mt7925_connac_mcu_set_wow_ctrl(struct mt76_phy *phy, struct ieee80211_vif *vif, bool suspend, struct cfg80211_wowlan *wowlan) { struct mt76_vif_link *mvif = (struct mt76_vif_link *)vif->drv_priv; + struct ieee80211_scan_ies ies = {}; struct mt76_dev *dev = phy->dev; struct { struct { @@ -227,7 +228,7 @@ mt7925_connac_mcu_set_wow_ctrl(struct mt76_phy *phy, struct ieee80211_vif *vif, req.wow_ctrl_tlv.trigger |= (UNI_WOW_DETECT_TYPE_DISCONNECT | UNI_WOW_DETECT_TYPE_BCN_LOST); if (wowlan->nd_config) { - mt7925_mcu_sched_scan_req(phy, vif, wowlan->nd_config); + mt7925_mcu_sched_scan_req(phy, vif, wowlan->nd_config, &ies); req.wow_ctrl_tlv.trigger |= UNI_WOW_DETECT_TYPE_SCH_SCAN_HIT; mt7925_mcu_sched_scan_enable(phy, vif, suspend); } @@ -3085,6 +3086,54 @@ int mt7925_mcu_set_dbdc(struct mt76_phy *phy, bool enable) return err; } +static void +mt7925_mcu_build_scan_ie_tlv(struct mt76_dev *mdev, + struct sk_buff *skb, + struct ieee80211_scan_ies *scan_ies) +{ + u32 max_len = sizeof(struct scan_ie_tlv) + MT76_CONNAC_SCAN_IE_LEN; + struct scan_ie_tlv *ie; + enum nl80211_band i; + struct tlv *tlv; + const u8 *ies; + u16 ies_len; + + for (i = 0; i <= NL80211_BAND_6GHZ; i++) { + if (i == NL80211_BAND_60GHZ) + continue; + + ies = scan_ies->ies[i]; + ies_len = scan_ies->len[i]; + + if (!ies || !ies_len) + continue; + + if (ies_len > max_len) + return; + + tlv = mt76_connac_mcu_add_tlv(skb, UNI_SCAN_IE, + sizeof(*ie) + ies_len); + ie = (struct scan_ie_tlv *)tlv; + + memcpy(ie->ies, ies, ies_len); + ie->ies_len = cpu_to_le16(ies_len); + + switch (i) { + case NL80211_BAND_2GHZ: + ie->band = 1; + break; + case NL80211_BAND_6GHZ: + ie->band = 3; + break; + default: + ie->band = 2; + break; + } + + max_len -= (sizeof(*ie) + ies_len); + } +} + int mt7925_mcu_hw_scan(struct mt76_phy *phy, struct ieee80211_vif *vif, struct ieee80211_scan_request *scan_req) { @@ -3110,7 +3159,8 @@ int mt7925_mcu_hw_scan(struct mt76_phy *phy, struct ieee80211_vif *vif, max_len = sizeof(*hdr) + sizeof(*req) + sizeof(*ssid) + sizeof(*bssid) * MT7925_RNR_SCAN_MAX_BSSIDS + - sizeof(*chan_info) + sizeof(*misc) + sizeof(*ie); + sizeof(*chan_info) + sizeof(*misc) + sizeof(*ie) + + MT76_CONNAC_SCAN_IE_LEN; skb = mt76_mcu_msg_alloc(mdev, NULL, max_len); if (!skb) @@ -3192,13 +3242,6 @@ int mt7925_mcu_hw_scan(struct mt76_phy *phy, struct ieee80211_vif *vif, } chan_info->channel_type = sreq->n_channels ? 4 : 0; - tlv = mt76_connac_mcu_add_tlv(skb, UNI_SCAN_IE, sizeof(*ie)); - ie = (struct scan_ie_tlv *)tlv; - if (sreq->ie_len > 0) { - memcpy(ie->ies, sreq->ie, sreq->ie_len); - ie->ies_len = cpu_to_le16(sreq->ie_len); - } - req->scan_func |= SCAN_FUNC_SPLIT_SCAN; tlv = mt76_connac_mcu_add_tlv(skb, UNI_SCAN_MISC, sizeof(*misc)); @@ -3209,6 +3252,9 @@ int mt7925_mcu_hw_scan(struct mt76_phy *phy, struct ieee80211_vif *vif, req->scan_func |= SCAN_FUNC_RANDOM_MAC; } + /* Append scan probe IEs as the last tlv */ + mt7925_mcu_build_scan_ie_tlv(mdev, skb, &scan_req->ies); + err = mt76_mcu_skb_send_msg(mdev, skb, MCU_UNI_CMD(SCAN_REQ), true); if (err < 0) @@ -3220,7 +3266,8 @@ EXPORT_SYMBOL_GPL(mt7925_mcu_hw_scan); int mt7925_mcu_sched_scan_req(struct mt76_phy *phy, struct ieee80211_vif *vif, - struct cfg80211_sched_scan_request *sreq) + struct cfg80211_sched_scan_request *sreq, + struct ieee80211_scan_ies *ies) { struct mt76_vif_link *mvif = (struct mt76_vif_link *)vif->drv_priv; struct ieee80211_channel **scan_list = sreq->channels; @@ -3308,12 +3355,8 @@ int mt7925_mcu_sched_scan_req(struct mt76_phy *phy, } chan_info->channel_type = sreq->n_channels ? 4 : 0; - tlv = mt76_connac_mcu_add_tlv(skb, UNI_SCAN_IE, sizeof(*ie)); - ie = (struct scan_ie_tlv *)tlv; - if (sreq->ie_len > 0) { - memcpy(ie->ies, sreq->ie, sreq->ie_len); - ie->ies_len = cpu_to_le16(sreq->ie_len); - } + /* Append scan probe IEs as the last tlv */ + mt7925_mcu_build_scan_ie_tlv(mdev, skb, ies); return mt76_mcu_skb_send_msg(mdev, skb, MCU_UNI_CMD(SCAN_REQ), true); diff --git a/drivers/net/wireless/mediatek/mt76/mt7925/mcu.h b/drivers/net/wireless/mediatek/mt76/mt7925/mcu.h index 3dddadb7457b..c34b28769782 100644 --- a/drivers/net/wireless/mediatek/mt76/mt7925/mcu.h +++ b/drivers/net/wireless/mediatek/mt76/mt7925/mcu.h @@ -269,7 +269,7 @@ struct scan_ie_tlv { __le16 ies_len; u8 band; u8 pad; - u8 ies[MT76_CONNAC_SCAN_IE_LEN]; + u8 ies[]; }; struct scan_misc_tlv { @@ -674,7 +674,8 @@ int mt7925_mcu_cancel_hw_scan(struct mt76_phy *phy, struct ieee80211_vif *vif); int mt7925_mcu_sched_scan_req(struct mt76_phy *phy, struct ieee80211_vif *vif, - struct cfg80211_sched_scan_request *sreq); + struct cfg80211_sched_scan_request *sreq, + struct ieee80211_scan_ies *ies); int mt7925_mcu_sched_scan_enable(struct mt76_phy *phy, struct ieee80211_vif *vif, bool enable); -- 2.34.1

3 weeks, 1 day

1
0
0 0

[PATCH v2 0/5] drm/xe: enable driver usage on non-4KiB kernels

by Mingcong Bai via B4 Relay

This patch series attempts to enable the use of xe DRM driver on non-4KiB kernel page platforms. This involves fixing the ttm/bo interface, as well as parts of the userspace API to make use of kernel `PAGE_SIZE' for alignment instead of the assumed `SZ_4K', it also fixes incorrect usage of `PAGE_SIZE' in the GuC and ring buffer interface code to make sure all instructions/commands were aligned to 4KiB barriers (per the Programmer's Manual for the GPUs covered by this DRM driver). This issue was first discovered and reported by members of the LoongArch user communities, whose hardware commonly ran on 16KiB-page kernels. The patch series began on an unassuming branch of a downstream kernel tree maintained by Shang Yatsen.[^1] It worked well but remained sparsely documented, a lot of the work done here relied on Shang Yatsen's original patch. AOSC OS then picked it up[^2] to provide Intel Xe/Arc support for users of its LoongArch port, for which I worked extensively on. After months of positive user feedback and from encouragement from Kexy Biscuit, my colleague at the community, I decided to examine its potential for upstreaming, cross-reference kernel and Intel documentation to better document and revise this patch. Now that this series has been tested good (for boot up, OpenGL, and playback of a standardised set of video samples[^3] on the following platforms (motherboard + GPU model): - x86-64, 4KiB kernel page: - MS-7D42 + Intel Arc A580 - COLORFIRE B760M-MEOW WIFI D5 + Intel Arc B580 - LoongArch, 16KiB kernel page: - XA61200 + GUNNIR DG1 Blue Halberd (Intel DG1) - XA61200 + GUNNIR Iris Xe Index 4 (Intel DG1) - XA61200 + GUNNIR Intel Iris Xe Max Index V2 (Intel DG1) - XA61200 + GUNNIR Intel Arc A380 Index 6G (Intel Arc A380) - XA61200 + ASRock Arc A380 Challenger ITX OC (Intel Arc A380) - XA61200 + Intel Arc A580 - XA61200 + GUNNIR Intel Arc A750 Photon 8G OC (Intel Arc A750) - XA61200 + Intel Arc B580 - XB612B0 + GUNNIR Intel Iris Xe Max Index V2 (Intel DG1) - XB612B0 + GUNNIR Intel Arc A380 Index 6G (Intel Arc A380) - ASUS XC-LS3A6M + GUNNIR Intel Arc B580 INDEX 12G (Intel Arc B580) On these platforms, basic functionalities tested good but the driver was unstable with occasional resets (I do suspect however, that this platform suffers from PCIe coherence issues, as instability only occurs under heavy VRAM I/O load): - AArch64, 4KiB/64KiB kernel pages: - ERUN-FD3000 (Phytium D3000) + GUNNIR Intel Iris Xe Max Index V2 (Intel DG1) - ERUN-FD3000 (Phytium D3000) + GUNNIR Intel Arc A380 Index 6G (Intel Arc A380) - ERUN-FD3000 (Phytium D3000) + GUNNIR Intel Arc A750 Photon 8G OC (Intel Arc A750) I think that this patch series is now ready for your comment and review. Please forgive me if I made any simple mistake or used wrong terminologies, but I have never worked on a patch for the DRM subsystem and my experience is still quite thin. But anyway, just letting you all know that Intel Xe/Arc works on non-4KiB kernel page platforms (and honestly, it's great to use, especially for games and media playback)! [^1]: https://github.com/FanFansfan/loongson-linux/tree/loongarch-xe [^2]: We maintained Shang Yatsen's patch until our v6.13.3 tree, until we decided to test and send this series upstream, https://github.com/AOSC-Tracking/linux/tree/aosc/v6.13.3 [^3]: Delicious hot pot! https://repo.aosc.io/ahvl/sample-videos-20250223.tar.zst --- Matthew(s), Lucas, and Francois: Thanks again for your patience and review. I recently had a job change and it put me off this series for months, but I'm back (and should be a lot more responsive now) - sorry! Let's get this ball rolling again. I was unfortunately unable to revise 1/5 from v1 as you requested, neither of your suggestions to allow allocation of VRAM smaller than page size worked... So I kept that part as is. As for the your comment in 5/5, I'm not sure about what the right approach to implement a SZ_64K >= PAGE_SIZE assert was, as there are many other instances of similar ternary conditional operators in the xe code. Correct me if I'm wrong but I felt that it might be better handled in a separate patch series? --- Changes in v2: - Define `GUC_ALIGN' and use them in GuC code to improve clarity. - Update documentation on `DRM_XE_QUERY_CONFIG_MIN_ALIGNMENT'. - Rebase, and other minor changes. - Link to v1: https://lore.kernel.org/all/20250226-xe-non-4k-fix-v1-0-80f23b5ee40e@aosc.i… To: Lucas De Marchi <lucas.demarchi(a)intel.com> To: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> To: Rodrigo Vivi <rodrigo.vivi(a)intel.com> To: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com> To: Maxime Ripard <mripard(a)kernel.org> To: Thomas Zimmermann <tzimmermann(a)suse.de> To: David Airlie <airlied(a)gmail.com> To: Simona Vetter <simona(a)ffwll.ch> To: José Roberto de Souza <jose.souza(a)intel.com> To: Francois Dugast <francois.dugast(a)intel.com> To: Matthew Brost <matthew.brost(a)intel.com> To: Alan Previn <alan.previn.teres.alexis(a)intel.com> To: Zhanjun Dong <zhanjun.dong(a)intel.com> To: Matt Roper <matthew.d.roper(a)intel.com> To: Mateusz Naklicki <mateusz.naklicki(a)intel.com> Cc: Mauro Carvalho Chehab <mauro.chehab(a)linux.intel.com> Cc: Zbigniew Kempczyński <zbigniew.kempczynski(a)intel.com> Cc: intel-xe(a)lists.freedesktop.org Cc: dri-devel(a)lists.freedesktop.org Cc: linux-kernel(a)vger.kernel.org Suggested-by: Kexy Biscuit <kexybiscuit(a)aosc.io> Co-developed-by: Shang Yatsen <429839446(a)qq.com> Signed-off-by: Shang Yatsen <429839446(a)qq.com> Signed-off-by: Mingcong Bai <jeffbai(a)aosc.io> --- Mingcong Bai (5): drm/xe/bo: fix alignment with non-4KiB kernel page sizes drm/xe/guc: use GUC_SIZE (SZ_4K) for alignment drm/xe/regs: fix RING_CTL_SIZE(size) calculation drm/xe: use 4KiB alignment for cursor jumps drm/xe/query: use PAGE_SIZE as the minimum page alignment drivers/gpu/drm/xe/regs/xe_engine_regs.h | 2 +- drivers/gpu/drm/xe/xe_bo.c | 8 ++++---- drivers/gpu/drm/xe/xe_guc.c | 4 ++-- drivers/gpu/drm/xe/xe_guc.h | 3 +++ drivers/gpu/drm/xe/xe_guc_ads.c | 32 ++++++++++++++++---------------- drivers/gpu/drm/xe/xe_guc_capture.c | 8 ++++---- drivers/gpu/drm/xe/xe_guc_ct.c | 2 +- drivers/gpu/drm/xe/xe_guc_log.c | 5 +++-- drivers/gpu/drm/xe/xe_guc_pc.c | 4 ++-- drivers/gpu/drm/xe/xe_migrate.c | 4 ++-- drivers/gpu/drm/xe/xe_query.c | 2 +- include/uapi/drm/xe_drm.h | 7 +++++-- 12 files changed, 44 insertions(+), 37 deletions(-) --- base-commit: 546b1c9e93c2bb8cf5ed24e0be1c86bb089b3253 change-id: 20250603-upstream-xe-non-4k-v2-4acf253c9bfd Best regards, -- Mingcong Bai <jeffbai(a)aosc.io>

3 weeks, 1 day

2
6
0 0

+ bcache-remove-unnecessary-select-min_heap.patch added to mm-hotfixes-unstable branch

by Andrew Morton

The patch titled Subject: bcache: remove unnecessary select MIN_HEAP has been added to the -mm mm-hotfixes-unstable branch. Its filename is bcache-remove-unnecessary-select-min_heap.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patche… This patch will later appear in the mm-hotfixes-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Kuan-Wei Chiu <visitorckw(a)gmail.com> Subject: bcache: remove unnecessary select MIN_HEAP Date: Sun, 15 Jun 2025 04:23:53 +0800 After reverting the transition to the generic min heap library, bcache no longer depends on MIN_HEAP. The select entry can be removed to reduce code size and shrink the kernel's attack surface. This change effectively reverts the bcache-related part of commit 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions"). This is part of a series of changes to address a performance regression caused by the use of the generic min_heap implementation. As reported by Robert, bcache now suffers from latency spikes, with P100 (max) latency increasing from 600 ms to 2.4 seconds every 5 minutes. These regressions degrade bcache's effectiveness as a low-latency cache layer and lead to frequent timeouts and application stalls in production environments. Link: https://lore.kernel.org/lkml/CAJhEC05+0S69z+3+FB2Cd0hD+pCRyWTKLEOsc8BOmH73p… Link: https://lkml.kernel.org/r/20250614202353.1632957-4-visitorckw@gmail.com Fixes: 866898efbb25 ("bcache: remove heap-related macros and switch to generic min_heap") Fixes: 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions") Signed-off-by: Kuan-Wei Chiu <visitorckw(a)gmail.com> Reported-by: Robert Pang <robertpang(a)google.com> Closes: https://lore.kernel.org/linux-bcache/CAJhEC06F_AtrPgw2-7CvCqZgeStgCtitbD-ry… Acked-by: Coly Li <colyli(a)kernel.org> Cc: Ching-Chun (Jim) Huang <jserv(a)ccns.ncku.edu.tw> Cc: Kent Overstreet <kent.overstreet(a)linux.dev> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- drivers/md/bcache/Kconfig | 1 - 1 file changed, 1 deletion(-) --- a/drivers/md/bcache/Kconfig~bcache-remove-unnecessary-select-min_heap +++ a/drivers/md/bcache/Kconfig @@ -5,7 +5,6 @@ config BCACHE select BLOCK_HOLDER_DEPRECATED if SYSFS select CRC64 select CLOSURES - select MIN_HEAP help Allows a block device to be used as cache for other devices; uses a btree for indexing and the layout is optimized for SSDs. _ Patches currently in -mm which might be from visitorckw(a)gmail.com are revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap.patch revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap.patch bcache-remove-unnecessary-select-min_heap.patch lib-math-gcd-use-static-key-to-select-implementation-at-runtime.patch riscv-optimize-gcd-code-size-when-config_riscv_isa_zbb-is-disabled.patch riscv-optimize-gcd-performance-on-risc-v-without-zbb-extension.patch

3 weeks, 1 day

1
0
0 0

+ revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap.patch added to mm-hotfixes-unstable branch

by Andrew Morton

The patch titled Subject: Revert "bcache: remove heap-related macros and switch to generic min_heap" has been added to the -mm mm-hotfixes-unstable branch. Its filename is revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patche… This patch will later appear in the mm-hotfixes-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Kuan-Wei Chiu <visitorckw(a)gmail.com> Subject: Revert "bcache: remove heap-related macros and switch to generic min_heap" Date: Sun, 15 Jun 2025 04:23:52 +0800 This reverts commit 866898efbb25bb44fd42848318e46db9e785973a. The generic bottom-up min_heap implementation causes performance regression in invalidate_buckets_lru(), a hot path in bcache. Before the cache is fully populated, new_bucket_prio() often returns zero, leading to many equal comparisons. In such cases, bottom-up sift_down performs up to 2 * log2(n) comparisons, while the original top-down approach completes with just O() comparisons, resulting in a measurable performance gap. The performance degradation is further worsened by the non-inlined min_heap API functions introduced in commit 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions"), adding function call overhead to this critical path. As reported by Robert, bcache now suffers from latency spikes, with P100 (max) latency increasing from 600 ms to 2.4 seconds every 5 minutes. These regressions degrade bcache's effectiveness as a low-latency cache layer and lead to frequent timeouts and application stalls in production environments. This revert aims to restore bcache's original low-latency behavior. Link: https://lore.kernel.org/lkml/CAJhEC05+0S69z+3+FB2Cd0hD+pCRyWTKLEOsc8BOmH73p… Link: https://lkml.kernel.org/r/20250614202353.1632957-3-visitorckw@gmail.com Fixes: 866898efbb25 ("bcache: remove heap-related macros and switch to generic min_heap") Fixes: 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions") Signed-off-by: Kuan-Wei Chiu <visitorckw(a)gmail.com> Reported-by: Robert Pang <robertpang(a)google.com> Closes: https://lore.kernel.org/linux-bcache/CAJhEC06F_AtrPgw2-7CvCqZgeStgCtitbD-ry… Acked-by: Coly Li <colyli(a)kernel.org> Cc: Ching-Chun (Jim) Huang <jserv(a)ccns.ncku.edu.tw> Cc: Kent Overstreet <kent.overstreet(a)linux.dev> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- drivers/md/bcache/alloc.c | 64 ++++------------ drivers/md/bcache/bcache.h | 2 drivers/md/bcache/bset.c | 124 +++++++++++--------------------- drivers/md/bcache/bset.h | 42 ++++++---- drivers/md/bcache/btree.c | 69 +++++++---------- drivers/md/bcache/extents.c | 51 ++++--------- drivers/md/bcache/movinggc.c | 41 ++-------- drivers/md/bcache/super.c | 3 drivers/md/bcache/sysfs.c | 4 - drivers/md/bcache/util.h | 67 ++++++++++++++++- drivers/md/bcache/writeback.c | 13 +-- 11 files changed, 217 insertions(+), 263 deletions(-) --- a/drivers/md/bcache/alloc.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/alloc.c @@ -164,68 +164,40 @@ static void bch_invalidate_one_bucket(st * prio is worth 1/8th of what INITIAL_PRIO is worth. */ -static inline unsigned int new_bucket_prio(struct cache *ca, struct bucket *b) -{ - unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; - - return (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); -} - -static inline bool new_bucket_max_cmp(const void *l, const void *r, void *args) -{ - struct bucket **lhs = (struct bucket **)l; - struct bucket **rhs = (struct bucket **)r; - struct cache *ca = args; - - return new_bucket_prio(ca, *lhs) > new_bucket_prio(ca, *rhs); -} - -static inline bool new_bucket_min_cmp(const void *l, const void *r, void *args) -{ - struct bucket **lhs = (struct bucket **)l; - struct bucket **rhs = (struct bucket **)r; - struct cache *ca = args; - - return new_bucket_prio(ca, *lhs) < new_bucket_prio(ca, *rhs); -} - -static inline void new_bucket_swap(void *l, void *r, void __always_unused *args) -{ - struct bucket **lhs = l, **rhs = r; +#define bucket_prio(b) \ +({ \ + unsigned int min_prio = (INITIAL_PRIO - ca->set->min_prio) / 8; \ + \ + (b->prio - ca->set->min_prio + min_prio) * GC_SECTORS_USED(b); \ +}) - swap(*lhs, *rhs); -} +#define bucket_max_cmp(l, r) (bucket_prio(l) < bucket_prio(r)) +#define bucket_min_cmp(l, r) (bucket_prio(l) > bucket_prio(r)) static void invalidate_buckets_lru(struct cache *ca) { struct bucket *b; - const struct min_heap_callbacks bucket_max_cmp_callback = { - .less = new_bucket_max_cmp, - .swp = new_bucket_swap, - }; - const struct min_heap_callbacks bucket_min_cmp_callback = { - .less = new_bucket_min_cmp, - .swp = new_bucket_swap, - }; + ssize_t i; - ca->heap.nr = 0; + ca->heap.used = 0; for_each_bucket(b, ca) { if (!bch_can_invalidate_bucket(ca, b)) continue; - if (!min_heap_full(&ca->heap)) - min_heap_push(&ca->heap, &b, &bucket_max_cmp_callback, ca); - else if (!new_bucket_max_cmp(&b, min_heap_peek(&ca->heap), ca)) { + if (!heap_full(&ca->heap)) + heap_add(&ca->heap, b, bucket_max_cmp); + else if (bucket_max_cmp(b, heap_peek(&ca->heap))) { ca->heap.data[0] = b; - min_heap_sift_down(&ca->heap, 0, &bucket_max_cmp_callback, ca); + heap_sift(&ca->heap, 0, bucket_max_cmp); } } - min_heapify_all(&ca->heap, &bucket_min_cmp_callback, ca); + for (i = ca->heap.used / 2 - 1; i >= 0; --i) + heap_sift(&ca->heap, i, bucket_min_cmp); while (!fifo_full(&ca->free_inc)) { - if (!ca->heap.nr) { + if (!heap_pop(&ca->heap, b, bucket_min_cmp)) { /* * We don't want to be calling invalidate_buckets() * multiple times when it can't do anything @@ -234,8 +206,6 @@ static void invalidate_buckets_lru(struc wake_up_gc(ca->set); return; } - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &bucket_min_cmp_callback, ca); bch_invalidate_one_bucket(ca, b); } --- a/drivers/md/bcache/bcache.h~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/bcache.h @@ -458,7 +458,7 @@ struct cache { /* Allocation stuff: */ struct bucket *buckets; - DEFINE_MIN_HEAP(struct bucket *, cache_heap) heap; + DECLARE_HEAP(struct bucket *, heap); /* * If nonzero, we know we aren't going to find any buckets to invalidate --- a/drivers/md/bcache/bset.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/bset.c @@ -54,11 +54,9 @@ void bch_dump_bucket(struct btree_keys * int __bch_count_data(struct btree_keys *b) { unsigned int ret = 0; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - if (b->ops->is_extents) for_each_key(b, k, &iter) ret += KEY_SIZE(k); @@ -69,11 +67,9 @@ void __bch_check_keys(struct btree_keys { va_list args; struct bkey *k, *p = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; const char *err; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - for_each_key(b, k, &iter) { if (b->ops->is_extents) { err = "Keys out of order"; @@ -114,9 +110,9 @@ bug: static void bch_btree_iter_next_check(struct btree_iter *iter) { - struct bkey *k = iter->heap.data->k, *next = bkey_next(k); + struct bkey *k = iter->data->k, *next = bkey_next(k); - if (next < iter->heap.data->end && + if (next < iter->data->end && bkey_cmp(k, iter->b->ops->is_extents ? &START_KEY(next) : next) > 0) { bch_dump_bucket(iter->b); @@ -883,14 +879,12 @@ unsigned int bch_btree_insert_key(struct unsigned int status = BTREE_INSERT_STATUS_NO_INSERT; struct bset *i = bset_tree_last(b)->data; struct bkey *m, *prev = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey preceding_key_on_stack = ZERO_KEY; struct bkey *preceding_key_p = &preceding_key_on_stack; BUG_ON(b->ops->is_extents && !KEY_SIZE(k)); - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* * If k has preceding key, preceding_key_p will be set to address * of k's preceding key; otherwise preceding_key_p will be set @@ -901,9 +895,9 @@ unsigned int bch_btree_insert_key(struct else preceding_key(k, &preceding_key_p); - m = bch_btree_iter_init(b, &iter, preceding_key_p); + m = bch_btree_iter_stack_init(b, &iter, preceding_key_p); - if (b->ops->insert_fixup(b, k, &iter, replace_key)) + if (b->ops->insert_fixup(b, k, &iter.iter, replace_key)) return status; status = BTREE_INSERT_STATUS_INSERT; @@ -1083,102 +1077,79 @@ struct bkey *__bch_bset_search(struct bt /* Btree iterator */ -typedef bool (new_btree_iter_cmp_fn)(const void *, const void *, void *); - -static inline bool new_btree_iter_cmp(const void *l, const void *r, void __always_unused *args) -{ - const struct btree_iter_set *_l = l; - const struct btree_iter_set *_r = r; - - return bkey_cmp(_l->k, _r->k) <= 0; -} +typedef bool (btree_iter_cmp_fn)(struct btree_iter_set, + struct btree_iter_set); -static inline void new_btree_iter_swap(void *iter1, void *iter2, void __always_unused *args) +static inline bool btree_iter_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - struct btree_iter_set *_iter1 = iter1; - struct btree_iter_set *_iter2 = iter2; - - swap(*_iter1, *_iter2); + return bkey_cmp(l.k, r.k) > 0; } static inline bool btree_iter_end(struct btree_iter *iter) { - return !iter->heap.nr; + return !iter->used; } void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, struct bkey *end) { - const struct min_heap_callbacks callbacks = { - .less = new_btree_iter_cmp, - .swp = new_btree_iter_swap, - }; - if (k != end) - BUG_ON(!min_heap_push(&iter->heap, - &((struct btree_iter_set) { k, end }), - &callbacks, - NULL)); + BUG_ON(!heap_add(iter, + ((struct btree_iter_set) { k, end }), + btree_iter_cmp)); } -static struct bkey *__bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, - struct bkey *search, - struct bset_tree *start) +static struct bkey *__bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, + struct bkey *search, + struct bset_tree *start) { struct bkey *ret = NULL; - iter->heap.size = ARRAY_SIZE(iter->heap.preallocated); - iter->heap.nr = 0; + iter->iter.size = ARRAY_SIZE(iter->stack_data); + iter->iter.used = 0; #ifdef CONFIG_BCACHE_DEBUG - iter->b = b; + iter->iter.b = b; #endif for (; start <= bset_tree_last(b); start++) { ret = bch_bset_search(b, start, search); - bch_btree_iter_push(iter, ret, bset_bkey_last(start->data)); + bch_btree_iter_push(&iter->iter, ret, bset_bkey_last(start->data)); } return ret; } -struct bkey *bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, +struct bkey *bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, struct bkey *search) { - return __bch_btree_iter_init(b, iter, search, b->set); + return __bch_btree_iter_stack_init(b, iter, search, b->set); } static inline struct bkey *__bch_btree_iter_next(struct btree_iter *iter, - new_btree_iter_cmp_fn *cmp) + btree_iter_cmp_fn *cmp) { struct btree_iter_set b __maybe_unused; struct bkey *ret = NULL; - const struct min_heap_callbacks callbacks = { - .less = cmp, - .swp = new_btree_iter_swap, - }; if (!btree_iter_end(iter)) { bch_btree_iter_next_check(iter); - ret = iter->heap.data->k; - iter->heap.data->k = bkey_next(iter->heap.data->k); + ret = iter->data->k; + iter->data->k = bkey_next(iter->data->k); - if (iter->heap.data->k > iter->heap.data->end) { + if (iter->data->k > iter->data->end) { WARN_ONCE(1, "bset was corrupt!\n"); - iter->heap.data->k = iter->heap.data->end; + iter->data->k = iter->data->end; } - if (iter->heap.data->k == iter->heap.data->end) { - if (iter->heap.nr) { - b = min_heap_peek(&iter->heap)[0]; - min_heap_pop(&iter->heap, &callbacks, NULL); - } - } + if (iter->data->k == iter->data->end) + heap_pop(iter, b, cmp); else - min_heap_sift_down(&iter->heap, 0, &callbacks, NULL); + heap_sift(iter, 0, cmp); } return ret; @@ -1186,7 +1157,7 @@ static inline struct bkey *__bch_btree_i struct bkey *bch_btree_iter_next(struct btree_iter *iter) { - return __bch_btree_iter_next(iter, new_btree_iter_cmp); + return __bch_btree_iter_next(iter, btree_iter_cmp); } @@ -1224,18 +1195,16 @@ static void btree_mergesort(struct btree struct btree_iter *iter, bool fixup, bool remove_stale) { + int i; struct bkey *k, *last = NULL; BKEY_PADDED(k) tmp; bool (*bad)(struct btree_keys *, const struct bkey *) = remove_stale ? bch_ptr_bad : bch_ptr_invalid; - const struct min_heap_callbacks callbacks = { - .less = b->ops->sort_cmp, - .swp = new_btree_iter_swap, - }; /* Heapify the iterator, using our comparison function */ - min_heapify_all(&iter->heap, &callbacks, NULL); + for (i = iter->used / 2 - 1; i >= 0; --i) + heap_sift(iter, i, b->ops->sort_cmp); while (!btree_iter_end(iter)) { if (b->ops->sort_fixup && fixup) @@ -1324,11 +1293,10 @@ void bch_btree_sort_partial(struct btree struct bset_sort_state *state) { size_t order = b->page_order, keys = 0; - struct btree_iter iter; + struct btree_iter_stack iter; int oldsize = bch_count_data(b); - min_heap_init(&iter.heap, NULL, MAX_BSETS); - __bch_btree_iter_init(b, &iter, NULL, &b->set[start]); + __bch_btree_iter_stack_init(b, &iter, NULL, &b->set[start]); if (start) { unsigned int i; @@ -1339,7 +1307,7 @@ void bch_btree_sort_partial(struct btree order = get_order(__set_bytes(b->set->data, keys)); } - __btree_sort(b, &iter, start, order, false, state); + __btree_sort(b, &iter.iter, start, order, false, state); EBUG_ON(oldsize >= 0 && bch_count_data(b) != oldsize); } @@ -1355,13 +1323,11 @@ void bch_btree_sort_into(struct btree_ke struct bset_sort_state *state) { uint64_t start_time = local_clock(); - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; - bch_btree_iter_init(b, &iter, NULL); + bch_btree_iter_stack_init(b, &iter, NULL); - btree_mergesort(b, new->set->data, &iter, false, true); + btree_mergesort(b, new->set->data, &iter.iter, false, true); bch_time_stats_update(&state->time, start_time); --- a/drivers/md/bcache/bset.h~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/bset.h @@ -187,9 +187,8 @@ struct bset_tree { }; struct btree_keys_ops { - bool (*sort_cmp)(const void *l, - const void *r, - void *args); + bool (*sort_cmp)(struct btree_iter_set l, + struct btree_iter_set r); struct bkey *(*sort_fixup)(struct btree_iter *iter, struct bkey *tmp); bool (*insert_fixup)(struct btree_keys *b, @@ -313,17 +312,23 @@ enum { BTREE_INSERT_STATUS_FRONT_MERGE, }; -struct btree_iter_set { - struct bkey *k, *end; -}; - /* Btree key iteration */ struct btree_iter { + size_t size, used; #ifdef CONFIG_BCACHE_DEBUG struct btree_keys *b; #endif - MIN_HEAP_PREALLOCATED(struct btree_iter_set, btree_iter_heap, MAX_BSETS) heap; + struct btree_iter_set { + struct bkey *k, *end; + } data[]; +}; + +/* Fixed-size btree_iter that can be allocated on the stack */ + +struct btree_iter_stack { + struct btree_iter iter; + struct btree_iter_set stack_data[MAX_BSETS]; }; typedef bool (*ptr_filter_fn)(struct btree_keys *b, const struct bkey *k); @@ -335,9 +340,9 @@ struct bkey *bch_btree_iter_next_filter( void bch_btree_iter_push(struct btree_iter *iter, struct bkey *k, struct bkey *end); -struct bkey *bch_btree_iter_init(struct btree_keys *b, - struct btree_iter *iter, - struct bkey *search); +struct bkey *bch_btree_iter_stack_init(struct btree_keys *b, + struct btree_iter_stack *iter, + struct bkey *search); struct bkey *__bch_bset_search(struct btree_keys *b, struct bset_tree *t, const struct bkey *search); @@ -352,13 +357,14 @@ static inline struct bkey *bch_bset_sear return search ? __bch_bset_search(b, t, search) : t->data->start; } -#define for_each_key_filter(b, k, iter, filter) \ - for (bch_btree_iter_init((b), (iter), NULL); \ - ((k) = bch_btree_iter_next_filter((iter), (b), filter));) - -#define for_each_key(b, k, iter) \ - for (bch_btree_iter_init((b), (iter), NULL); \ - ((k) = bch_btree_iter_next(iter));) +#define for_each_key_filter(b, k, stack_iter, filter) \ + for (bch_btree_iter_stack_init((b), (stack_iter), NULL); \ + ((k) = bch_btree_iter_next_filter(&((stack_iter)->iter), (b), \ + filter));) + +#define for_each_key(b, k, stack_iter) \ + for (bch_btree_iter_stack_init((b), (stack_iter), NULL); \ + ((k) = bch_btree_iter_next(&((stack_iter)->iter)));) /* Sorting */ --- a/drivers/md/bcache/btree.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/btree.c @@ -148,19 +148,19 @@ void bch_btree_node_read_done(struct btr { const char *err = "bad btree header"; struct bset *i = btree_bset_first(b); - struct btree_iter iter; + struct btree_iter *iter; /* * c->fill_iter can allocate an iterator with more memory space * than static MAX_BSETS. * See the comment arount cache_set->fill_iter. */ - iter.heap.data = mempool_alloc(&b->c->fill_iter, GFP_NOIO); - iter.heap.size = b->c->cache->sb.bucket_size / b->c->cache->sb.block_size; - iter.heap.nr = 0; + iter = mempool_alloc(&b->c->fill_iter, GFP_NOIO); + iter->size = b->c->cache->sb.bucket_size / b->c->cache->sb.block_size; + iter->used = 0; #ifdef CONFIG_BCACHE_DEBUG - iter.b = &b->keys; + iter->b = &b->keys; #endif if (!i->seq) @@ -198,7 +198,7 @@ void bch_btree_node_read_done(struct btr if (i != b->keys.set[0].data && !i->keys) goto err; - bch_btree_iter_push(&iter, i->start, bset_bkey_last(i)); + bch_btree_iter_push(iter, i->start, bset_bkey_last(i)); b->written += set_blocks(i, block_bytes(b->c->cache)); } @@ -210,7 +210,7 @@ void bch_btree_node_read_done(struct btr if (i->seq == b->keys.set[0].data->seq) goto err; - bch_btree_sort_and_fix_extents(&b->keys, &iter, &b->c->sort); + bch_btree_sort_and_fix_extents(&b->keys, iter, &b->c->sort); i = b->keys.set[0].data; err = "short btree key"; @@ -222,7 +222,7 @@ void bch_btree_node_read_done(struct btr bch_bset_init_next(&b->keys, write_block(b), bset_magic(&b->c->cache->sb)); out: - mempool_free(iter.heap.data, &b->c->fill_iter); + mempool_free(iter, &b->c->fill_iter); return; err: set_btree_node_io_error(b); @@ -1306,11 +1306,9 @@ static bool btree_gc_mark_node(struct bt uint8_t stale = 0; unsigned int keys = 0, good_keys = 0; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; struct bset_tree *t; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - gc->nodes++; for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) { @@ -1569,11 +1567,9 @@ static int btree_gc_rewrite_node(struct static unsigned int btree_gc_count_keys(struct btree *b) { struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; unsigned int ret = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad) ret += bkey_u64s(k); @@ -1612,18 +1608,18 @@ static int btree_gc_recurse(struct btree int ret = 0; bool should_rewrite; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; struct gc_merge_info r[GC_MERGE_NODES]; struct gc_merge_info *i, *last = r + ARRAY_SIZE(r) - 1; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, &b->c->gc_done); + bch_btree_iter_stack_init(&b->keys, &iter, &b->c->gc_done); for (i = r; i < r + ARRAY_SIZE(r); i++) i->b = ERR_PTR(-EINTR); while (1) { - k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad); + k = bch_btree_iter_next_filter(&iter.iter, &b->keys, + bch_ptr_bad); if (k) { r->b = bch_btree_node_get(b->c, op, k, b->level - 1, true, b); @@ -1918,9 +1914,7 @@ static int bch_btree_check_recurse(struc { int ret = 0; struct bkey *k, *p = NULL; - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) bch_initial_mark_key(b->c, b->level, k); @@ -1928,10 +1922,10 @@ static int bch_btree_check_recurse(struc bch_initial_mark_key(b->c, b->level + 1, &b->key); if (b->level) { - bch_btree_iter_init(&b->keys, &iter, NULL); + bch_btree_iter_stack_init(&b->keys, &iter, NULL); do { - k = bch_btree_iter_next_filter(&iter, &b->keys, + k = bch_btree_iter_next_filter(&iter.iter, &b->keys, bch_ptr_bad); if (k) { btree_node_prefetch(b, k); @@ -1959,7 +1953,7 @@ static int bch_btree_check_thread(void * struct btree_check_info *info = arg; struct btree_check_state *check_state = info->state; struct cache_set *c = check_state->c; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k, *p; int cur_idx, prev_idx, skip_nr; @@ -1967,11 +1961,9 @@ static int bch_btree_check_thread(void * cur_idx = prev_idx = 0; ret = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* root node keys are checked before thread created */ - bch_btree_iter_init(&c->root->keys, &iter, NULL); - k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + bch_btree_iter_stack_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); BUG_ON(!k); p = k; @@ -1989,7 +1981,7 @@ static int bch_btree_check_thread(void * skip_nr = cur_idx - prev_idx; while (skip_nr) { - k = bch_btree_iter_next_filter(&iter, + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); if (k) @@ -2062,11 +2054,9 @@ int bch_btree_check(struct cache_set *c) int ret = 0; int i; struct bkey *k = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct btree_check_state check_state; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - /* check and mark root node keys */ for_each_key_filter(&c->root->keys, k, &iter, bch_ptr_invalid) bch_initial_mark_key(c, c->root->level, k); @@ -2560,12 +2550,11 @@ static int bch_btree_map_nodes_recurse(s if (b->level) { struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, from); + bch_btree_iter_stack_init(&b->keys, &iter, from); - while ((k = bch_btree_iter_next_filter(&iter, &b->keys, + while ((k = bch_btree_iter_next_filter(&iter.iter, &b->keys, bch_ptr_bad))) { ret = bcache_btree(map_nodes_recurse, k, b, op, from, fn, flags); @@ -2594,12 +2583,12 @@ int bch_btree_map_keys_recurse(struct bt { int ret = MAP_CONTINUE; struct bkey *k; - struct btree_iter iter; + struct btree_iter_stack iter; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&b->keys, &iter, from); + bch_btree_iter_stack_init(&b->keys, &iter, from); - while ((k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad))) { + while ((k = bch_btree_iter_next_filter(&iter.iter, &b->keys, + bch_ptr_bad))) { ret = !b->level ? fn(op, b, k) : bcache_btree(map_keys_recurse, k, --- a/drivers/md/bcache/extents.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/extents.c @@ -33,16 +33,15 @@ static void sort_key_next(struct btree_i i->k = bkey_next(i->k); if (i->k == i->end) - *i = iter->heap.data[--iter->heap.nr]; + *i = iter->data[--iter->used]; } -static bool new_bch_key_sort_cmp(const void *l, const void *r, void *args) +static bool bch_key_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - struct btree_iter_set *_l = (struct btree_iter_set *)l; - struct btree_iter_set *_r = (struct btree_iter_set *)r; - int64_t c = bkey_cmp(_l->k, _r->k); + int64_t c = bkey_cmp(l.k, r.k); - return !(c ? c > 0 : _l->k < _r->k); + return c ? c > 0 : l.k < r.k; } static bool __ptr_invalid(struct cache_set *c, const struct bkey *k) @@ -239,7 +238,7 @@ static bool bch_btree_ptr_insert_fixup(s } const struct btree_keys_ops bch_btree_keys_ops = { - .sort_cmp = new_bch_key_sort_cmp, + .sort_cmp = bch_key_sort_cmp, .insert_fixup = bch_btree_ptr_insert_fixup, .key_invalid = bch_btree_ptr_invalid, .key_bad = bch_btree_ptr_bad, @@ -256,36 +255,22 @@ const struct btree_keys_ops bch_btree_ke * Necessary for btree_sort_fixup() - if there are multiple keys that compare * equal in different sets, we have to process them newest to oldest. */ - -static bool new_bch_extent_sort_cmp(const void *l, const void *r, void __always_unused *args) -{ - struct btree_iter_set *_l = (struct btree_iter_set *)l; - struct btree_iter_set *_r = (struct btree_iter_set *)r; - int64_t c = bkey_cmp(&START_KEY(_l->k), &START_KEY(_r->k)); - - return !(c ? c > 0 : _l->k < _r->k); -} - -static inline void new_btree_iter_swap(void *iter1, void *iter2, void __always_unused *args) +static bool bch_extent_sort_cmp(struct btree_iter_set l, + struct btree_iter_set r) { - struct btree_iter_set *_iter1 = iter1; - struct btree_iter_set *_iter2 = iter2; + int64_t c = bkey_cmp(&START_KEY(l.k), &START_KEY(r.k)); - swap(*_iter1, *_iter2); + return c ? c > 0 : l.k < r.k; } static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, struct bkey *tmp) { - const struct min_heap_callbacks callbacks = { - .less = new_bch_extent_sort_cmp, - .swp = new_btree_iter_swap, - }; - while (iter->heap.nr > 1) { - struct btree_iter_set *top = iter->heap.data, *i = top + 1; + while (iter->used > 1) { + struct btree_iter_set *top = iter->data, *i = top + 1; - if (iter->heap.nr > 2 && - !new_bch_extent_sort_cmp(&i[0], &i[1], NULL)) + if (iter->used > 2 && + bch_extent_sort_cmp(i[0], i[1])) i++; if (bkey_cmp(top->k, &START_KEY(i->k)) <= 0) @@ -293,7 +278,7 @@ static struct bkey *bch_extent_sort_fixu if (!KEY_SIZE(i->k)) { sort_key_next(iter, i); - min_heap_sift_down(&iter->heap, i - top, &callbacks, NULL); + heap_sift(iter, i - top, bch_extent_sort_cmp); continue; } @@ -303,7 +288,7 @@ static struct bkey *bch_extent_sort_fixu else bch_cut_front(top->k, i->k); - min_heap_sift_down(&iter->heap, i - top, &callbacks, NULL); + heap_sift(iter, i - top, bch_extent_sort_cmp); } else { /* can't happen because of comparison func */ BUG_ON(!bkey_cmp(&START_KEY(top->k), &START_KEY(i->k))); @@ -313,7 +298,7 @@ static struct bkey *bch_extent_sort_fixu bch_cut_back(&START_KEY(i->k), tmp); bch_cut_front(i->k, top->k); - min_heap_sift_down(&iter->heap, 0, &callbacks, NULL); + heap_sift(iter, 0, bch_extent_sort_cmp); return tmp; } else { @@ -633,7 +618,7 @@ static bool bch_extent_merge(struct btre } const struct btree_keys_ops bch_extent_keys_ops = { - .sort_cmp = new_bch_extent_sort_cmp, + .sort_cmp = bch_extent_sort_cmp, .sort_fixup = bch_extent_sort_fixup, .insert_fixup = bch_extent_insert_fixup, .key_invalid = bch_extent_invalid, --- a/drivers/md/bcache/movinggc.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/movinggc.c @@ -182,27 +182,16 @@ err: if (!IS_ERR_OR_NULL(w->private)) closure_sync(&cl); } -static bool new_bucket_cmp(const void *l, const void *r, void __always_unused *args) +static bool bucket_cmp(struct bucket *l, struct bucket *r) { - struct bucket **_l = (struct bucket **)l; - struct bucket **_r = (struct bucket **)r; - - return GC_SECTORS_USED(*_l) >= GC_SECTORS_USED(*_r); -} - -static void new_bucket_swap(void *l, void *r, void __always_unused *args) -{ - struct bucket **_l = l; - struct bucket **_r = r; - - swap(*_l, *_r); + return GC_SECTORS_USED(l) < GC_SECTORS_USED(r); } static unsigned int bucket_heap_top(struct cache *ca) { struct bucket *b; - return (b = min_heap_peek(&ca->heap)[0]) ? GC_SECTORS_USED(b) : 0; + return (b = heap_peek(&ca->heap)) ? GC_SECTORS_USED(b) : 0; } void bch_moving_gc(struct cache_set *c) @@ -210,10 +199,6 @@ void bch_moving_gc(struct cache_set *c) struct cache *ca = c->cache; struct bucket *b; unsigned long sectors_to_move, reserve_sectors; - const struct min_heap_callbacks callbacks = { - .less = new_bucket_cmp, - .swp = new_bucket_swap, - }; if (!c->copy_gc_enabled) return; @@ -224,7 +209,7 @@ void bch_moving_gc(struct cache_set *c) reserve_sectors = ca->sb.bucket_size * fifo_used(&ca->free[RESERVE_MOVINGGC]); - ca->heap.nr = 0; + ca->heap.used = 0; for_each_bucket(b, ca) { if (GC_MARK(b) == GC_MARK_METADATA || @@ -233,31 +218,25 @@ void bch_moving_gc(struct cache_set *c) atomic_read(&b->pin)) continue; - if (!min_heap_full(&ca->heap)) { + if (!heap_full(&ca->heap)) { sectors_to_move += GC_SECTORS_USED(b); - min_heap_push(&ca->heap, &b, &callbacks, NULL); - } else if (!new_bucket_cmp(&b, min_heap_peek(&ca->heap), ca)) { + heap_add(&ca->heap, b, bucket_cmp); + } else if (bucket_cmp(b, heap_peek(&ca->heap))) { sectors_to_move -= bucket_heap_top(ca); sectors_to_move += GC_SECTORS_USED(b); ca->heap.data[0] = b; - min_heap_sift_down(&ca->heap, 0, &callbacks, NULL); + heap_sift(&ca->heap, 0, bucket_cmp); } } while (sectors_to_move > reserve_sectors) { - if (ca->heap.nr) { - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &callbacks, NULL); - } + heap_pop(&ca->heap, b, bucket_cmp); sectors_to_move -= GC_SECTORS_USED(b); } - while (ca->heap.nr) { - b = min_heap_peek(&ca->heap)[0]; - min_heap_pop(&ca->heap, &callbacks, NULL); + while (heap_pop(&ca->heap, b, bucket_cmp)) SET_GC_MOVE(b, 1); - } mutex_unlock(&c->bucket_lock); --- a/drivers/md/bcache/super.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/super.c @@ -1912,7 +1912,8 @@ struct cache_set *bch_cache_set_alloc(st INIT_LIST_HEAD(&c->btree_cache_freed); INIT_LIST_HEAD(&c->data_buckets); - iter_size = ((meta_bucket_pages(sb) * PAGE_SECTORS) / sb->block_size) * + iter_size = sizeof(struct btree_iter) + + ((meta_bucket_pages(sb) * PAGE_SECTORS) / sb->block_size) * sizeof(struct btree_iter_set); c->devices = kcalloc(c->nr_uuids, sizeof(void *), GFP_KERNEL); --- a/drivers/md/bcache/sysfs.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/sysfs.c @@ -660,9 +660,7 @@ static unsigned int bch_root_usage(struc unsigned int bytes = 0; struct bkey *k; struct btree *b; - struct btree_iter iter; - - min_heap_init(&iter.heap, NULL, MAX_BSETS); + struct btree_iter_stack iter; goto lock_root; --- a/drivers/md/bcache/util.h~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/util.h @@ -9,7 +9,6 @@ #include <linux/kernel.h> #include <linux/sched/clock.h> #include <linux/llist.h> -#include <linux/min_heap.h> #include <linux/ratelimit.h> #include <linux/vmalloc.h> #include <linux/workqueue.h> @@ -31,10 +30,16 @@ struct closure; #endif +#define DECLARE_HEAP(type, name) \ + struct { \ + size_t size, used; \ + type *data; \ + } name + #define init_heap(heap, _size, gfp) \ ({ \ size_t _bytes; \ - (heap)->nr = 0; \ + (heap)->used = 0; \ (heap)->size = (_size); \ _bytes = (heap)->size * sizeof(*(heap)->data); \ (heap)->data = kvmalloc(_bytes, (gfp) & GFP_KERNEL); \ @@ -47,6 +52,64 @@ do { \ (heap)->data = NULL; \ } while (0) +#define heap_swap(h, i, j) swap((h)->data[i], (h)->data[j]) + +#define heap_sift(h, i, cmp) \ +do { \ + size_t _r, _j = i; \ + \ + for (; _j * 2 + 1 < (h)->used; _j = _r) { \ + _r = _j * 2 + 1; \ + if (_r + 1 < (h)->used && \ + cmp((h)->data[_r], (h)->data[_r + 1])) \ + _r++; \ + \ + if (cmp((h)->data[_r], (h)->data[_j])) \ + break; \ + heap_swap(h, _r, _j); \ + } \ +} while (0) + +#define heap_sift_down(h, i, cmp) \ +do { \ + while (i) { \ + size_t p = (i - 1) / 2; \ + if (cmp((h)->data[i], (h)->data[p])) \ + break; \ + heap_swap(h, i, p); \ + i = p; \ + } \ +} while (0) + +#define heap_add(h, d, cmp) \ +({ \ + bool _r = !heap_full(h); \ + if (_r) { \ + size_t _i = (h)->used++; \ + (h)->data[_i] = d; \ + \ + heap_sift_down(h, _i, cmp); \ + heap_sift(h, _i, cmp); \ + } \ + _r; \ +}) + +#define heap_pop(h, d, cmp) \ +({ \ + bool _r = (h)->used; \ + if (_r) { \ + (d) = (h)->data[0]; \ + (h)->used--; \ + heap_swap(h, 0, (h)->used); \ + heap_sift(h, 0, cmp); \ + } \ + _r; \ +}) + +#define heap_peek(h) ((h)->used ? (h)->data[0] : NULL) + +#define heap_full(h) ((h)->used == (h)->size) + #define DECLARE_FIFO(type, name) \ struct { \ size_t front, back, size, mask; \ --- a/drivers/md/bcache/writeback.c~revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap +++ a/drivers/md/bcache/writeback.c @@ -908,16 +908,15 @@ static int bch_dirty_init_thread(void *a struct dirty_init_thrd_info *info = arg; struct bch_dirty_init_state *state = info->state; struct cache_set *c = state->c; - struct btree_iter iter; + struct btree_iter_stack iter; struct bkey *k, *p; int cur_idx, prev_idx, skip_nr; k = p = NULL; prev_idx = 0; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - bch_btree_iter_init(&c->root->keys, &iter, NULL); - k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad); + bch_btree_iter_stack_init(&c->root->keys, &iter, NULL); + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); BUG_ON(!k); p = k; @@ -931,7 +930,7 @@ static int bch_dirty_init_thread(void *a skip_nr = cur_idx - prev_idx; while (skip_nr) { - k = bch_btree_iter_next_filter(&iter, + k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad); if (k) @@ -980,13 +979,11 @@ void bch_sectors_dirty_init(struct bcach int i; struct btree *b = NULL; struct bkey *k = NULL; - struct btree_iter iter; + struct btree_iter_stack iter; struct sectors_dirty_init op; struct cache_set *c = d->c; struct bch_dirty_init_state state; - min_heap_init(&iter.heap, NULL, MAX_BSETS); - retry_lock: b = c->root; rw_lock(0, b, b->level); _ Patches currently in -mm which might be from visitorckw(a)gmail.com are revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap.patch revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap.patch bcache-remove-unnecessary-select-min_heap.patch lib-math-gcd-use-static-key-to-select-implementation-at-runtime.patch riscv-optimize-gcd-code-size-when-config_riscv_isa_zbb-is-disabled.patch riscv-optimize-gcd-performance-on-risc-v-without-zbb-extension.patch

3 weeks, 1 day

1
0
0 0

+ revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap.patch added to mm-hotfixes-unstable branch

by Andrew Morton

The patch titled Subject: Revert "bcache: update min_heap_callbacks to use default builtin swap" has been added to the -mm mm-hotfixes-unstable branch. Its filename is revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patche… This patch will later appear in the mm-hotfixes-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Kuan-Wei Chiu <visitorckw(a)gmail.com> Subject: Revert "bcache: update min_heap_callbacks to use default builtin swap" Date: Sun, 15 Jun 2025 04:23:51 +0800 Patch series "bcache: Revert min_heap migration due to performance regression". This patch series reverts the migration of bcache from its original heap implementation to the generic min_heap library. While the original change aimed to simplify the code and improve maintainability, it introduced a severe performance regression in real-world scenarios. As reported by Robert, systems using bcache now suffer from periodic latency spikes, with P100 (max) latency increasing from 600 ms to 2.4 seconds every 5 minutes. This degrades bcache's value as a low-latency caching layer, and leads to frequent timeouts and application stalls in production environments. The primary cause of this regression is the behavior of the generic min_heap implementation's bottom-up sift_down, which performs up to 2 * log2(n) comparisons when many elements are equal. The original top-down variant used by bcache only required O(1) comparisons in such cases. The issue was further exacerbated by commit 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions"), which introduced non-inlined versions of the min_heap API, adding function call overhead to a performance-critical hot path. This patch (of 3): This reverts commit 3d8a9a1c35227c3f1b0bd132c9f0a80dbda07b65. Although removing the custom swap function simplified the code, this change is part of a broader migration to the generic min_heap API that introduced significant performance regressions in bcache. As reported by Robert, bcache now suffers from latency spikes, with P100 (max) latency increasing from 600 ms to 2.4 seconds every 5 minutes. These regressions degrade bcache's effectiveness as a low-latency cache layer and lead to frequent timeouts and application stalls in production environments. This revert is part of a series of changes to restore previous performance by undoing the min_heap transition. Link: https://lkml.kernel.org/r/20250614202353.1632957-1-visitorckw@gmail.com Link: https://lore.kernel.org/lkml/CAJhEC05+0S69z+3+FB2Cd0hD+pCRyWTKLEOsc8BOmH73p… Link: https://lkml.kernel.org/r/20250614202353.1632957-2-visitorckw@gmail.com Fixes: 866898efbb25 ("bcache: remove heap-related macros and switch to generic min_heap") Fixes: 92a8b224b833 ("lib/min_heap: introduce non-inline versions of min heap API functions") Signed-off-by: Kuan-Wei Chiu <visitorckw(a)gmail.com> Reported-by: Robert Pang <robertpang(a)google.com> Closes: https://lore.kernel.org/linux-bcache/CAJhEC06F_AtrPgw2-7CvCqZgeStgCtitbD-ry… Acked-by: Coly Li <colyli(a)kernel.org> Cc: Ching-Chun (Jim) Huang <jserv(a)ccns.ncku.edu.tw> Cc: Kent Overstreet <kent.overstreet(a)linux.dev> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- drivers/md/bcache/alloc.c | 11 +++++++++-- drivers/md/bcache/bset.c | 14 +++++++++++--- drivers/md/bcache/extents.c | 10 +++++++++- drivers/md/bcache/movinggc.c | 10 +++++++++- 4 files changed, 38 insertions(+), 7 deletions(-) --- a/drivers/md/bcache/alloc.c~revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap +++ a/drivers/md/bcache/alloc.c @@ -189,16 +189,23 @@ static inline bool new_bucket_min_cmp(co return new_bucket_prio(ca, *lhs) < new_bucket_prio(ca, *rhs); } +static inline void new_bucket_swap(void *l, void *r, void __always_unused *args) +{ + struct bucket **lhs = l, **rhs = r; + + swap(*lhs, *rhs); +} + static void invalidate_buckets_lru(struct cache *ca) { struct bucket *b; const struct min_heap_callbacks bucket_max_cmp_callback = { .less = new_bucket_max_cmp, - .swp = NULL, + .swp = new_bucket_swap, }; const struct min_heap_callbacks bucket_min_cmp_callback = { .less = new_bucket_min_cmp, - .swp = NULL, + .swp = new_bucket_swap, }; ca->heap.nr = 0; --- a/drivers/md/bcache/bset.c~revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap +++ a/drivers/md/bcache/bset.c @@ -1093,6 +1093,14 @@ static inline bool new_btree_iter_cmp(co return bkey_cmp(_l->k, _r->k) <= 0; } +static inline void new_btree_iter_swap(void *iter1, void *iter2, void __always_unused *args) +{ + struct btree_iter_set *_iter1 = iter1; + struct btree_iter_set *_iter2 = iter2; + + swap(*_iter1, *_iter2); +} + static inline bool btree_iter_end(struct btree_iter *iter) { return !iter->heap.nr; @@ -1103,7 +1111,7 @@ void bch_btree_iter_push(struct btree_it { const struct min_heap_callbacks callbacks = { .less = new_btree_iter_cmp, - .swp = NULL, + .swp = new_btree_iter_swap, }; if (k != end) @@ -1149,7 +1157,7 @@ static inline struct bkey *__bch_btree_i struct bkey *ret = NULL; const struct min_heap_callbacks callbacks = { .less = cmp, - .swp = NULL, + .swp = new_btree_iter_swap, }; if (!btree_iter_end(iter)) { @@ -1223,7 +1231,7 @@ static void btree_mergesort(struct btree : bch_ptr_invalid; const struct min_heap_callbacks callbacks = { .less = b->ops->sort_cmp, - .swp = NULL, + .swp = new_btree_iter_swap, }; /* Heapify the iterator, using our comparison function */ --- a/drivers/md/bcache/extents.c~revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap +++ a/drivers/md/bcache/extents.c @@ -266,12 +266,20 @@ static bool new_bch_extent_sort_cmp(cons return !(c ? c > 0 : _l->k < _r->k); } +static inline void new_btree_iter_swap(void *iter1, void *iter2, void __always_unused *args) +{ + struct btree_iter_set *_iter1 = iter1; + struct btree_iter_set *_iter2 = iter2; + + swap(*_iter1, *_iter2); +} + static struct bkey *bch_extent_sort_fixup(struct btree_iter *iter, struct bkey *tmp) { const struct min_heap_callbacks callbacks = { .less = new_bch_extent_sort_cmp, - .swp = NULL, + .swp = new_btree_iter_swap, }; while (iter->heap.nr > 1) { struct btree_iter_set *top = iter->heap.data, *i = top + 1; --- a/drivers/md/bcache/movinggc.c~revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap +++ a/drivers/md/bcache/movinggc.c @@ -190,6 +190,14 @@ static bool new_bucket_cmp(const void *l return GC_SECTORS_USED(*_l) >= GC_SECTORS_USED(*_r); } +static void new_bucket_swap(void *l, void *r, void __always_unused *args) +{ + struct bucket **_l = l; + struct bucket **_r = r; + + swap(*_l, *_r); +} + static unsigned int bucket_heap_top(struct cache *ca) { struct bucket *b; @@ -204,7 +212,7 @@ void bch_moving_gc(struct cache_set *c) unsigned long sectors_to_move, reserve_sectors; const struct min_heap_callbacks callbacks = { .less = new_bucket_cmp, - .swp = NULL, + .swp = new_bucket_swap, }; if (!c->copy_gc_enabled) _ Patches currently in -mm which might be from visitorckw(a)gmail.com are revert-bcache-update-min_heap_callbacks-to-use-default-builtin-swap.patch revert-bcache-remove-heap-related-macros-and-switch-to-generic-min_heap.patch bcache-remove-unnecessary-select-min_heap.patch lib-math-gcd-use-static-key-to-select-implementation-at-runtime.patch riscv-optimize-gcd-code-size-when-config_riscv_isa_zbb-is-disabled.patch riscv-optimize-gcd-performance-on-risc-v-without-zbb-extension.patch

3 weeks, 1 day

1
0
0 0

[PATCH] kallsyms: fix build without execinfo

by Achill Gilgenast

Signed-off-by: Achill Gilgenast <fossdd(a)pwned.life> Cc: stable(a)vger.kernel.org --- tools/include/linux/kallsyms.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tools/include/linux/kallsyms.h b/tools/include/linux/kallsyms.h index 5a37ccbec54f..f61a01dd7eb7 100644 --- a/tools/include/linux/kallsyms.h +++ b/tools/include/linux/kallsyms.h @@ -18,6 +18,7 @@ static inline const char *kallsyms_lookup(unsigned long addr, return NULL; } +#ifdef HAVE_BACKTRACE_SUPPORT #include <execinfo.h> #include <stdlib.h> static inline void print_ip_sym(const char *loglvl, unsigned long ip) @@ -30,5 +31,8 @@ static inline void print_ip_sym(const char *loglvl, unsigned long ip) free(name); } +#else +static inline void print_ip_sym(const char *loglvl, unsigned long ip) {} +#endif #endif -- 2.49.0

3 weeks, 1 day

3
2
0 0

[PATCH v2] kallsyms: fix build without execinfo

by Achill Gilgenast

Some libc's like musl libc don't provide execinfo.h since it's not part of POSIX. In order to fix compilation on musl, only include execinfo.h if available (HAVE_BACKTRACE_SUPPORT) This was discovered with c104c16073b7 ("Kunit to check the longest symbol length") which starts to include linux/kallsyms.h with Alpine Linux' configs. Signed-off-by: Achill Gilgenast <fossdd(a)pwned.life> Cc: stable(a)vger.kernel.org --- tools/include/linux/kallsyms.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/tools/include/linux/kallsyms.h b/tools/include/linux/kallsyms.h index 5a37ccbec54f..f61a01dd7eb7 100644 --- a/tools/include/linux/kallsyms.h +++ b/tools/include/linux/kallsyms.h @@ -18,6 +18,7 @@ static inline const char *kallsyms_lookup(unsigned long addr, return NULL; } +#ifdef HAVE_BACKTRACE_SUPPORT #include <execinfo.h> #include <stdlib.h> static inline void print_ip_sym(const char *loglvl, unsigned long ip) @@ -30,5 +31,8 @@ static inline void print_ip_sym(const char *loglvl, unsigned long ip) free(name); } +#else +static inline void print_ip_sym(const char *loglvl, unsigned long ip) {} +#endif #endif -- 2.50.0.rc2

3 weeks, 1 day

1
0
0 0

[PATCH 5.10.y 0/5] Backport few sfq fixes

by Harshit Mogalapalli

commit: 10685681bafc ("net_sched: sch_sfq: don't allow 1 packet limit") fixes CVE-2024-57996 and commit: b3bf8f63e617 ("net_sched: sch_sfq: move the limit validation") fixes CVE-2025-37752. Patches 3 and 5 are CVE fixes for above mentioned CVEs. Patch 1,2 and 4 are pulled in as stable-deps. Testing performed on the patched 5.10.238 kernel with the above 5 patches: (Used latest upstream kselftests for tc-testing) # uname -a Linux hamogala-vm-6 5.10.238+ #2 SMP Sun Jun 15 17:27:54 GMT 2025 x86_64 x86_64 x86_64 GNU/Linux # ./tdc.py -f tc-tests/qdiscs/sfq.json -- ns/SubPlugin.__init__ Test 7482: Create SFQ with default setting Test c186: Create SFQ with limit setting Test ae23: Create SFQ with perturb setting Test a430: Create SFQ with quantum setting Test 4539: Create SFQ with divisor setting Test b089: Create SFQ with flows setting Test 99a0: Create SFQ with depth setting Test 7389: Create SFQ with headdrop setting Test 6472: Create SFQ with redflowlimit setting Test 8929: Show SFQ class Test 4d6f: Check that limit of 1 is rejected Test 7f8f: Check that a derived limit of 1 is rejected (limit 2 depth 1 flows 1) Test 5168: Check that a derived limit of 1 is rejected (limit 2 depth 1 divisor 1) All test results: 1..13 ok 1 7482 - Create SFQ with default setting ok 2 c186 - Create SFQ with limit setting ok 3 ae23 - Create SFQ with perturb setting ok 4 a430 - Create SFQ with quantum setting ok 5 4539 - Create SFQ with divisor setting ok 6 b089 - Create SFQ with flows setting ok 7 99a0 - Create SFQ with depth setting ok 8 7389 - Create SFQ with headdrop setting ok 9 6472 - Create SFQ with redflowlimit setting ok 10 8929 - Show SFQ class ok 11 4d6f - Check that limit of 1 is rejected ok 12 7f8f - Check that a derived limit of 1 is rejected (limit 2 depth 1 flows 1) ok 13 5168 - Check that a derived limit of 1 is rejected (limit 2 depth 1 divisor 1) Thanks, Harshit Eric Dumazet (2): net_sched: sch_sfq: annotate data-races around q->perturb_period net_sched: sch_sfq: handle bigger packets Octavian Purdila (3): net_sched: sch_sfq: don't allow 1 packet limit net_sched: sch_sfq: use a temporary work area for validating configuration net_sched: sch_sfq: move the limit validation net/sched/sch_sfq.c | 112 ++++++++++++++++++++++++++++---------------- 1 file changed, 71 insertions(+), 41 deletions(-) -- 2.47.1

3 weeks, 1 day

1
5
0 0

[PATCH 1/1] phy: tegra: xusb: Fix unbalanced regulator disable in UTMI PHY mode

by Wayne Chang

When transitioning from USB_ROLE_DEVICE to USB_ROLE_NONE, the code assumed that the regulator should be disabled. However, if the regulator is marked as always-on, regulator_is_enabled() continues to return true, leading to an incorrect attempt to disable a regulator which is not enabled. This can result in warnings such as: [ 250.155624] WARNING: CPU: 1 PID: 7326 at drivers/regulator/core.c:3004 _regulator_disable+0xe4/0x1a0 [ 250.155652] unbalanced disables for VIN_SYS_5V0 To fix this, we move the regulator control logic into tegra186_xusb_padctl_id_override() function since it's directly related to the ID override state. The regulator is now only disabled when the role transitions from USB_ROLE_HOST to USB_ROLE_NONE, by checking the VBUS_ID register. This ensures that regulator enable/disable operations are properly balanced and only occur when actually transitioning to/from host mode. Fixes: 49d46e3c7e59 ("phy: tegra: xusb: Add set_mode support for UTMI phy on Tegra186") Cc: stable(a)vger.kernel.org Signed-off-by: Wayne Chang <waynec(a)nvidia.com> --- drivers/phy/tegra/xusb-tegra186.c | 59 +++++++++++++++++++------------ 1 file changed, 37 insertions(+), 22 deletions(-) diff --git a/drivers/phy/tegra/xusb-tegra186.c b/drivers/phy/tegra/xusb-tegra186.c index fae6242aa730..1b35d50821f7 100644 --- a/drivers/phy/tegra/xusb-tegra186.c +++ b/drivers/phy/tegra/xusb-tegra186.c @@ -774,13 +774,15 @@ static int tegra186_xusb_padctl_vbus_override(struct tegra_xusb_padctl *padctl, } static int tegra186_xusb_padctl_id_override(struct tegra_xusb_padctl *padctl, - bool status) + struct tegra_xusb_usb2_port *port, bool status) { - u32 value; + u32 value, id_override; + int err = 0; dev_dbg(padctl->dev, "%s id override\n", status ? "set" : "clear"); value = padctl_readl(padctl, USB2_VBUS_ID); + id_override = value & ID_OVERRIDE(~0); if (status) { if (value & VBUS_OVERRIDE) { @@ -791,15 +793,35 @@ static int tegra186_xusb_padctl_id_override(struct tegra_xusb_padctl *padctl, value = padctl_readl(padctl, USB2_VBUS_ID); } - value &= ~ID_OVERRIDE(~0); - value |= ID_OVERRIDE_GROUNDED; + if (id_override != ID_OVERRIDE_GROUNDED) { + value &= ~ID_OVERRIDE(~0); + value |= ID_OVERRIDE_GROUNDED; + padctl_writel(padctl, value, USB2_VBUS_ID); + + err = regulator_enable(port->supply); + if (err) { + dev_err(padctl->dev, "Failed to enable regulator: %d\n", err); + return err; + } + } } else { - value &= ~ID_OVERRIDE(~0); - value |= ID_OVERRIDE_FLOATING; + if (id_override == ID_OVERRIDE_GROUNDED) { + /* + * The regulator is disabled only when the role transitions + * from USB_ROLE_HOST to USB_ROLE_NONE. + */ + err = regulator_disable(port->supply); + if (err) { + dev_err(padctl->dev, "Failed to disable regulator: %d\n", err); + return err; + } + + value &= ~ID_OVERRIDE(~0); + value |= ID_OVERRIDE_FLOATING; + padctl_writel(padctl, value, USB2_VBUS_ID); + } } - padctl_writel(padctl, value, USB2_VBUS_ID); - return 0; } @@ -818,27 +840,20 @@ static int tegra186_utmi_phy_set_mode(struct phy *phy, enum phy_mode mode, if (mode == PHY_MODE_USB_OTG) { if (submode == USB_ROLE_HOST) { - tegra186_xusb_padctl_id_override(padctl, true); - - err = regulator_enable(port->supply); + err = tegra186_xusb_padctl_id_override(padctl, port, true); + if (err) + goto out; } else if (submode == USB_ROLE_DEVICE) { tegra186_xusb_padctl_vbus_override(padctl, true); } else if (submode == USB_ROLE_NONE) { - /* - * When port is peripheral only or role transitions to - * USB_ROLE_NONE from USB_ROLE_DEVICE, regulator is not - * enabled. - */ - if (regulator_is_enabled(port->supply)) - regulator_disable(port->supply); - - tegra186_xusb_padctl_id_override(padctl, false); + err = tegra186_xusb_padctl_id_override(padctl, port, false); + if (err) + goto out; tegra186_xusb_padctl_vbus_override(padctl, false); } } - +out: mutex_unlock(&padctl->lock); - return err; } -- 2.25.1

3 weeks, 1 day

3
3
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror