tools/power turbostat: Fix added raw MSR output
Subsystems (13)
Sasha Levin
maintainerPATCH 1/19 tools/power turbostat: Fix added raw MSR output
From: Doug Smythies <[email protected]>
[ Upstream commit e5f4e68eed85fa8495d78cd966eecc2b27bb9e53 ]
When using --Summary mode, added MSRs in raw mode always print zeros. Print the actual register contents.
Example, with patch:
note the added column:
--add msr0x64f,u32,package,raw,REASON
Where:
0x64F is MSR_CORE_PERF_LIMIT_REASONS
Busy% Bzy_MHz PkgTmp PkgWatt CorWatt REASON 0.00 4800 35 1.42 0.76 0x00000000 0.00 4801 34 1.42 0.76 0x00000000 80.08 4531 66 108.17 107.52 0x08000000 98.69 4530 66 133.21 132.54 0x08000000 99.28 4505 66 128.26 127.60 0x0c000400 99.65 4486 68 124.91 124.25 0x0c000400 99.63 4483 68 124.90 124.25 0x0c000400 79.34 4481 41 99.80 99.13 0x0c000000 0.00 4801 41 1.40 0.73 0x0c000000
Where, for the test processor (i5-10600K):
PKG Limit #1: 125.000 Watts, 8.000000 sec
MSR bit 26 = log; bit 10 = status
PKG Limit #2: 136.000 Watts, 0.002441 sec
MSR bit 27 = log; bit 11 = status
Example, without patch:
Busy% Bzy_MHz PkgTmp PkgWatt CorWatt REASON 0.01 4800 35 1.43 0.77 0x00000000 0.00 4801 35 1.39 0.73 0x00000000 83.49 4531 66 112.71 112.06 0x00000000 98.69 4530 68 133.35 132.69 0x00000000 99.31 4500 67 127.96 127.30 0x00000000 99.63 4483 69 124.91 124.25 0x00000000 99.61 4481 69 124.90 124.25 0x00000000 99.61 4481 71 124.92 124.25 0x00000000 59.35 4479 42 75.03 74.37 0x00000000 0.00 4800 42 1.39 0.73 0x00000000 0.00 4801 42 1.42 0.76 0x00000000
c000000
[lenb: simplified patch to apply only to package scope]
--- tools/power/x86/turbostat/turbostat.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
tools/power/x86/turbostat/turbostat.c +4 -3
@@ -1811,9 +1811,10 @@ int sum_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) average.packages.rapl_dram_perf_status += p->rapl_dram_perf_status;for (i =0 , mp = sys.pp; mp; i++, mp = mp->next) { -if (mp->format ==FORMAT_RAW ) -continue ; - average.packages.counter[i] += p->counter[i]; +if ((mp->format ==FORMAT_RAW ) && (topo.num_packages ==0 )) + average.packages.counter[i] = p->counter[i]; +else + average.packages.counter[i] += p->counter[i]; }return 0 ; }
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 2/19 tools/power turbostat: Increase the limit for fd opened
From: Wyes Karny <[email protected]>
[ Upstream commit 3ac1d14d0583a2de75d49a5234d767e2590384dd ]
When running turbostat, a system with 512 cpus reaches the limit for maximum number of file descriptors that can be opened. To solve this problem, the limit is raised to 2^15, which is a large enough number.
Below data is collected from AMD server systems while running turbostat:
|-----------+-------------------------------| | # of cpus | # of opened fds for turbostat | |-----------+-------------------------------| | 128 | 260 | |-----------+-------------------------------| | 192 | 388 | |-----------+-------------------------------| | 512 | 1028 | |-----------+-------------------------------|
So, the new max limit would be sufficient up to 2^14 cpus (but this also depends on how many counters are enabled).
--- tools/power/x86/turbostat/turbostat.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+)
tools/power/x86/turbostat/turbostat.c +21 -0
@@ -53,6 +53,8 @@#define NAME_BYTES 20#define PATH_BYTES 128 +#define MAX_NOFILE 0x8000 +enum counter_scope {SCOPE_CPU ,SCOPE_CORE ,SCOPE_PACKAGE };enum counter_type {COUNTER_ITEMS ,COUNTER_CYCLES ,COUNTER_SECONDS ,COUNTER_USEC };enum counter_format {FORMAT_RAW ,FORMAT_DELTA ,FORMAT_PERCENT };
@@ -6719,6 +6721,22 @@ void cmdline(int argc, char **argv) } } +void set_rlimit (void ) +{ +struct rlimit limit; + +if (getrlimit (RLIMIT_NOFILE , &limit) <0 ) +err (1 ,"Failed to get rlimit" ); + +if (limit.rlim_max <MAX_NOFILE ) + limit.rlim_max =MAX_NOFILE ; +if (limit.rlim_cur <MAX_NOFILE ) + limit.rlim_cur =MAX_NOFILE ; + +if (setrlimit (RLIMIT_NOFILE , &limit) <0 ) +err (1 ,"Failed to set rlimit" ); +} +int main (int argc,char **argv) { outf = stderr;
@@ -6731,6 +6749,9 @@ int main(int argc, char **argv)probe_sysfs (); +if (!getuid ()) +set_rlimit (); +turbostat_init ();msr_sum_record ();
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 3/19 tools/power turbostat: Fix Bzy_MHz documentation typo
In reply to Sasha LevinFrom: Peng Liu <[email protected]>
[ Upstream commit 0b13410b52c4636aacb6964a4253a797c0fa0d16 ]
The code calculates Bzy_MHz by multiplying TSC_delta * APERF_delta/MPERF_delta The man page erroneously showed that TSC_delta was divided.
--- tools/power/x86/turbostat/turbostat.8 | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/power/x86/turbostat/turbostat.8 +1 -1
@@ -370,7 +370,7 @@ below the processor's base frequency. Busy% = MPERF_delta/TSC_delta -Bzy_MHz = TSC_delta/APERF_delta/MPERF_delta/measurement_interval +Bzy_MHz = TSC_delta*APERF_delta/MPERF_delta/measurement_interval Note that these calculations depend on TSC_delta, so they are not reliable during intervals when TSC_MHz is not running at the base frequency.
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 4/19 btrfs: make btrfs_clear_delalloc_extent() free delalloc reserve
In reply to Sasha LevinFrom: Boris Burkov <[email protected]>
[ Upstream commit 3c6f0c5ecc8910d4ffb0dfe85609ebc0c91c8f34 ]
Currently, this call site in btrfs_clear_delalloc_extent() only converts the reservation. We are marking it not delalloc, so I don't think it makes sense to keep the rsv around. This is a path where we are not sure to join a transaction, so it leads to incorrect free-ing during umount.
Helps with the pass rate of generic/269 and generic/475.
--- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/btrfs/inode.c +1 -1
@@ -2472,7 +2472,7 @@ void btrfs_clear_delalloc_extent(struct inode *vfs_inode,*/ if (bits &EXTENT_CLEAR_META_RESV && root != fs_info->tree_root) -btrfs_delalloc_release_metadata (inode, len, false); +btrfs_delalloc_release_metadata (inode, len, true);/* For sanity tests. */ if (btrfs_is_testing (fs_info))
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 5/19 btrfs: always clear PERTRANS metadata during commit
In reply to Sasha LevinFrom: Boris Burkov <[email protected]>
[ Upstream commit 6e68de0bb0ed59e0554a0c15ede7308c47351e2d ]
It is possible to clear a root's IN_TRANS tag from the radix tree, but not clear its PERTRANS, if there is some error in between. Eliminate that possibility by moving the free up to where we clear the tag.
--- fs/btrfs/transaction.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/btrfs/transaction.c +1 -1
@@ -1440,6 +1440,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans)radix_tree_tag_clear (&fs_info->fs_roots_radix, (unsigned long )root->root_key.objectid,BTRFS_ROOT_TRANS_TAG ); +btrfs_qgroup_free_meta_all_pertrans (root);spin_unlock (&fs_info->fs_roots_radix_lock);btrfs_free_log (trans, root);
@@ -1464,7 +1465,6 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans)if (ret2)return ret2;spin_lock (&fs_info->fs_roots_radix_lock); -btrfs_qgroup_free_meta_all_pertrans (root); } }spin_unlock (&fs_info->fs_roots_radix_lock);
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 6/19 memblock tests: fix undefined reference to `early_pfn_to_nid'
In reply to Sasha LevinFrom: Wei Yang <[email protected]>
[ Upstream commit 7d8ed162e6a92268d4b2b84d364a931216102c8e ]
commit 6a9531c3a880 ("memblock: fix crash when reserved memory is not added to memory") introduce the usage of early_pfn_to_nid, which is not defined in memblock tests.
The original definition of early_pfn_to_nid is defined in mm.h, so let add this in the corresponding mm.h.
--- tools/include/linux/mm.h | 5 +++++ 1 file changed, 5 insertions(+)
tools/include/linux/mm.h +5 -0
@@ -37,4 +37,9 @@ static inline void totalram_pages_add(long count) { } +static inline int early_pfn_to_nid (unsigned long pfn) +{ +return 0 ; +} +#endif
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 7/19 memblock tests: fix undefined reference to `panic'
In reply to Sasha LevinFrom: Wei Yang <[email protected]>
[ Upstream commit e0f5a8e74be88f2476e58b25d3b49a9521bdc4ec ]
commit e96c6b8f212a ("memblock: report failures when memblock_can_resize is not set") introduced the usage of panic, which is not defined in memblock test.
Let's define it directly in panic.h to fix it.
--- tools/include/linux/kernel.h | 1 + tools/include/linux/panic.h | 19 +++++++++++++++++++ 2 files changed, 20 insertions(+) create mode 100644 tools/include/linux/panic.h
tools/include/linux/kernel.h +1 -0
@@ -8,6 +8,7 @@#include <linux/build_bug.h> #include <linux/compiler.h> #include <linux/math.h> +#include <linux/panic.h> #include <endian.h> #include <byteswap.h>
tools/include/linux/panic.h new file +19 -0
@@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _TOOLS_LINUX_PANIC_H +#define _TOOLS_LINUX_PANIC_H + +#include <stdarg.h> +#include <stdio.h> +#include <stdlib.h> + +static inline void panic (const char *fmt, ...) +{ +va_list argp; + +va_start (argp, fmt); +vfprintf (stderr, fmt, argp); +va_end (argp); +exit (-1 ); +} + +#endif
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 8/19 memblock tests: fix undefined reference to `BIT'
In reply to Sasha LevinFrom: Wei Yang <[email protected]>
[ Upstream commit 592447f6cb3c20d606d6c5d8e6af68e99707b786 ]
commit 772dd0342727 ("mm: enumerate all gfp flags") define gfp flags with the help of BIT, while gfp_types.h doesn't include header file for the definition. This through an error on building memblock tests.
Let's include linux/bits.h to fix it.
--- include/linux/gfp_types.h | 2 ++ 1 file changed, 2 insertions(+)
include/linux/gfp_types.h +2 -0
@@ -2,6 +2,8 @@#ifndef __LINUX_GFP_TYPES_H#define __LINUX_GFP_TYPES_H +#include <linux/bits.h> +/* The typedef is in types.h but we want the documentation here */ #if 0 /**
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 9/19 scsi: target: Fix SELinux error when systemd-modules loads the target module
In reply to Sasha LevinFrom: Maurizio Lombardi <[email protected]>
[ Upstream commit 97a54ef596c3fd24ec2b227ba8aaf2cf5415e779 ]
If the systemd-modules service loads the target module, the credentials of that userspace process will be used to validate the access to the target db directory. SELinux will prevent it, reporting an error like the following:
kernel: audit: type=1400 audit(1676301082.205:4): avc: denied { read } for pid=1020 comm="systemd-modules" name="target" dev="dm-3" ino=4657583 scontext=system_u:system_r:systemd_modules_load_t:s0 tcontext=system_u:object_r:targetd_etc_rw_t:s0 tclass=dir permissive=0
Fix the error by using the kernel credentials to access the db directory
--- drivers/target/target_core_configfs.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
drivers/target/target_core_configfs.c +12 -0
@@ -3587,6 +3587,8 @@ static int __init target_core_init_configfs(void) {struct configfs_subsystem *subsys = &target_core_fabrics;struct t10_alua_lu_gp *lu_gp; +struct cred *kern_cred; +const struct cred *old_cred;int ret; pr_debug("TARGET_CORE[0]: Loading Generic Kernel Storage"
@@ -3663,11 +3665,21 @@ static int __init target_core_init_configfs(void)if (ret <0 ) goto out; +/* We use the kernel credentials to access the target directory */ + kern_cred =prepare_kernel_cred (&init_task); +if (!kern_cred) { + ret = -ENOMEM ; + goto out; + } + old_cred =override_creds (kern_cred);target_init_dbroot (); +revert_creds (old_cred); +put_cred (kern_cred);return 0 ; out: +target_xcopy_release_pt ();configfs_unregister_subsystem (subsys);core_dev_release_virtual_lun0 ();rd_module_exit ();
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 10/19 blk-iocost: avoid out of bounds shift
In reply to Sasha LevinFrom: Rik van Riel <[email protected]>
[ Upstream commit beaa51b36012fad5a4d3c18b88a617aea7a9b96d ]
UBSAN catches undefined behavior in blk-iocost, where sometimes iocg->delay is shifted right by a number that is too large, resulting in undefined behavior on some architectures.
[ 186.556576] ------------[ cut here ]------------ UBSAN: shift-out-of-bounds in block/blk-iocost.c:1366:23 shift exponent 64 is too large for 64-bit type 'u64' (aka 'unsigned long long') CPU: 16 PID: 0 Comm: swapper/16 Tainted: G S E N 6.9.0-0_fbk700_debug_rc2_kbuilder_0_gc85af715cac0 #1 Hardware name: Quanta Twin Lakes MP/Twin Lakes Passive MP, BIOS F09_3A23 12/08/2020 Call Trace: <IRQ> dump_stack_lvl+0x8f/0xe0 __ubsan_handle_shift_out_of_bounds+0x22c/0x280 iocg_kick_delay+0x30b/0x310 ioc_timer_fn+0x2fb/0x1f80 __run_timer_base+0x1b6/0x250
...
Avoid that undefined behavior by simply taking the "delay = 0" branch if the shift is too large.
I am not sure what the symptoms of an undefined value delay will be, but I suspect it could be more than a little annoying to debug.
--- block/blk-iocost.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
block/blk-iocost.c +4 -3
@@ -1331,7 +1331,7 @@ static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now) {struct ioc *ioc = iocg->ioc;struct blkcg_gq *blkg =iocg_to_blkg (iocg); -u64 tdelta, delay, new_delay; +u64 tdelta, delay, new_delay, shift;s64 vover, vover_pct;u32 hwa;
@@ -1346,8 +1346,9 @@ static bool iocg_kick_delay(struct ioc_gq *iocg, struct ioc_now *now)/* calculate the current delay in effect - 1/2 every second */ tdelta = now->now - iocg->delay_at; -if (iocg->delay) - delay = iocg->delay >>div64_u64 (tdelta,USEC_PER_SEC ); + shift =div64_u64 (tdelta,USEC_PER_SEC ); +if (iocg->delay && shift <BITS_PER_LONG ) + delay = iocg->delay >> shift;else delay =0 ;
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 11/19 gpu: host1x: Do not setup DMA for virtual devices
In reply to Sasha LevinFrom: Thierry Reding <[email protected]>
[ Upstream commit 8ab58f6841b19423231c5db3378691ec80c778f8 ]
The host1x devices are virtual compound devices and do not perform DMA accesses themselves, so they do not need to be set up for DMA.
Ideally we would also not need to set up DMA masks for the virtual devices, but we currently still need those for legacy support on old hardware.
--- drivers/gpu/host1x/bus.c | 8 -------- 1 file changed, 8 deletions(-)
drivers/gpu/host1x/bus.c +0 -8
@@ -368,11 +368,6 @@ static int host1x_device_uevent(struct device *dev,return 0 ; } -static int host1x_dma_configure (struct device *dev) -{ -return of_dma_configure (dev, dev->of_node, true); -} -static const struct dev_pm_ops host1x_device_pm_ops = { .suspend = pm_generic_suspend, .resume = pm_generic_resume,
@@ -386,7 +381,6 @@ struct bus_type host1x_bus_type = { .name = "host1x", .match = host1x_device_match, .uevent = host1x_device_uevent, - .dma_configure = host1x_dma_configure, .pm = &host1x_device_pm_ops, };
@@ -475,8 +469,6 @@ static int host1x_device_add(struct host1x *host1x, device->dev.bus = &host1x_bus_type; device->dev.parent = host1x->dev; -of_dma_configure (&device->dev, host1x->dev->of_node, true); - device->dev.dma_parms = &device->dma_parms;dma_set_max_seg_size (&device->dev,UINT_MAX );
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 12/19 MIPS: scall: Save thread_info.syscall unconditionally on entry
In reply to Sasha LevinFrom: Jiaxun Yang <[email protected]>
[ Upstream commit 4370b673ccf240bf7587b0cb8e6726a5ccaf1f17 ]
thread_info.syscall is used by syscall_get_nr to supply syscall nr over a thread stack frame.
Previously, thread_info.syscall is only saved at syscall_trace_enter when syscall tracing is enabled. However rest of the kernel code do expect syscall_get_nr to be available without syscall tracing. The previous design breaks collect_syscall.
Move saving process to syscall entry to fix it.
--- arch/mips/include/asm/ptrace.h | 2 +- arch/mips/kernel/asm-offsets.c | 1 + arch/mips/kernel/ptrace.c | 15 ++++++--------- arch/mips/kernel/scall32-o32.S | 23 +++++++++++++---------- arch/mips/kernel/scall64-n32.S | 3 ++- arch/mips/kernel/scall64-n64.S | 3 ++- arch/mips/kernel/scall64-o32.S | 33 +++++++++++++++++---------------- 7 files changed, 42 insertions(+), 38 deletions(-)
arch/mips/include/asm/ptrace.h +1 -1
@@ -157,7 +157,7 @@ static inline long regs_return_value(struct pt_regs *regs)#define instruction_pointer (regs) ((regs)->cp0_epc)#define profile_pc (regs) instruction_pointer(regs) -extern asmlinkage long syscall_trace_enter (struct pt_regs *regs,long syscall); +extern asmlinkage long syscall_trace_enter (struct pt_regs *regs); extern asmlinkage void syscall_trace_leave(struct pt_regs *regs);extern void die (const char *,struct pt_regs *)__noreturn ;
arch/mips/kernel/asm-offsets.c +1 -0
@@ -98,6 +98,7 @@ void output_thread_info_defines(void)OFFSET (TI_CPU , thread_info, cpu);OFFSET (TI_PRE_COUNT , thread_info, preempt_count);OFFSET (TI_REGS , thread_info, regs); +OFFSET (TI_SYSCALL , thread_info, syscall);DEFINE (_THREAD_SIZE,THREAD_SIZE );DEFINE (_THREAD_MASK,THREAD_MASK );DEFINE (_IRQ_STACK_SIZE,IRQ_STACK_SIZE );
arch/mips/kernel/ptrace.c +6 -9
@@ -1309,16 +1309,13 @@ long arch_ptrace(struct task_struct *child, long request,* Notification of system call entry/exit * - triggered by current->work.syscall_trace */ -asmlinkage long syscall_trace_enter (struct pt_regs *regs,long syscall) +asmlinkage long syscall_trace_enter (struct pt_regs *regs) {user_exit (); -current_thread_info ()->syscall = syscall; -if (test_thread_flag (TIF_SYSCALL_TRACE )) {if (ptrace_report_syscall_entry (regs))return -1 ; - syscall =current_thread_info ()->syscall; }#ifdef CONFIG_SECCOMP
@@ -1327,7 +1324,7 @@ asmlinkage long syscall_trace_enter(struct pt_regs *regs, long syscall)struct seccomp_data sd;unsigned long args[6 ]; - sd.nr = syscall; + sd.nr =current_thread_info ()->syscall; sd.arch =syscall_get_arch (current);syscall_get_arguments (current, regs, args);for (i =0 ; i <6 ; i++)
@@ -1337,23 +1334,23 @@ asmlinkage long syscall_trace_enter(struct pt_regs *regs, long syscall) ret =__secure_computing (&sd);if (ret ==-1 )return ret; - syscall = current_thread_info()->syscall; }#endif if (unlikely (test_thread_flag (TIF_SYSCALL_TRACEPOINT )))trace_sys_enter (regs, regs->regs[2 ]); -audit_syscall_entry (syscall, regs->regs[4 ], regs->regs[5 ], +audit_syscall_entry (current_thread_info ()->syscall, + regs->regs[4 ], regs->regs[5 ], regs->regs[6 ], regs->regs[7 ]);/* * Negative syscall numbers are mistaken for rejected syscalls, but * won't have had the return value set appropriately, so we do so now. */ -if (syscall <0 ) +if (current_thread_info ()->syscall <0 )syscall_set_return_value (current, regs, -ENOSYS ,0 ); -return syscall; +return current_thread_info ()->syscall; }/*
arch/mips/kernel/scall32-o32.S +13 -10
@@ -77,6 +77,18 @@ loads_done: PTR_WD load_a7, bad_stack_a7 .previous + /* + * syscall number is in v0 unless we called syscall(__NR_###) + * where the real syscall number is in a0 + */ + subu t2, v0, __NR_O32_Linux + bnez t2, 1f /* __NR_syscall at offset 0 */ + LONG_S a0, TI_SYSCALL($28) # Save a0 as syscall number + b 2f +1: + LONG_S v0, TI_SYSCALL($28) # Save v0 as syscall number +2: + lw t0, TI_FLAGS($28) # syscall tracing enabled? li t1, _TIF_WORK_SYSCALL_ENTRY and t0, t1
@@ -114,16 +126,7 @@ syscall_trace_entry: SAVE_STATIC move a0, sp - /* - * syscall number is in v0 unless we called syscall(__NR_###) - * where the real syscall number is in a0 - */ - move a1, v0 - subu t2, v0, __NR_O32_Linux - bnez t2, 1f /* __NR_syscall at offset 0 */ - lw a1, PT_R4(sp) - -1: jal syscall_trace_enter + jal syscall_trace_enter bltz v0, 1f # seccomp failed? Skip syscall
arch/mips/kernel/scall64-n32.S +2 -1
@@ -44,6 +44,8 @@ NESTED(handle_sysn32, PT_SIZE, sp) sd a3, PT_R26(sp) # save a3 for syscall restarting + LONG_S v0, TI_SYSCALL($28) # Store syscall number + li t1, _TIF_WORK_SYSCALL_ENTRY LONG_L t0, TI_FLAGS($28) # syscall tracing enabled? and t0, t1, t0
@@ -72,7 +74,6 @@ syscall_common: n32_syscall_trace_entry: SAVE_STATIC move a0, sp - move a1, v0 jal syscall_trace_enter bltz v0, 1f # seccomp failed? Skip syscall
arch/mips/kernel/scall64-n64.S +2 -1
@@ -46,6 +46,8 @@ NESTED(handle_sys64, PT_SIZE, sp) sd a3, PT_R26(sp) # save a3 for syscall restarting + LONG_S v0, TI_SYSCALL($28) # Store syscall number + li t1, _TIF_WORK_SYSCALL_ENTRY LONG_L t0, TI_FLAGS($28) # syscall tracing enabled? and t0, t1, t0
@@ -82,7 +84,6 @@ n64_syscall_exit: syscall_trace_entry: SAVE_STATIC move a0, sp - move a1, v0 jal syscall_trace_enter bltz v0, 1f # seccomp failed? Skip syscall
arch/mips/kernel/scall64-o32.S +17 -16
@@ -79,6 +79,22 @@ loads_done: PTR_WD load_a7, bad_stack_a7 .previous + /* + * absolute syscall number is in v0 unless we called syscall(__NR_###) + * where the real syscall number is in a0 + * note: NR_syscall is the first O32 syscall but the macro is + * only defined when compiling with -mabi=32 (CONFIG_32BIT) + * therefore __NR_O32_Linux is used (4000) + */ + + subu t2, v0, __NR_O32_Linux + bnez t2, 1f /* __NR_syscall at offset 0 */ + LONG_S a0, TI_SYSCALL($28) # Save a0 as syscall number + b 2f +1: + LONG_S v0, TI_SYSCALL($28) # Save v0 as syscall number +2: + li t1, _TIF_WORK_SYSCALL_ENTRY LONG_L t0, TI_FLAGS($28) # syscall tracing enabled? and t0, t1, t0
@@ -113,22 +129,7 @@ trace_a_syscall: sd a7, PT_R11(sp) # For indirect syscalls move a0, sp - /* - * absolute syscall number is in v0 unless we called syscall(__NR_###) - * where the real syscall number is in a0 - * note: NR_syscall is the first O32 syscall but the macro is - * only defined when compiling with -mabi=32 (CONFIG_32BIT) - * therefore __NR_O32_Linux is used (4000) - */ - .set push - .set reorder - subu t1, v0, __NR_O32_Linux - move a1, v0 - bnez t1, 1f /* __NR_syscall at offset 0 */ - ld a1, PT_R4(sp) /* Arg1 for __NR_syscall case */ - .set pop - -1: jal syscall_trace_enter + jal syscall_trace_enter bltz v0, 1f # seccomp failed? Skip syscall
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 13/19 tools/power/turbostat: Fix uncore frequency file string
In reply to Sasha LevinFrom: Justin Ernst <[email protected]>
[ Upstream commit 60add818ab2543b7e4f2bfeaacf2504743c1eb50 ]
Running turbostat on a 16 socket HPE Scale-up Compute 3200 (SapphireRapids) fails with:
turbostat: /sys/devices/system/cpu/intel_uncore_frequency/package_010_die_00/current_freq_khz: open failed: No such file or directory
We observe the sysfs uncore frequency directories named:
...
package_09_die_00/
package_10_die_00/
package_11_die_00/
...
package_15_die_00/
The culprit is an incorrect sprintf format string "package_0%d_die_0%d" used with each instance of reading uncore frequency files. uncore-frequency-common.c creates the sysfs directory with the format "package_%02d_die_%02d". Once the package value reaches double digits, the formats diverge.
Change each instance of "package_0%d_die_0%d" to "package_%02d_die_%02d".
[lenb: deleted the probe part of this patch, as it was already fixed]
--- tools/power/x86/turbostat/turbostat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
tools/power/x86/turbostat/turbostat.c +1 -1
@@ -1969,7 +1969,7 @@ unsigned long long get_uncore_mhz(int package, int die) {char path[128 ]; -sprintf (path,"/sys/devices/system/cpu/intel_uncore_frequency/package_0%d_die_0%d/current_freq_khz" , package, +sprintf (path,"/sys/devices/system/cpu/intel_uncore_frequency/package_%02d_die_%02d/current_freq_khz" , package, die);return (snapshot_sysfs_counter (path) /1000 );
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 14/19 drm/amdgpu: Refine IB schedule error logging
In reply to Sasha LevinFrom: Lijo Lazar <[email protected]>
[ Upstream commit 4b18a91faf1752f9bd69a4ed3aed2c8f6e5b0528 ]
Downgrade to debug information when IBs are skipped. Also, use dev_* to identify the device.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_job.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +5 -2
@@ -301,12 +301,15 @@ static struct dma_fence *amdgpu_job_run(struct drm_sched_job *sched_job)dma_fence_set_error (finished, -ECANCELED );if (finished->error <0 ) { -DRM_INFO ("Skip scheduling IBs!\n" ); +dev_dbg (adev->dev,"Skip scheduling IBs in ring(%s)" , + ring->name); }else { r =amdgpu_ib_schedule (ring, job->num_ibs, job->ibs, job, &fence);if (r) -DRM_ERROR ("Error scheduling IBs (%d)\n" , r); +dev_err (adev->dev, +"Error scheduling IBs (%d) in ring(%s)" , r, + ring->name); } job->job_run_counter++;
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 15/19 selftests: timers: Fix valid-adjtimex signed left-shift undefined behavior
In reply to Sasha LevinFrom: John Stultz <[email protected]>
[ Upstream commit 076361362122a6d8a4c45f172ced5576b2d4a50d ]
The struct adjtimex freq field takes a signed value who's units are in shifted (<<16) parts-per-million.
Unfortunately for negative adjustments, the straightforward use of:
freq = ppm << 16 trips undefined behavior warnings with clang:
valid-adjtimex.c:66:6: warning: shifting a negative signed value is undefined [-Wshift-negative-value]
-499<<16,
~~~~^
valid-adjtimex.c:67:6: warning: shifting a negative signed value is undefined [-Wshift-negative-value]
-450<<16,
~~~~^..
Fix it by using a multiply by (1 << 16) instead of shifting negative values in the valid-adjtimex test case. Align the values for better readability.
--- .../testing/selftests/timers/valid-adjtimex.c | 73 +++++++++---------- 1 file changed, 36 insertions(+), 37 deletions(-)
tools/testing/selftests/timers/valid-adjtimex.c +36 -37
@@ -21,9 +21,6 @@* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. */ - - -#include <stdio.h> #include <stdlib.h> #include <time.h>
@@ -62,45 +59,47 @@ int clear_time_state(void)#define NUM_FREQ_OUTOFRANGE 4#define NUM_FREQ_INVALID 2 +#define SHIFTED_PPM (1 << 16) +long valid_freq[NUM_FREQ_VALID ] = { --499 <<16 , --450 <<16 , --400 <<16 , --350 <<16 , --300 <<16 , --250 <<16 , --200 <<16 , --150 <<16 , --100 <<16 , --75 <<16 , --50 <<16 , --25 <<16 , --10 <<16 , --5 <<16 , --1 <<16 , +-499 *SHIFTED_PPM , +-450 *SHIFTED_PPM , +-400 *SHIFTED_PPM , +-350 *SHIFTED_PPM , +-300 *SHIFTED_PPM , +-250 *SHIFTED_PPM , +-200 *SHIFTED_PPM , +-150 *SHIFTED_PPM , +-100 *SHIFTED_PPM , +-75 *SHIFTED_PPM , +-50 *SHIFTED_PPM , +-25 *SHIFTED_PPM , +-10 *SHIFTED_PPM , +-5 *SHIFTED_PPM , +-1 *SHIFTED_PPM ,-1000 , -1 <<16 , -5 <<16 , -10 <<16 , -25 <<16 , -50 <<16 , -75 <<16 , -100 <<16 , -150 <<16 , -200 <<16 , -250 <<16 , -300 <<16 , -350 <<16 , -400 <<16 , -450 <<16 , -499 <<16 , +1 *SHIFTED_PPM , +5 *SHIFTED_PPM , +10 *SHIFTED_PPM , +25 *SHIFTED_PPM , +50 *SHIFTED_PPM , +75 *SHIFTED_PPM , +100 *SHIFTED_PPM , +150 *SHIFTED_PPM , +200 *SHIFTED_PPM , +250 *SHIFTED_PPM , +300 *SHIFTED_PPM , +350 *SHIFTED_PPM , +400 *SHIFTED_PPM , +450 *SHIFTED_PPM , +499 *SHIFTED_PPM , };long outofrange_freq[NUM_FREQ_OUTOFRANGE ] = { --1000 <<16 , --550 <<16 , -550 <<16 , -1000 <<16 , +-1000 *SHIFTED_PPM , +-550 *SHIFTED_PPM , +550 *SHIFTED_PPM , +1000 *SHIFTED_PPM , };#define LONG_MAX (~0UL>>1)
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 16/19 Drivers: hv: vmbus: Track decrypted status in vmbus_gpadl
In reply to Sasha LevinFrom: Rick Edgecombe <[email protected]>
[ Upstream commit 211f514ebf1ef5de37b1cf6df9d28a56cfd242ca ]
In CoCo VMs it is possible for the untrusted host to cause set_memory_encrypted() or set_memory_decrypted() to fail such that an error is returned and the resulting memory is shared. Callers need to take care to handle these errors to avoid returning decrypted (shared) memory to the page allocator, which could lead to functional or security issues.
In order to make sure callers of vmbus_establish_gpadl() and vmbus_teardown_gpadl() don't return decrypted/shared pages to allocators, add a field in struct vmbus_gpadl to keep track of the decryption status of the buffers. This will allow the callers to know if they should free or leak the pages.
Message-ID: <[email protected]>
--- drivers/hv/channel.c | 25 +++++++++++++++++++++---- include/linux/hyperv.h | 1 + 2 files changed, 22 insertions(+), 4 deletions(-)
drivers/hv/channel.c +21 -4
@@ -472,9 +472,18 @@ static int __vmbus_establish_gpadl(struct vmbus_channel *channel, (atomic_inc_return (&vmbus_connection.next_gpadl_handle) -1 ); ret =create_gpadl_header (type, kbuffer, size, send_offset, &msginfo); -if (ret) +if (ret) { + gpadl->decrypted = false;return ret; + } +/* +* Set the "decrypted" flag to true for the set_memory_decrypted() +* success case. In the failure case, the encryption state of the +* memory is unknown. Leave "decrypted" as true to ensure the +* memory will be leaked instead of going back on the free list. +*/ + gpadl->decrypted = true; ret =set_memory_decrypted ((unsigned long )kbuffer,PFN_UP (size));if (ret) {
@@ -563,9 +572,15 @@ static int __vmbus_establish_gpadl(struct vmbus_channel *channel,kfree (msginfo); -if (ret) -set_memory_encrypted ((unsigned long )kbuffer, -PFN_UP (size)); +if (ret) { +/* +* If set_memory_encrypted() fails, the decrypted flag is +* left as true so the memory is leaked instead of being +* put back on the free list. +*/ +if (!set_memory_encrypted ((unsigned long )kbuffer,PFN_UP (size))) + gpadl->decrypted = false; + }return ret; }
@@ -886,6 +901,8 @@ int vmbus_teardown_gpadl(struct vmbus_channel *channel, struct vmbus_gpadl *gpadif (ret)pr_warn ("Fail to set mem host visibility in GPADL teardown %d.\n" , ret); + gpadl->decrypted = ret; +return ret; }EXPORT_SYMBOL_GPL (vmbus_teardown_gpadl);
include/linux/hyperv.h +1 -0
@@ -832,6 +832,7 @@ struct vmbus_gpadl {u32 gpadl_handle;u32 size;void *buffer; +bool decrypted; };struct vmbus_channel {
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 17/19 uio_hv_generic: Don't free decrypted memory
In reply to Sasha LevinFrom: Rick Edgecombe <[email protected]>
[ Upstream commit 3d788b2fbe6a1a1a9e3db09742b90809d51638b7 ]
In CoCo VMs it is possible for the untrusted host to cause set_memory_encrypted() or set_memory_decrypted() to fail such that an error is returned and the resulting memory is shared. Callers need to take care to handle these errors to avoid returning decrypted (shared) memory to the page allocator, which could lead to functional or security issues.
The VMBus device UIO driver could free decrypted/shared pages if set_memory_decrypted() fails. Check the decrypted field in the gpadl to decide whether to free the memory.
Message-ID: <[email protected]>
--- drivers/uio/uio_hv_generic.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-)
drivers/uio/uio_hv_generic.c +8 -4
@@ -181,12 +181,14 @@ hv_uio_cleanup(struct hv_device *dev, struct hv_uio_private_data *pdata) {if (pdata->send_gpadl.gpadl_handle) {vmbus_teardown_gpadl (dev->channel, &pdata->send_gpadl); -vfree (pdata->send_buf); +if (!pdata->send_gpadl.decrypted) +vfree (pdata->send_buf); }if (pdata->recv_gpadl.gpadl_handle) {vmbus_teardown_gpadl (dev->channel, &pdata->recv_gpadl); -vfree (pdata->recv_buf); +if (!pdata->recv_gpadl.decrypted) +vfree (pdata->recv_buf); } }
@@ -295,7 +297,8 @@ hv_uio_probe(struct hv_device *dev, ret =vmbus_establish_gpadl (channel, pdata->recv_buf,RECV_BUFFER_SIZE , &pdata->recv_gpadl);if (ret) { -vfree (pdata->recv_buf); +if (!pdata->recv_gpadl.decrypted) +vfree (pdata->recv_buf); goto fail_close; }
@@ -317,7 +320,8 @@ hv_uio_probe(struct hv_device *dev, ret =vmbus_establish_gpadl (channel, pdata->send_buf,SEND_BUFFER_SIZE , &pdata->send_gpadl);if (ret) { -vfree (pdata->send_buf); +if (!pdata->send_gpadl.decrypted) +vfree (pdata->send_buf); goto fail_close; }
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 18/19 Drivers: hv: vmbus: Don't free ring buffers that couldn't be re-encrypted
In reply to Sasha LevinFrom: Michael Kelley <[email protected]>
[ Upstream commit 30d18df6567be09c1433e81993e35e3da573ac48 ]
In CoCo VMs it is possible for the untrusted host to cause set_memory_encrypted() or set_memory_decrypted() to fail such that an error is returned and the resulting memory is shared. Callers need to take care to handle these errors to avoid returning decrypted (shared) memory to the page allocator, which could lead to functional or security issues.
The VMBus ring buffer code could free decrypted/shared pages if set_memory_decrypted() fails. Check the decrypted field in the struct vmbus_gpadl for the ring buffers to decide whether to free the memory.
Message-ID: <[email protected]>
--- drivers/hv/channel.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/hv/channel.c +3 -1
@@ -153,7 +153,9 @@ void vmbus_free_ring(struct vmbus_channel *channel)hv_ringbuffer_cleanup (&channel->inbound);if (channel->ringbuffer_page) { -__free_pages (channel->ringbuffer_page, +/* In a CoCo VM leak the memory if it didn't get re-encrypted */ +if (!channel->ringbuffer_gpadlhandle.decrypted) +__free_pages (channel->ringbuffer_page,get_order (channel->ringbuffer_pagecount <<PAGE_SHIFT )); channel->ringbuffer_page =NULL ;
Signature · 1 line
2.43.0
Sasha Levin
maintainerPATCH 19/19 iommu: mtk: fix module autoloading
In reply to Sasha LevinFrom: Krzysztof Kozlowski <[email protected]>
[ Upstream commit 7537e31df80cb58c27f3b6fef702534ea87a5957 ]
Add MODULE_DEVICE_TABLE(), so modules could be properly autoloaded based on the alias from of_device_id table.
--- drivers/iommu/mtk_iommu.c | 1 + drivers/iommu/mtk_iommu_v1.c | 1 + 2 files changed, 2 insertions(+)
drivers/iommu/mtk_iommu.c +1 -0
@@ -1570,6 +1570,7 @@ static const struct of_device_id mtk_iommu_of_ids[] = { { .compatible = "mediatek,mt8195-iommu-vpp", .data = &mt8195_data_vpp}, {} }; +MODULE_DEVICE_TABLE (of, mtk_iommu_of_ids);static struct platform_driver mtk_iommu_driver = { .probe = mtk_iommu_probe,
drivers/iommu/mtk_iommu_v1.c +1 -0
@@ -602,6 +602,7 @@ static const struct of_device_id mtk_iommu_v1_of_ids[] = { { .compatible = "mediatek,mt2701-m4u", }, {} }; +MODULE_DEVICE_TABLE (of, mtk_iommu_v1_of_ids);static const struct component_master_ops mtk_iommu_v1_com_ops = { .bind = mtk_iommu_v1_bind,
Signature · 1 line
2.43.0