July 2024 - Linux-stable-mirror

[PATCH v6 4/8] RISC-V: Scalar unaligned access emulated on hotplug CPUs

by Jesse Taube

The check_unaligned_access_emulated() function should have been called during CPU hotplug to ensure that if all CPUs had emulated unaligned accesses, the new CPU also does. This patch adds the call to check_unaligned_access_emulated() in the hotplug path. Fixes: 55e0bf49a0d0 ("RISC-V: Probe misaligned access speed in parallel") Signed-off-by: Jesse Taube <jesse(a)rivosinc.com> Cc: stable(a)vger.kernel.org --- V5 -> V6: - New patch --- arch/riscv/kernel/unaligned_access_speed.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/riscv/kernel/unaligned_access_speed.c b/arch/riscv/kernel/unaligned_access_speed.c index a9a6bcb02acf..b67db1fc3740 100644 --- a/arch/riscv/kernel/unaligned_access_speed.c +++ b/arch/riscv/kernel/unaligned_access_speed.c @@ -191,6 +191,7 @@ static int riscv_online_cpu(unsigned int cpu) if (per_cpu(misaligned_access_speed, cpu) != RISCV_HWPROBE_MISALIGNED_UNKNOWN) goto exit; + check_unaligned_access_emulated(NULL); buf = alloc_pages(GFP_KERNEL, MISALIGNED_BUFFER_ORDER); if (!buf) { pr_warn("Allocation failure, not measuring misaligned performance\n"); -- 2.45.2

11 months, 1 week

2
1
0 0

[PATCH v6 3/8] RISC-V: Check scalar unaligned access on all CPUs

by Jesse Taube

Originally, the check_unaligned_access_emulated_all_cpus function only checked the boot hart. This fixes the function to check all harts. Fixes: 71c54b3d169d ("riscv: report misaligned accesses emulation to hwprobe") Signed-off-by: Jesse Taube <jesse(a)rivosinc.com> Reviewed-by: Charlie Jenkins <charlie(a)rivosinc.com> Cc: stable(a)vger.kernel.org --- V1 -> V2: - New patch V2 -> V3: - Split patch V3 -> V4: - Re-add check for a system where a heterogeneous CPU is hotplugged into a previously homogenous system. V4 -> V5: - Change work_struct *unused to work_struct *work __always_unused V5 -> V6: - Change check_unaligned_access_emulated to extern --- arch/riscv/include/asm/cpufeature.h | 2 ++ arch/riscv/kernel/traps_misaligned.c | 14 +++++++------- 2 files changed, 9 insertions(+), 7 deletions(-) diff --git a/arch/riscv/include/asm/cpufeature.h b/arch/riscv/include/asm/cpufeature.h index 347805446151..3b24342c7d2a 100644 --- a/arch/riscv/include/asm/cpufeature.h +++ b/arch/riscv/include/asm/cpufeature.h @@ -8,6 +8,7 @@ #include <linux/bitmap.h> #include <linux/jump_label.h> +#include <linux/workqueue.h> #include <asm/hwcap.h> #include <asm/alternative-macros.h> #include <asm/errno.h> @@ -35,6 +36,7 @@ void riscv_user_isa_enable(void); #if defined(CONFIG_RISCV_MISALIGNED) bool check_unaligned_access_emulated_all_cpus(void); +void check_unaligned_access_emulated(struct work_struct *work __always_unused); void unaligned_emulation_finish(void); bool unaligned_ctl_available(void); DECLARE_PER_CPU(long, misaligned_access_speed); diff --git a/arch/riscv/kernel/traps_misaligned.c b/arch/riscv/kernel/traps_misaligned.c index b62d5a2f4541..bb09357778c5 100644 --- a/arch/riscv/kernel/traps_misaligned.c +++ b/arch/riscv/kernel/traps_misaligned.c @@ -526,11 +526,11 @@ int handle_misaligned_store(struct pt_regs *regs) return 0; } -static bool check_unaligned_access_emulated(int cpu) +void check_unaligned_access_emulated(struct work_struct *work __always_unused) { + int cpu = smp_processor_id(); long *mas_ptr = per_cpu_ptr(&misaligned_access_speed, cpu); unsigned long tmp_var, tmp_val; - bool misaligned_emu_detected; *mas_ptr = RISCV_HWPROBE_MISALIGNED_UNKNOWN; @@ -538,19 +538,16 @@ static bool check_unaligned_access_emulated(int cpu) " "REG_L" %[tmp], 1(%[ptr])\n" : [tmp] "=r" (tmp_val) : [ptr] "r" (&tmp_var) : "memory"); - misaligned_emu_detected = (*mas_ptr == RISCV_HWPROBE_MISALIGNED_EMULATED); /* * If unaligned_ctl is already set, this means that we detected that all * CPUS uses emulated misaligned access at boot time. If that changed * when hotplugging the new cpu, this is something we don't handle. */ - if (unlikely(unaligned_ctl && !misaligned_emu_detected)) { + if (unlikely(unaligned_ctl && (*mas_ptr != RISCV_HWPROBE_MISALIGNED_EMULATED))) { pr_crit("CPU misaligned accesses non homogeneous (expected all emulated)\n"); while (true) cpu_relax(); } - - return misaligned_emu_detected; } bool check_unaligned_access_emulated_all_cpus(void) @@ -562,8 +559,11 @@ bool check_unaligned_access_emulated_all_cpus(void) * accesses emulated since tasks requesting such control can run on any * CPU. */ + schedule_on_each_cpu(check_unaligned_access_emulated); + for_each_online_cpu(cpu) - if (!check_unaligned_access_emulated(cpu)) + if (per_cpu(misaligned_access_speed, cpu) + != RISCV_HWPROBE_MISALIGNED_EMULATED) return false; unaligned_ctl = true; -- 2.45.2

11 months, 1 week

2
1
0 0

[PATCH wireless] wifi: mac80211: use monitor sdata with driver only if desired

by Johannes Berg

From: Johannes Berg <johannes.berg(a)intel.com> In commit 0d9c2beed116 ("wifi: mac80211: fix monitor channel with chanctx emulation") I changed mac80211 to always have an internal monitor_sdata to have something to have the chanctx bound to. However, if the driver didn't also have the WANT_MONITOR flag this would cause mac80211 to allocate it without telling the driver (which was intentional) but also use it for later APIs to the driver without it ever having known about it which was _not_ intentional. Check through the code and only use the monitor_sdata in the relevant places (TX, MU-MIMO follow settings, TX power, and interface iteration) when the WANT_MONITOR flag is set. Cc: stable(a)vger.kernel.org Fixes: 0d9c2beed116 ("wifi: mac80211: fix monitor channel with chanctx emulation") Reported-by: ZeroBeat <ZeroBeat(a)gmx.de> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219086 Tested-by: Lorenzo Bianconi <lorenzo(a)kernel.org> Signed-off-by: Johannes Berg <johannes.berg(a)intel.com> --- net/mac80211/cfg.c | 7 +++++-- net/mac80211/tx.c | 5 +++-- net/mac80211/util.c | 2 +- 3 files changed, 9 insertions(+), 5 deletions(-) diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c index 85cb71de370f..b02b84ce2130 100644 --- a/net/mac80211/cfg.c +++ b/net/mac80211/cfg.c @@ -114,7 +114,7 @@ static int ieee80211_set_mon_options(struct ieee80211_sub_if_data *sdata, /* apply all changes now - no failures allowed */ - if (monitor_sdata) + if (monitor_sdata && ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF)) ieee80211_set_mu_mimo_follow(monitor_sdata, params); if (params->flags) { @@ -3053,6 +3053,9 @@ static int ieee80211_set_tx_power(struct wiphy *wiphy, sdata = IEEE80211_WDEV_TO_SUB_IF(wdev); if (sdata->vif.type == NL80211_IFTYPE_MONITOR) { + if (!ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF)) + return -EOPNOTSUPP; + sdata = wiphy_dereference(local->hw.wiphy, local->monitor_sdata); if (!sdata) @@ -3115,7 +3118,7 @@ static int ieee80211_set_tx_power(struct wiphy *wiphy, if (has_monitor) { sdata = wiphy_dereference(local->hw.wiphy, local->monitor_sdata); - if (sdata) { + if (sdata && ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF)) { sdata->deflink.user_power_level = local->user_power_level; if (txp_type != sdata->vif.bss_conf.txpower_type) update_txp_type = true; diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c index 72a9ba8bc5fd..edba4a31844f 100644 --- a/net/mac80211/tx.c +++ b/net/mac80211/tx.c @@ -1768,7 +1768,7 @@ static bool __ieee80211_tx(struct ieee80211_local *local, break; } sdata = rcu_dereference(local->monitor_sdata); - if (sdata) { + if (sdata && ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF)) { vif = &sdata->vif; info->hw_queue = vif->hw_queue[skb_get_queue_mapping(skb)]; @@ -3957,7 +3957,8 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw, break; } tx.sdata = rcu_dereference(local->monitor_sdata); - if (tx.sdata) { + if (tx.sdata && + ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF)) { vif = &tx.sdata->vif; info->hw_queue = vif->hw_queue[skb_get_queue_mapping(skb)]; diff --git a/net/mac80211/util.c b/net/mac80211/util.c index ced19ce7c51a..c7ad9bc5973a 100644 --- a/net/mac80211/util.c +++ b/net/mac80211/util.c @@ -776,7 +776,7 @@ static void __iterate_interfaces(struct ieee80211_local *local, sdata = rcu_dereference_check(local->monitor_sdata, lockdep_is_held(&local->iflist_mtx) || lockdep_is_held(&local->hw.wiphy->mtx)); - if (sdata && + if (sdata && ieee80211_hw_check(&local->hw, WANT_MONITOR_VIF) && (iter_flags & IEEE80211_IFACE_ITER_RESUME_ALL || !active_only || sdata->flags & IEEE80211_SDATA_IN_DRIVER)) iterator(data, sdata->vif.addr, &sdata->vif); -- 2.45.2

11 months, 1 week

2
1
0 0

[PATCH] tools/nolibc: include arch.h from string.h

by Thomas Weißschuh

string.h tests for the macros NOLIBC_ARCH_HAS_$FUNC to use the architecture-optimized function variants. However if string.h is included before arch.h header than that check does not work, leading to duplicate function definitions. Fixes: 553845eebd60 ("tools/nolibc: x86-64: Use `rep movsb` for `memcpy()` and `memmove()`") Fixes: 12108aa8c1a1 ("tools/nolibc: x86-64: Use `rep stosb` for `memset()`") Cc: stable(a)vger.kernel.org Signed-off-by: Thomas Weißschuh <linux(a)weissschuh.net> --- If nobody complains I'll apply this after v6.11-rc1 is released. --- tools/include/nolibc/string.h | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/include/nolibc/string.h b/tools/include/nolibc/string.h index f9ab28421e6d..9ec9c24f38c0 100644 --- a/tools/include/nolibc/string.h +++ b/tools/include/nolibc/string.h @@ -7,6 +7,7 @@ #ifndef _NOLIBC_STRING_H #define _NOLIBC_STRING_H +#include "arch.h" #include "std.h" static void *malloc(size_t len); --- base-commit: 6ca8f2e20bd1ced8a7cd12b3ae4b1ceca85cfc2b change-id: 20240725-arch-has-func-59b6c92ce935 Best regards, -- Thomas Weißschuh <linux(a)weissschuh.net>

11 months, 1 week

2
1
0 0

[PATCH 2/3] spmi: pmic-arb: Pass the correct of_node to irq_domain_add_tree

by Stephen Boyd

From: Konrad Dybcio <konrad.dybcio(a)linaro.org> Currently, irqchips for all of the subnodes (which represent a given bus master) point to the parent wrapper node. This is no bueno, as no interrupts arrive, ever (because nothing references that node). Fix that by passing a reference to the respective master's of_node. Worth noting, this is a NOP for devices with only a single master described. Signed-off-by: Konrad Dybcio <konrad.dybcio(a)linaro.org> Link: https://lore.kernel.org/r/20240522-topic-spmi_multi_master_irqfix-v2-1-7ec9… Reviewed-by: Abel Vesa <abel.vesa(a)linaro.org> Tested-by: Dmitry Baryshkov <dmitry.baryshkov(a)linaro.org> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov(a)linaro.org> Fixes: 02922ccbb330 ("spmi: pmic-arb: Register controller for bus instead of arbiter") Cc: stable(a)vger.kernel.org Signed-off-by: Stephen Boyd <sboyd(a)kernel.org> --- drivers/spmi/spmi-pmic-arb.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c index f240fcc5a4e1..b6880c13163c 100644 --- a/drivers/spmi/spmi-pmic-arb.c +++ b/drivers/spmi/spmi-pmic-arb.c @@ -1737,8 +1737,7 @@ static int spmi_pmic_arb_bus_init(struct platform_device *pdev, dev_dbg(&pdev->dev, "adding irq domain for bus %d\n", bus_index); - bus->domain = irq_domain_add_tree(dev->of_node, - &pmic_arb_irq_domain_ops, bus); + bus->domain = irq_domain_add_tree(node, &pmic_arb_irq_domain_ops, bus); if (!bus->domain) { dev_err(&pdev->dev, "unable to create irq_domain\n"); return -ENOMEM; -- https://git.kernel.org/pub/scm/linux/kernel/git/clk/linux.git/ https://git.kernel.org/pub/scm/linux/kernel/git/sboyd/spmi.git

11 months, 1 week

1
0
0 0

[PATCH 6.1.y] udf: Convert udf_mkdir() to new directory iteration code

by Sergio González Collado

From: Jan Kara <jack(a)suse.cz> [ Upstream commit 00bce6f792caccefa73daeaf9bde82d24d50037f ] Convert udf_mkdir() to new directory iteration code. Signed-off-by: Jan Kara <jack(a)suse.cz> (cherry picked from commit 00bce6f792caccefa73daeaf9bde82d24d50037f) Signed-off-by: Sergio González Collado <sergio.collado(a)gmail.com> Reported-by: syzbot+600a32676df180ebc4af(a)syzkaller.appspotmail.com --- fs/udf/namei.c | 48 +++++++++++++++++++++--------------------------- 1 file changed, 21 insertions(+), 27 deletions(-) diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 7c95c549dd64..a33e6d762716 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -665,8 +665,7 @@ static int udf_mkdir(struct user_namespace *mnt_userns, struct inode *dir, struct dentry *dentry, umode_t mode) { struct inode *inode; - struct udf_fileident_bh fibh; - struct fileIdentDesc cfi, *fi; + struct udf_fileident_iter iter; int err; struct udf_inode_info *dinfo = UDF_I(dir); struct udf_inode_info *iinfo; @@ -678,47 +677,42 @@ static int udf_mkdir(struct user_namespace *mnt_userns, struct inode *dir, iinfo = UDF_I(inode); inode->i_op = &udf_dir_inode_operations; inode->i_fop = &udf_dir_operations; - fi = udf_add_entry(inode, NULL, &fibh, &cfi, &err); - if (!fi) { - inode_dec_link_count(inode); + err = udf_fiiter_add_entry(inode, NULL, &iter); + if (err) { + clear_nlink(inode); discard_new_inode(inode); - goto out; + return err; } set_nlink(inode, 2); - cfi.icb.extLength = cpu_to_le32(inode->i_sb->s_blocksize); - cfi.icb.extLocation = cpu_to_lelb(dinfo->i_location); - *(__le32 *)((struct allocDescImpUse *)cfi.icb.impUse)->impUse = + iter.fi.icb.extLength = cpu_to_le32(inode->i_sb->s_blocksize); + iter.fi.icb.extLocation = cpu_to_lelb(dinfo->i_location); + *(__le32 *)((struct allocDescImpUse *)iter.fi.icb.impUse)->impUse = cpu_to_le32(dinfo->i_unique & 0x00000000FFFFFFFFUL); - cfi.fileCharacteristics = + iter.fi.fileCharacteristics = FID_FILE_CHAR_DIRECTORY | FID_FILE_CHAR_PARENT; - udf_write_fi(inode, &cfi, fi, &fibh, NULL, NULL); - brelse(fibh.sbh); + udf_fiiter_write_fi(&iter, NULL); + udf_fiiter_release(&iter); mark_inode_dirty(inode); - fi = udf_add_entry(dir, dentry, &fibh, &cfi, &err); - if (!fi) { + err = udf_fiiter_add_entry(dir, dentry, &iter); + if (err) { clear_nlink(inode); - mark_inode_dirty(inode); discard_new_inode(inode); - goto out; + return err; } - cfi.icb.extLength = cpu_to_le32(inode->i_sb->s_blocksize); - cfi.icb.extLocation = cpu_to_lelb(iinfo->i_location); - *(__le32 *)((struct allocDescImpUse *)cfi.icb.impUse)->impUse = + iter.fi.icb.extLength = cpu_to_le32(inode->i_sb->s_blocksize); + iter.fi.icb.extLocation = cpu_to_lelb(iinfo->i_location); + *(__le32 *)((struct allocDescImpUse *)iter.fi.icb.impUse)->impUse = cpu_to_le32(iinfo->i_unique & 0x00000000FFFFFFFFUL); - cfi.fileCharacteristics |= FID_FILE_CHAR_DIRECTORY; - udf_write_fi(dir, &cfi, fi, &fibh, NULL, NULL); + iter.fi.fileCharacteristics |= FID_FILE_CHAR_DIRECTORY; + udf_fiiter_write_fi(&iter, NULL); + udf_fiiter_release(&iter); inc_nlink(dir); dir->i_ctime = dir->i_mtime = current_time(dir); mark_inode_dirty(dir); d_instantiate_new(dentry, inode); - if (fibh.sbh != fibh.ebh) - brelse(fibh.ebh); - brelse(fibh.sbh); - err = 0; -out: - return err; + return 0; } static int empty_dir(struct inode *dir) -- 2.39.2

11 months, 1 week

2
3
0 0

[REGRESSION] ALSA: firewire-lib: heavy digital distortion with Fireface 800

by edmund.raile

Bisection revealed that the bitcrushing distortion with RME FireFace 800 was caused by 1d717123bb1a7555 ("ALSA: firewire-lib: Avoid -Wflex-array-member-not-at-end warning"). Reverting this commit yields restoration of clear audio output. I will send in a patch reverting this commit for now, soonTM. #regzbot introduced: 1d717123bb1a7555

11 months, 1 week

4
6
0 0

FAILED: patch "[PATCH] s390/mm: Fix VM_FAULT_HWPOISON handling in do_exception()" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.6-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y git checkout FETCH_HEAD git cherry-pick -x df39038cd89525d465c2c8827eb64116873f141a # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2024072535-synergy-struggle-8ecc@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^.. Possible dependencies: df39038cd895 ("s390/mm: Fix VM_FAULT_HWPOISON handling in do_exception()") b20c8216c1e0 ("s390/mm,fault: move VM_FAULT_ERROR handling to do_exception()") 7c194d84a9ce ("s390/mm,fault: remove VM_FAULT_BADMAP and VM_FAULT_BADACCESS") b61a0922b6dc ("s390/mm,fault: remove VM_FAULT_SIGNAL") 0f86ac4ba713 ("s390/mm,fault: remove VM_FAULT_BADCONTEXT") 0f1a14e0348e ("s390/mm,fault: simplify kfence fault handling") 64ea33fb09f8 ("s390/mm,fault: call do_fault_error() only from do_exception()") 5db06565cad6 ("s390/mm,fault: get rid of do_low_address()") cca12b427d43 ("s390/mm,fault: remove VM_FAULT_PFAULT") 5be05c35e72f ("s390/mm,fault: improve readability by using teid union") 4416d2ed8166 ("s390/mm,fault: use static key for store indication") 9641613f48bb ("s390/mm,fault: use get_fault_address() everywhere") 5c845de331d9 ("s390/mm,fault: remove noinline attribute from all functions") 8dbc33dc8163 ("s390/mm,fault: have balanced braces, remove unnecessary blanks") 7c915a84e5e2 ("s390/mm,fault: reverse x-mas tree coding style") 3aad8c044297 ("s390/mm,fault: remove and improve comments, adjust whitespace") thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From df39038cd89525d465c2c8827eb64116873f141a Mon Sep 17 00:00:00 2001 From: Gerald Schaefer <gerald.schaefer(a)linux.ibm.com> Date: Mon, 15 Jul 2024 20:04:16 +0200 Subject: [PATCH] s390/mm: Fix VM_FAULT_HWPOISON handling in do_exception() There is no support for HWPOISON, MEMORY_FAILURE, or ARCH_HAS_COPY_MC on s390. Therefore we do not expect to see VM_FAULT_HWPOISON in do_exception(). However, since commit af19487f00f3 ("mm: make PTE_MARKER_SWAPIN_ERROR more general"), it is possible to see VM_FAULT_HWPOISON in combination with PTE_MARKER_POISONED, even on architectures that do not support HWPOISON otherwise. In this case, we will end up on the BUG() in do_exception(). Fix this by treating VM_FAULT_HWPOISON the same as VM_FAULT_SIGBUS, similar to x86 when MEMORY_FAILURE is not configured. Also print unexpected fault flags, for easier debugging. Note that VM_FAULT_HWPOISON_LARGE is not expected, because s390 cannot support swap entries on other levels than PTE level. Cc: stable(a)vger.kernel.org # 6.6+ Fixes: af19487f00f3 ("mm: make PTE_MARKER_SWAPIN_ERROR more general") Reported-by: Yunseong Kim <yskelg(a)gmail.com> Tested-by: Yunseong Kim <yskelg(a)gmail.com> Acked-by: Alexander Gordeev <agordeev(a)linux.ibm.com> Signed-off-by: Gerald Schaefer <gerald.schaefer(a)linux.ibm.com> Message-ID: <20240715180416.3632453-1-gerald.schaefer(a)linux.ibm.com> Signed-off-by: Vasily Gorbik <gor(a)linux.ibm.com> diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c index 6b19a33c49c2..8e149ef5e89b 100644 --- a/arch/s390/mm/fault.c +++ b/arch/s390/mm/fault.c @@ -433,12 +433,13 @@ static void do_exception(struct pt_regs *regs, int access) handle_fault_error_nolock(regs, 0); else do_sigsegv(regs, SEGV_MAPERR); - } else if (fault & VM_FAULT_SIGBUS) { + } else if (fault & (VM_FAULT_SIGBUS | VM_FAULT_HWPOISON)) { if (!user_mode(regs)) handle_fault_error_nolock(regs, 0); else do_sigbus(regs); } else { + pr_emerg("Unexpected fault flags: %08x\n", fault); BUG(); } }

11 months, 1 week

2
1
0 0

[PATCH net v2] gve: Fix an edge case for TSO skb validity check

by Praveen Kaligineedi

From: Bailey Forrest <bcf(a)google.com> The NIC requires each TSO segment to not span more than 10 descriptors. NIC further requires each descriptor to not exceed 16KB - 1 (GVE_TX_MAX_BUF_SIZE_DQO). The descriptors for an skb are generated by gve_tx_add_skb_no_copy_dqo() for DQO RDA queue format. gve_tx_add_skb_no_copy_dqo() loops through each skb frag and generates a descriptor for the entire frag if the frag size is not greater than GVE_TX_MAX_BUF_SIZE_DQO. If the frag size is greater than GVE_TX_MAX_BUF_SIZE_DQO, it is split into descriptor(s) of size GVE_TX_MAX_BUF_SIZE_DQO and a descriptor is generated for the remainder (frag size % GVE_TX_MAX_BUF_SIZE_DQO). gve_can_send_tso() checks if the descriptors thus generated for an skb would meet the requirement that each TSO-segment not span more than 10 descriptors. However, the current code misses an edge case when a TSO segment spans multiple descriptors within a large frag. This change fixes the edge case. gve_can_send_tso() relies on the assumption that max gso size (9728) is less than GVE_TX_MAX_BUF_SIZE_DQO and therefore within an skb fragment a TSO segment can never span more than 2 descriptors. Fixes: a57e5de476be ("gve: DQO: Add TX path") Signed-off-by: Praveen Kaligineedi <pkaligineedi(a)google.com> Signed-off-by: Bailey Forrest <bcf(a)google.com> Reviewed-by: Jeroen de Borst <jeroendb(a)google.com> Cc: stable(a)vger.kernel.org --- Changes from v1: - Added 'stable tag' - Added more explanation in the commit message - Modified comments to clarify the changes made - Changed variable names 'last_frag_size' to 'prev_frag_size' and 'last_frag_remain' to 'prev_frag_remain' - Removed parentheses around single line statement drivers/net/ethernet/google/gve/gve_tx_dqo.c | 22 +++++++++++++++++++- 1 file changed, 21 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/google/gve/gve_tx_dqo.c b/drivers/net/ethernet/google/gve/gve_tx_dqo.c index 0b3cca3fc792..f879426cb552 100644 --- a/drivers/net/ethernet/google/gve/gve_tx_dqo.c +++ b/drivers/net/ethernet/google/gve/gve_tx_dqo.c @@ -866,22 +866,42 @@ static bool gve_can_send_tso(const struct sk_buff *skb) const int header_len = skb_tcp_all_headers(skb); const int gso_size = shinfo->gso_size; int cur_seg_num_bufs; + int prev_frag_size; int cur_seg_size; int i; cur_seg_size = skb_headlen(skb) - header_len; + prev_frag_size = skb_headlen(skb); cur_seg_num_bufs = cur_seg_size > 0; for (i = 0; i < shinfo->nr_frags; i++) { if (cur_seg_size >= gso_size) { cur_seg_size %= gso_size; cur_seg_num_bufs = cur_seg_size > 0; + + if (prev_frag_size > GVE_TX_MAX_BUF_SIZE_DQO) { + int prev_frag_remain = prev_frag_size % + GVE_TX_MAX_BUF_SIZE_DQO; + + /* If the last descriptor of the previous frag + * is less than cur_seg_size, the segment will + * span two descriptors in the previous frag. + * Since max gso size (9728) is less than + * GVE_TX_MAX_BUF_SIZE_DQO, it is impossible + * for the segment to span more than two + * descriptors. + */ + if (prev_frag_remain && + cur_seg_size > prev_frag_remain) + cur_seg_num_bufs++; + } } if (unlikely(++cur_seg_num_bufs > max_bufs_per_seg)) return false; - cur_seg_size += skb_frag_size(&shinfo->frags[i]); + prev_frag_size = skb_frag_size(&shinfo->frags[i]); + cur_seg_size += prev_frag_size; } return true; -- 2.45.2.1089.g2a221341d9-goog

11 months, 1 week

3
2
0 0

[PATCH 6.1.y] bpf: Synchronize dispatcher update with bpf_dispatcher_xdp_func

by Sergio González Collado

From: Jiri Olsa <jolsa(a)kernel.org> [ Upstream commit 4121d4481b72501aa4d22680be4ea1096d69d133 ] Hao Sun reported crash in dispatcher image [1]. Currently we don't have any sync between bpf_dispatcher_update and bpf_dispatcher_xdp_func, so following race is possible: cpu 0: cpu 1: bpf_prog_run_xdp ... bpf_dispatcher_xdp_func in image at offset 0x0 bpf_dispatcher_update update image at offset 0x800 bpf_dispatcher_update update image at offset 0x0 in image at offset 0x0 -> crash Fixing this by synchronizing dispatcher image update (which is done in bpf_dispatcher_update function) with bpf_dispatcher_xdp_func that reads and execute the dispatcher image. Calling synchronize_rcu after updating and installing new image ensures that readers leave old image before it's changed in the next dispatcher update. The update itself is locked with dispatcher's mutex. The bpf_prog_run_xdp is called under local_bh_disable and synchronize_rcu will wait for it to leave [2]. [1] https://lore.kernel.org/bpf/Y5SFho7ZYXr9ifRn@krava/T/#m00c29ece654bc9f332a1… [2] https://lore.kernel.org/bpf/0B62D35A-E695-4B7A-A0D4-774767544C1A@gmail.com/… Reported-by: Hao Sun <sunhao.th(a)gmail.com> Signed-off-by: Jiri Olsa <jolsa(a)kernel.org> Acked-by: Yonghong Song <yhs(a)fb.com> Acked-by: Paul E. McKenney <paulmck(a)kernel.org> Link: https://lore.kernel.org/r/20221214123542.1389719-1-jolsa@kernel.org Signed-off-by: Martin KaFai Lau <martin.lau(a)kernel.org> (cherry picked from commit 4121d4481b72501aa4d22680be4ea1096d69d133) Signed-off-by: Sergio González Collado <sergio.collado(a)gmail.com> Reported-by: syzbot+08ba1e474d350b613604(a)syzkaller.appspotmail.com --- kernel/bpf/dispatcher.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kernel/bpf/dispatcher.c b/kernel/bpf/dispatcher.c index c19719f48ce0..fa3e9225aedc 100644 --- a/kernel/bpf/dispatcher.c +++ b/kernel/bpf/dispatcher.c @@ -125,6 +125,11 @@ static void bpf_dispatcher_update(struct bpf_dispatcher *d, int prev_num_progs) __BPF_DISPATCHER_UPDATE(d, new ?: (void *)&bpf_dispatcher_nop_func); + /* Make sure all the callers executing the previous/old half of the + * image leave it, so following update call can modify it safely. + */ + synchronize_rcu(); + if (new) d->image_off = noff; } -- 2.39.2

11 months, 1 week

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror July 2024