parisc: Prevent interrupts during reboot
Subsystems (6)
Sasha Levin
maintainerPATCH 1/1 parisc: Prevent interrupts during reboot
From: Helge Deller <[email protected]>
[ Upstream commit 35ac5a728c878594f2ea6c43b57652a16be3c968 ]
---
LLM Generated explanations, may be completely bogus:
Now I have all the information I need for a thorough analysis. Here is my complete assessment:
---
## Detailed Analysis: "parisc: Prevent interrupts during reboot"
### 1. COMMIT MESSAGE ANALYSIS
The commit message is minimal: "parisc: Prevent interrupts during reboot". It clearly states the problem it's fixing -- interrupts are not properly disabled during the reboot path on PA-RISC systems. The author is **Helge Deller**, the long-time parisc maintainer and the most authoritative developer for this architecture. The commit has no Fixes: tag or Cc: stable (as expected for autosel candidates).
### 2. CODE CHANGE ANALYSIS
The change is **one single effective line of code** plus a comment:
```c /* prevent interrupts during reboot */ set_eiem(0); ```
This is inserted into machine_restart() in arch/parisc/kernel/process.c immediately after pdc_chassis_send_status(PDC_CHASSIS_DIRECT_SHUTDOWN) and before pdc_do_reset().
**What set_eiem(0) does:** On PA-RISC, the EIEM (External Interrupt Enable Mask, Control Register 15) controls which external interrupts can fire. Setting it to 0 **masks all external interrupts at the hardware level**, preventing any interrupt from being delivered to the CPU. This is defined as:
```82:82:arch/parisc/include/asm/special_insns.h #define set_eiem(val) mtctl(val, CR_EIEM) ```
**The bug:** Without this line, external interrupts remain enabled during the entire reboot sequence. This means:
a) **Deadlock risk in pdc_do_reset()**: The pdc_do_reset() function acquires pdc_lock via spin_lock_irqsave():
```1236:1246:arch/parisc/kernel/firmware.c
int pdc_do_reset(void)
{
int retval;
unsigned long flags;
spin_lock_irqsave(&pdc_lock, flags);
retval = mem_pdc_call(PDC_BROADCAST_RESET, PDC_DO_RESET);
spin_unlock_irqrestore(&pdc_lock, flags);
return retval;
}
```While spin_lock_irqsave disables local interrupts, the PA-RISC EIEM hardware mask is a separate mechanism. On PA-RISC, the external interrupt delivery path goes through the EIEM -- an interrupt fires only if the corresponding EIEM bit is set AND the EIRR (External Interrupt Request Register) bit is set. If a hardware interrupt fires between pdc_chassis_send_status() (which also uses pdc_lock) and pdc_do_reset(), or during the firmware calls themselves, it could interfere with the reset process.
b) **Interference with firmware reset**: pdc_do_reset() calls into PDC firmware (mem_pdc_call(PDC_BROADCAST_RESET, PDC_DO_RESET)). Firmware calls on PA-RISC are sensitive to the processor state. An interrupt arriving during or between firmware calls can corrupt the reset sequence, potentially causing the machine to **hang instead of rebooting**.
c) **The gsc_writel(CMD_RESET, COMMAND_GLOBAL) fallback**: If pdc_do_reset() returns (on machines that don't implement PDC_BROADCAST_RESET), the code tries a hardware reset via gsc_writel. Interrupts during this path are equally problematic.
### 3. ESTABLISHED PATTERN IN PARISC AND OTHER ARCHITECTURES
**PA-RISC internal precedent:**
- parisc_terminate() in traps.c uses the exact same pattern:
set_eiem(0) followed by local_irq_disable() before critical shutdown operations (line 428-429)
- The SMP CPU hotplug code (smp.c:481) uses set_eiem(0) to disable all external interrupts when taking a CPU offline
**Other architectures ALL disable interrupts before reset:**
- ARM: local_irq_disable() at line 136 of arch/arm/kernel/reboot.c
- ARM64: local_irq_disable() at line 141 of arch/arm64/kernel/process.c
- x86: local_irq_disable() at line 100 of arch/x86/kernel/reboot.c
- xtensa: local_irq_disable() at line 524 of arch/xtensa/kernel/setup.c
- nios2: local_irq_disable() at line 49 of arch/nios2/kernel/process.c
- csky: local_irq_disable() at line 25 of arch/csky/kernel/power.c
- MIPS falcon: local_irq_disable() at line 37 of arch/mips/lantiq/falcon/reset.c
PA-RISC was the **outlier** in not disabling interrupts before reboot. This commit fixes that deficiency.
### 4. CLASSIFICATION
This is a **bug fix** -- specifically fixing a potential hang/crash during reboot caused by unmasked interrupts during firmware calls. It is NOT a new feature, optimization, or cleanup.
### 5. SCOPE AND RISK ASSESSMENT
- **Lines changed**: 3 (1 blank line, 1 comment, 1 functional line)
- **Files touched**: 1 (arch/parisc/kernel/process.c)
- **Subsystem**: parisc architecture, reboot path
- **Risk**: **Extremely low**. Disabling interrupts before a system reset is universally accepted as correct and necessary. The set_eiem(0) pattern is already used in the same codebase for similar critical paths.
### 6. USER IMPACT
- **Who is affected**: All PA-RISC Linux users who reboot their machines
- **What happens without the fix**: Potential hang during reboot, requiring a hard power cycle. This could be intermittent and timing- dependent, making it hard to reproduce but very real.
- **Severity if triggered**: System hang requiring manual intervention (power cycle)
### 7. STABILITY INDICATORS
- Author is the subsystem maintainer (highest trust level)
- Pattern is well-established across all other architectures
- The exact same set_eiem(0) call is used in the parisc parisc_terminate() function already
- The change is trivially small and obviously correct
### 8. DEPENDENCY CHECK
- No dependencies on other commits
- The set_eiem() macro has existed since the earliest parisc code
- The machine_restart() function is largely unchanged across all stable trees
- This will apply cleanly to any kernel version that has parisc support
### Summary
This is a trivially small, obviously correct fix that prevents interrupts during the reboot sequence on PA-RISC systems. Without it, an untimely interrupt can interfere with PDC firmware calls during reset, potentially causing the machine to hang instead of rebooting. Every other Linux architecture disables interrupts before reboot; parisc was the outlier. The fix uses the same set_eiem(0) pattern already used in parisc_terminate() and SMP CPU offline code. The risk is essentially zero -- there is no conceivable way that disabling interrupts right before a hardware reset could cause a regression.
**YES**
arch/parisc/kernel/process.c | 3 +++ 1 file changed, 3 insertions(+)
arch/parisc/kernel/process.c +3 -0
@@ -85,6 +85,9 @@ void machine_restart(char *cmd)#endif /* set up a new led state on systems shipped with a LED State panel */ pdc_chassis_send_status (PDC_CHASSIS_DIRECT_SHUTDOWN ); + +/* prevent interrupts during reboot */ +set_eiem (0 );/* "Normal" system reset */ pdc_do_reset ();
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 soc: imx8m: Fix error handling for clk_prepare_enable()
From: Peng Fan <[email protected]>
[ Upstream commit f6ef3d9ff81240e9bcc030f2da132eb0f8a761d7 ]
imx8m_soc_prepare() directly returns the result of clk_prepare_enable(), which skips proper cleanup if the clock enable fails. Check the return value of clk_prepare_enable() and release resources if failure.
---
LLM Generated explanations, may be completely bogus:
Interesting - in v6.15 and earlier, clk_prepare_enable() was called without checking its return value at all, but the code continued with proper cleanup on the success path. The v6.15 code has a different (arguably less severe) issue - it doesn't check clk_prepare_enable() return value but doesn't leak resources on normal exit.
Now I have a complete picture. Let me write the analysis.
---
## Detailed Analysis
### 1. Commit Message Analysis
The commit title explicitly says "Fix error handling for clk_prepare_enable()". The message clearly describes the bug:imx8m_soc_prepare() directly returns the result of clk_prepare_enable(), bypassing the err_clk cleanup label when the clock enable fails. This is a resource leak bug.
**Reporters:** Kernel test robot (lkp) and Dan Carpenter (well-known Smatch/static analysis developer). This means the bug was found via static analysis (Smatch), not a runtime crash report. The presence of "Closes:" link to a kernel test robot report confirms this.
**Reviews:** Two Reviewed-by tags (Marco Felsch and Daniel Baluta), indicating the fix was validated by subsystem developers.
### 2. Code Change Analysis
The bug is in imx8m_soc_prepare():
```131:156:drivers/soc/imx/soc-imx8m.c
static int imx8m_soc_prepare(struct platform_device *pdev, const char
*ocotp_compatible)
{
// ... setup code ...
drvdata->ocotp_base = of_iomap(np, 0); // resource 1: iomap
// ...
drvdata->clk = of_clk_get_by_name(np, NULL); // resource 2:
clock ref
// ...
return clk_prepare_enable(drvdata->clk); // BUG: if this
fails, ocotp_base is leaked!
err_clk:
iounmap(drvdata->ocotp_base);
return ret;
}
```**The bug mechanism:** When clk_prepare_enable() fails, the function returns directly with the error code, skipping the err_clk label which calls iounmap(drvdata->ocotp_base). This leaks:
1. The ocotp_base iomap mapping (definite leak)
2. The clock reference from of_clk_get_by_name() (also leaked, since clk_put() is not called - this is arguably still not fully fixed, but the err_clk label was designed for the of_clk_get_by_name failure case where clk_put should not be called)
**The fix:** Check the return value, jump to err_clk on failure, which properly unmaps ocotp_base:
```diff - return clk_prepare_enable(drvdata->clk); + ret = clk_prepare_enable(drvdata->clk); + if (ret) + goto err_clk; + return 0; ```
**Note:** There is a subtlety - when clk_prepare_enable() fails, the err_clk label only calls iounmap() but not clk_put(). The clock reference obtained from of_clk_get_by_name() is still leaked. However, this is a separate (pre-existing) concern. The fix as written is still an improvement: it's strictly better than the old code.
### 3. Dependency/Prerequisite Analysis
This is **critical** for the backport decision:
- The function imx8m_soc_prepare() was introduced in commit 390c01073f5d0 ("soc: imx8m: Cleanup with adding imx8m_soc_[un]prepare"), which first appeared in **v6.16-rc1**.
- This means the **buggy code does NOT exist** in v6.15 or any earlier stable tree (v6.14, v6.13, v6.12, v6.6, v6.1, 5.15, etc.).
- The bug was introduced in v6.16 and the fix only applies to **v6.16.y** stable and newer (v6.17.y, v6.18.y if applicable).
- I confirmed the buggy code exists identically in v6.16.12 (latest v6.16 stable), so the fix would apply cleanly.
### 4. Scope and Risk Assessment
- **Size:** 5 lines changed (+5/-1) in a single file - extremely small and surgical.
- **Risk:** Essentially zero. The fix only changes the error path behavior (adds proper cleanup). The success path (when clk_prepare_enable() works) is unchanged (returns 0 instead of the return value of clk_prepare_enable(), but those are equivalent on success).
- **Subsystem:** SoC driver for NXP i.MX8M family, a widely-used embedded platform.
### 5. User Impact
- **Affected hardware:** All i.MX8M family SoCs (i.MX8MQ, i.MX8MM, i.MX8MN, i.MX8MP) - very common in embedded/industrial/automotive use.
- **Trigger likelihood:** Low in practice. clk_prepare_enable() for the OCOTP clock is unlikely to fail on real hardware under normal conditions.
- **Severity if triggered:** Resource leak (iomap + clock reference). Not a crash, not security-critical, not data corruption. On repeated failures (e.g., in a test environment or if the clock framework has issues), could contribute to resource exhaustion.
- **Detection method:** Static analysis (Smatch/kernel test robot), not a user report.
### 6. Classification
This is a **real bug fix** - it fixes a resource leak on an error path. It falls into the category of "error path changes that fix resource leaks," which is explicitly identified as backport-worthy material despite being described as a "cleanup."
### 7. Stable Criteria Check
| Criterion | Assessment | |---|---| | Obviously correct and tested | Yes - trivially correct, reviewed by 2 developers | | Fixes a real bug | Yes - resource leak on error path | | Important issue | Moderate - resource leak, not crash/corruption | | Small and contained | Yes - 5 lines in 1 file | | No new features/APIs | Correct | | Applies cleanly | Yes - to v6.16.y |
### Conclusion
This is a small, obviously correct fix for a resource leak in the i.MX8M SoC driver's error handling path. The fix is surgical (5 lines), well- reviewed, and applies cleanly to v6.16.y stable (the only stable tree that has the affected code). While the bug is unlikely to trigger in practice (it requires clk_prepare_enable() to fail), the fix has essentially zero regression risk and properly handles an error path that should have been handled from the start. The fix was reported by kernel test robot and Dan Carpenter, two well-known automated and manual static analysis sources, and was reviewed by two subsystem developers.
The fix is appropriate for v6.16.y stable backport. It does NOT apply to older stable trees (v6.15.y and below) because the affected function was introduced in v6.16.
**YES**
drivers/soc/imx/soc-imx8m.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/soc/imx/soc-imx8m.c +5 -1
@@ -148,7 +148,11 @@ static int imx8m_soc_prepare(struct platform_device *pdev, const char *ocotp_com goto err_clk; } -return clk_prepare_enable (drvdata->clk); + ret =clk_prepare_enable (drvdata->clk); +if (ret) + goto err_clk; + +return 0 ; err_clk:iounmap (drvdata->ocotp_base);
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 soc/tegra: pmc: Fix unsafe generic_handle_irq() call
In reply to Sasha LevinFrom: Prathamesh Shete <[email protected]>
[ Upstream commit e6d96073af681780820c94079b978474a8a44413 ]
Currently, when resuming from system suspend on Tegra platforms, the following warning is observed:
WARNING: CPU: 0 PID: 14459 at kernel/irq/irqdesc.c:666 Call trace: handle_irq_desc+0x20/0x58 (P) tegra186_pmc_wake_syscore_resume+0xe4/0x15c syscore_resume+0x3c/0xb8 suspend_devices_and_enter+0x510/0x540 pm_suspend+0x16c/0x1d8
The warning occurs because generic_handle_irq() is being called from a non-interrupt context which is considered as unsafe.
Fix this warning by deferring generic_handle_irq() call to an IRQ work which gets executed in hard IRQ context where generic_handle_irq() can be called safely.
When PREEMPT_RT kernels are used, regular IRQ work (initialized with init_irq_work) is deferred to run in per-CPU kthreads in preemptible context rather than hard IRQ context. Hence, use the IRQ_WORK_INIT_HARD variant so that with PREEMPT_RT kernels, the IRQ work is processed in hardirq context instead of being deferred to a thread which is required for calling generic_handle_irq().
On non-PREEMPT_RT kernels, both init_irq_work() and IRQ_WORK_INIT_HARD() execute in IRQ context, so this change has no functional impact for standard kernel configurations.
---
LLM Generated explanations, may be completely bogus:
## Complete Analysis
### 1. COMMIT MESSAGE ANALYSIS
**Subject**: "soc/tegra: pmc: Fix unsafe generic_handle_irq() call"
The commit message is unambiguous - it's fixing a **real, observable WARNING** during system resume on Tegra platforms. The warning backtrace is included:
``` WARNING: CPU: 0 PID: 14459 at kernel/irq/irqdesc.c:666 Call trace: handle_irq_desc+0x20/0x58 (P) tegra186_pmc_wake_syscore_resume+0xe4/0x15c syscore_resume+0x3c/0xb8 suspend_devices_and_enter+0x510/0x540 pm_suspend+0x16c/0x1d8 ```
The fix has been **Reviewed-by** and **Tested-by** Jon Hunter (NVIDIA kernel engineer), and authored by NVIDIA engineers who are experts on Tegra PMC.
### 2. CODE CHANGE ANALYSIS - The Bug
**Root cause**: tegra186_pmc_wake_syscore_resume() is called by syscore_resume() during system resume. Looking at syscore_resume() in drivers/base/syscore.c:93-110:
```c
void syscore_resume(void)
{
WARN_ONCE(!irqs_disabled(), ...);
list_for_each_entry(syscore, &syscore_list, node)
if (syscore->ops->resume)
syscore->ops->resume(syscore->data);
}
```This runs with **interrupts disabled** but **NOT in hard IRQ context** (i.e., in_hardirq() returns false). However, when generic_handle_irq() is called, it goes through handle_irq_desc() (in kernel/irq/irqdesc.c:658-671):
```c
int handle_irq_desc(struct irq_desc *desc)
{
data = irq_desc_get_irq_data(desc);
if (WARN_ON_ONCE(!in_hardirq() &&
irqd_is_handle_enforce_irqctx(data)))
return -EPERM;
generic_handle_irq_desc(desc);
return 0;
}
```The ARM GIC v2 and GIC v3 interrupt controllers **set IRQD_HANDLE_ENFORCE_IRQCTX on ALL interrupts** they map (confirmed in drivers/irqchip/irq-gic.c:1077 and drivers/irqchip/irq- gic-v3.c:1585). Since Tegra SoCs use ARM GIC, any IRQ going through the Tegra wake domain will have this flag set. Thus, calling generic_handle_irq() from the syscore resume context (not hard IRQ) triggers the WARN_ON_ONCE.
**Why it matters**: Beyond the WARN, the function returns -EPERM without actually handling the interrupt. This means **edge-triggered wake IRQs are silently dropped**, potentially causing missed wakeup events and device misbehavior after resume.
### 3. THE FIX
The fix:
1. Adds struct irq_work wake_work and u32 *wake_status fields to struct tegra_pmc
2. Creates a new tegra186_pmc_wake_handler() function (the IRQ work handler) that contains the logic previously in tegra186_pmc_process_wake_events()
3. tegra186_pmc_wake_syscore_resume() now saves wake status to pmc->wake_status[] and queues the IRQ work via irq_work_queue()
4. The IRQ work handler runs in **hard IRQ context** where generic_handle_irq() is safe
5. Uses IRQ_WORK_INIT_HARD() to ensure hard IRQ context even on PREEMPT_RT kernels
6. Adds a check in tegra186_pmc_wake_syscore_suspend() for unhandled wake IRQs
This is the **correct** fix pattern - using irq_work to defer IRQ handling to hard IRQ context is a well-established kernel pattern. Multiple other drivers have been fixed similarly (see commits 94ec234a16cf3, 118c3ba24d04f, c6a91405ac5cd, f285de79569f9, f460c70125bcb which all converted to generic_handle_irq_safe()).
### 4. DEPENDENCY ANALYSIS
**Critical dependency**: This commit was written against the v6.19 tree which has the new struct syscore API from commit a97fbc3ee3e2a (v6.19-rc1 only). In stable trees (v6.6, v6.12), the functions use the old struct syscore_ops interface:
- tegra186_pmc_wake_syscore_resume(void) (no void *data parameter)
- tegra186_pmc_wake_syscore_suspend(void) (no void *data parameter)
- struct syscore_ops syscore (not struct syscore)
**However**, the core fix logic (adding irq_work, deferring generic_handle_irq()) does NOT depend on the syscore API change. The fix can be backported with minor adaptation:
- The function signatures need to match the old syscore_ops callbacks (no void *data)
- IRQ_WORK_INIT_HARD and irq_work_queue are available in v6.6+
- The irq_work and wake_status additions to struct tegra_pmc are self-contained
**Affected stable trees**: v6.6.y, v6.12.y (and any other active stable tree based on v6.2+, since commit 0474cc8489bda introduced the buggy code in v6.2).
### 5. SCOPE AND RISK
- **Files changed**: 1 file (drivers/soc/tegra/pmc.c)
- **Lines**: ~+60, ~-30 (net ~30 lines added)
- **Subsystem**: Tegra SoC PMC - a self-contained platform driver
- **Risk**: LOW - The change only affects Tegra platforms with wake events. The irq_work mechanism is well-tested kernel infrastructure. The fix has been reviewed and tested by NVIDIA engineers.
- **Regression risk**: Very low - non-Tegra systems are unaffected. The new code path is functionally equivalent but executes in the correct context.
### 6. USER IMPACT
- **Who**: All users of NVIDIA Tegra186+ SoCs (Tegra186, Tegra194, Tegra234) using suspend/resume
- **Severity**: Every system resume triggers a kernel WARNING. Beyond the warning itself, edge-triggered wake IRQs may be silently dropped because handle_irq_desc() returns -EPERM without processing the interrupt
- **Real-world**: Tegra platforms are used in NVIDIA Jetson embedded systems, automotive platforms, and gaming devices (Nintendo Switch)
### 7. CONCLUSION
This is a clear bug fix for a real, reproducible issue that fires on every resume on affected hardware. The fix uses a well-established kernel pattern (irq_work) and is self-contained within a single platform driver file. It has been reviewed and tested by the subsystem maintainers. The fix will need minor adaptation for stable trees (different syscore_ops API), but the core logic is fully applicable. The bug exists in all stable trees from v6.2 onward.
**YES**
drivers/soc/tegra/pmc.c | 104 ++++++++++++++++++++++++++++------------ 1 file changed, 74 insertions(+), 30 deletions(-)
drivers/soc/tegra/pmc.c +74 -30
@@ -28,6 +28,7 @@#include <linux/iopoll.h> #include <linux/irqdomain.h> #include <linux/irq.h> +#include <linux/irq_work.h> #include <linux/kernel.h> #include <linux/of_address.h> #include <linux/of_clk.h>
@@ -468,6 +469,10 @@ struct tegra_pmc {unsigned long *wake_sw_status_map;unsigned long *wake_cntrl_level_map;struct syscore syscore; + +/* Pending wake IRQ processing */ +struct irq_work wake_work; +u32 *wake_status; };static struct tegra_pmc *pmc = &(struct tegra_pmc ) {
@@ -1905,6 +1910,50 @@ static int tegra_pmc_parse_dt(struct tegra_pmc *pmc, struct device_node *np)return 0 ; } +/* translate sc7 wake sources back into IRQs to catch edge triggered wakeups */ +static void tegra186_pmc_wake_handler (struct irq_work *work) +{ + struct tegra_pmc *pmc = container_of(work, struct tegra_pmc, wake_work); +unsigned int i, wake; + +for (i =0 ; i < pmc->soc->max_wake_vectors; i++) { +unsigned long status = pmc->wake_status[i]; + +for_each_set_bit (wake, &status,32 ) { +irq_hw_number_t hwirq = wake + (i *32 ); +struct irq_desc *desc; +unsigned int irq; + + irq =irq_find_mapping (pmc->domain, hwirq); +if (!irq) { +dev_warn (pmc->dev, +"No IRQ found for WAKE#%lu!\n" , + hwirq); +continue ; + } + +dev_dbg (pmc->dev, +"Resume caused by WAKE#%lu mapped to IRQ#%u\n" , + hwirq, irq); + + desc =irq_to_desc (irq); +if (!desc) { +dev_warn (pmc->dev, +"No descriptor found for IRQ#%u\n" , + irq); +continue ; + } + +if (!desc->action || !desc->action->name) +continue ; + +generic_handle_irq (irq); + } + + pmc->wake_status[i] =0 ; + } +} +static int tegra_pmc_init (struct tegra_pmc *pmc) {if (pmc->soc->max_wake_events >0 ) {
@@ -1923,6 +1972,18 @@ static int tegra_pmc_init(struct tegra_pmc *pmc) pmc->wake_cntrl_level_map =bitmap_zalloc (pmc->soc->max_wake_events,GFP_KERNEL );if (!pmc->wake_cntrl_level_map)return -ENOMEM ; + + pmc->wake_status =kcalloc (pmc->soc->max_wake_vectors,sizeof (u32),GFP_KERNEL ); +if (!pmc->wake_status) +return -ENOMEM ; + +/* +* Initialize IRQ work for processing wake IRQs. Must use +* HARD_IRQ variant to run in hard IRQ context on PREEMPT_RT +* because we call generic_handle_irq() which requires hard +* IRQ context. +*/ + pmc->wake_work =IRQ_WORK_INIT_HARD (tegra186_pmc_wake_handler); }if (pmc->soc->init)
@@ -3129,47 +3190,30 @@ static void wke_clear_wake_status(struct tegra_pmc *pmc) } } -/* translate sc7 wake sources back into IRQs to catch edge triggered wakeups */ -static void tegra186_pmc_process_wake_events (struct tegra_pmc *pmc,unsigned int index, -unsigned long status) -{ -unsigned int wake; - -dev_dbg (pmc->dev,"Wake[%d:%d] status=%#lx\n" , (index *32 ) +31 , index *32 , status); - -for_each_set_bit (wake, &status,32 ) { -irq_hw_number_t hwirq = wake +32 * index; -struct irq_desc *desc; -unsigned int irq; - - irq =irq_find_mapping (pmc->domain, hwirq); - - desc =irq_to_desc (irq); -if (!desc || !desc->action || !desc->action->name) { -dev_dbg (pmc->dev,"Resume caused by WAKE%ld, IRQ %d\n" , hwirq, irq); -continue ; - } - -dev_dbg (pmc->dev,"Resume caused by WAKE%ld, %s\n" , hwirq, desc->action->name); -generic_handle_irq (irq); - } -} -static void tegra186_pmc_wake_syscore_resume (void *data) { -u32 status, mask;unsigned int i; +u32 mask;for (i =0 ; i < pmc->soc->max_wake_vectors; i++) { mask =readl (pmc->wake +WAKE_AOWAKE_TIER2_ROUTING (i)); - status =readl (pmc->wake +WAKE_AOWAKE_STATUS_R (i)) & mask; - -tegra186_pmc_process_wake_events (pmc, i, status); + pmc->wake_status[i] =readl (pmc->wake +WAKE_AOWAKE_STATUS_R (i)) & mask; } + +/* Schedule IRQ work to process wake IRQs (if any) */ +irq_work_queue (&pmc->wake_work); }static int tegra186_pmc_wake_syscore_suspend (void *data) { +unsigned int i; + +/* Check if there are unhandled wake IRQs */ +for (i =0 ; i < pmc->soc->max_wake_vectors; i++) +if (pmc->wake_status[i]) +dev_warn (pmc->dev, +"Unhandled wake IRQs pending vector[%u]: 0x%x\n" , + i, pmc->wake_status[i]);wke_read_sw_wake_status (pmc);/* flip the wakeup trigger for dual-edge triggered pads
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 x86/sev: Use kfree_sensitive() when freeing a SNP message descriptor
In reply to Sasha LevinFrom: "Borislav Petkov (AMD)" <[email protected]>
[ Upstream commit af05e558988ed004a20fc4de7d0f80cfbba663f0 ]
Use the proper helper instead of an open-coded variant.
---
LLM Generated explanations, may be completely bogus:
Now I have a thorough understanding. Let me compile my full analysis.
---
## Detailed Analysis
### 1. Commit Message Analysis
The commit message is intentionally understated: "Use the proper helper instead of an open-coded variant." However, this is a **security- relevant fix** flagged by automated static analysis tools. It was:
- **Reported-by: kernel test robot** (automated scanner from Intel)
- **Reported-by: Julia Lawall** (well-known for Coccinelle semantic patches that detect exactly these patterns)
- **Reviewed-by: Tom Lendacky** (AMD SEV subsystem maintainer)
- **Authored-by: Borislav Petkov** (x86 co-maintainer, AMD SEV expert)
### 2. Code Change Analysis
The change is in snp_msg_free() in arch/x86/coco/sev/core.c:
**Before:**
```c memset(mdesc, 0, sizeof(*mdesc)); kfree(mdesc); ```
**After:**
```c kfree_sensitive(mdesc); ```
This is a **critical difference** despite appearing cosmetic:
1. **memset() can be optimized away**: When memset() is followed immediately by kfree(), the compiler can perform dead store elimination (DSE). Since the memory is about to be freed and no code reads the zeroed values, the compiler may determine the memset() is a dead store and remove it entirely. This is increasingly common with modern GCC and Clang optimization levels, and especially with LTO.
2. **kfree_sensitive() uses memzero_explicit()**: This function calls memset() followed by barrier_data(), which is a compiler barrier that prevents the zero operation from being optimized away. This is the canonical kernel API specifically designed for clearing sensitive data before freeing.
3. **kfree_sensitive() clears the entire slab allocation**: Via ksize(), it zeros the whole allocated slab object (which may be larger than sizeof(*mdesc)), while the original code only zeroed sizeof(*mdesc). This provides more thorough clearing of any padding or slab slack space.
### 3. Sensitivity of the Data
The snp_msg_desc structure contains:
- **secret_request and secret_response** (embedded struct snp_guest_msg): These are **double-buffered encrypted/decrypted guest messages** used for communication between the SEV-SNP guest VM and the AMD Secure Processor (ASP). The comment in the struct says "Avoid information leakage by double-buffering shared messages in fields that are in regular encrypted memory." Each snp_guest_msg contains auth tags, sequence numbers, and a full-page payload.
- **vmpck**: Pointer to the VM Platform Communication Key.
- **ctx**: AES-GCM crypto context (freed separately via kfree()).
- **secrets**: Pointer to the SNP secrets page.
- **os_area_msg_seqno**: Message sequence number.
This is AMD SEV-SNP code. The entire purpose of SEV-SNP is to protect guest VM memory from the hypervisor. Leaving cryptographic state and message contents in freed slab memory directly undermines the security guarantees of the technology.
### 4. History and Stable Tree Applicability
- snp_msg_free() was introduced in commit c5529418d0507 ("x86/sev:
Carve out and export SNP guest messaging init routines"), which landed in **v6.14-rc1**.
- The buggy memset() + kfree() pattern was introduced by this same refactoring commit.
- The fix applies to all stable trees from **v6.14 onwards**: 6.14.y, 6.15.y, 6.16.y, 6.17.y, 6.18.y.
- I confirmed the vulnerable code exists in v6.14 and that kfree_sensitive() is available there.
- It does NOT apply to 6.12.y, 6.6.y, or older LTS trees (they use a different code structure in drivers/virt/coco/sev-guest/).
### 5. Risk Assessment
- **Size**: 3 lines changed to 2 lines (net -1 line). One of the smallest possible changes.
- **Risk**: Essentially zero. kfree_sensitive() is the standard kernel API for this exact use case, has been available since v5.9, and is used in 300+ call sites across the kernel.
- **No dependencies**: The fix is entirely self-contained.
- **No behavioral change** from a functional perspective — same memory is freed — only ensures the zeroing actually happens at runtime.
### 6. Classification
This is a **security fix** in **security-critical AMD SEV-SNP code**. The pattern of memset() + kfree() for sensitive data is a well-known anti-pattern that has been the subject of multiple CVEs in other software projects. While the compiler may or may not optimize it away in current builds, it is a latent security vulnerability that could manifest with different compiler versions, optimization flags, or LTO configurations.
### 7. Verdict
This commit:
- Fixes a real security issue (potential leakage of cryptographic data in freed memory)
- Is in security-critical code (AMD SEV-SNP guest messaging)
- Is extremely small and surgical (2 lines -> 1 line)
- Uses the canonical kernel API (kfree_sensitive())
- Has zero risk of regression
- Was reviewed by the AMD SEV maintainer
- Was flagged by two independent automated/semi-automated analysis tools
- Applies cleanly to stable trees v6.14+
- Has no dependencies on other commits
**YES**
arch/x86/coco/sev/core.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
arch/x86/coco/sev/core.c +1 -2
@@ -2008,8 +2008,7 @@ void snp_msg_free(struct snp_msg_desc *mdesc)free_shared_pages (mdesc->request,sizeof (struct snp_guest_msg ));iounmap ((__force void __iomem *)mdesc->secrets); -memset (mdesc,0 ,sizeof (*mdesc)); -kfree (mdesc); +kfree_sensitive (mdesc); }EXPORT_SYMBOL_GPL (snp_msg_free);
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 firmware: arm_ffa: Unmap Rx/Tx buffers on init failure
In reply to Sasha LevinFrom: Haoxiang Li <[email protected]>
[ Upstream commit 9fda364cb78c8b9e1abe4029f877300c94655742 ]
ffa_init() maps the Rx/Tx buffers via ffa_rxtx_map() but on the partition setup failure path it never unmaps them.
Add the missing ffa_rxtx_unmap() call in the error path so that the Rx/Tx buffers are properly released before freeing the backing pages.
Message-Id: <[email protected]>
---
LLM Generated explanations, may be completely bogus:
## Analysis
### 1. Commit Message Analysis
The commit message clearly describes a **resource leak on an error path**: ffa_init() maps Rx/Tx buffers via ffa_rxtx_map(), but when ffa_setup_partitions() fails, the error path frees the backing pages without first unmapping the buffers from the firmware. The fix adds the missing ffa_rxtx_unmap() call. The commit was accepted by the FF-A subsystem maintainer Sudeep Holla.
### 2. Code Change Analysis
The fix is a **single line addition**:
```2071:2071:drivers/firmware/arm_ffa/driver.c
ffa_rxtx_unmap(drv_info->vm_id);
```inserted between ffa_notifications_cleanup() and the free_pages: label.
**Understanding the bug mechanism:**
The ffa_init() function follows this flow:
1. Allocates Rx/Tx buffer pages (lines 2038-2048)
2. Maps them to firmware via ffa_rxtx_map() (line 2050) — this is an FF-A SMC/HVC call to the secure firmware (e.g., TrustZone / Hafnium hypervisor)
3. Sets up partitions via ffa_setup_partitions() (line 2065)
4. If partition setup fails, falls through to free_pages: which frees the backing pages
The problem: Step 4 **never calls ffa_rxtx_unmap()**, so the firmware retains stale references to the freed physical pages. The ffa_exit() function (line 2081-2089) correctly calls ffa_rxtx_unmap() before freeing pages, confirming the error path was simply missing this cleanup.
**Consequences of the bug:**
- The firmware (running in secure world or EL2 hypervisor) still holds mappings to physical addresses of freed kernel pages
- Those freed pages can be reallocated for other kernel purposes
- The firmware could write to these stale mappings, causing **memory corruption** of unrelated kernel data
- On systems with a stage-2 hypervisor, the mapping entries are also leaked
### 3. Bug Origin
The bug was introduced by commit 0c565d16b80074e57e3e56240d13fc6cd6ed0334 ("firmware: arm_ffa: Handle partitions setup failures"), which first made ffa_setup_partitions() return an error and added an error path in ffa_init(). Before this commit, ffa_setup_partitions() was void and never triggered error cleanup. This commit landed in **v6.8**. The error path was later restructured by be61da9385766 (v6.15), but the missing unmap persisted through both versions.
### 4. Classification
This is a **resource leak / firmware-side stale mapping fix** on an error path. It is NOT a new feature, API change, or refactoring — it is purely adding missing cleanup.
### 5. Scope and Risk Assessment
- **Lines changed**: 1 line added
- **Files changed**: 1 file (drivers/firmware/arm_ffa/driver.c)
- **Risk**: Extremely low. The ffa_rxtx_unmap() function is already used in ffa_exit() for normal cleanup. Adding it to the error path is the obvious correct fix. The function itself just issues an FF-A call to the firmware.
- **Regression potential**: Near zero. This only affects the failure path of ffa_setup_partitions(), and calling unmap before freeing is strictly correct.
### 6. User Impact
This affects ARM platforms using FF-A (Firmware Framework for Arm) — notably ARM TrustZone/OP-TEE systems, Arm CCA systems, and virtualized ARM environments with Hafnium. While ffa_setup_partitions() failure is not the most common path, when it does occur, the consequences (firmware holding stale page references, potential memory corruption) are severe.
### 7. Stable Kernel Criteria Check
| Criterion | Met? | |---|---| | Obviously correct and tested | Yes — mirrors `ffa_exit()` cleanup exactly | | Fixes a real bug | Yes — resource leak + potential memory corruption | | Important issue | Yes — firmware-side stale mapping to freed kernel pages | | Small and contained | Yes — 1 line in 1 file | | No new features | Yes — purely error-path cleanup | | Applies cleanly | Yes — for v6.8+ trees (may need minor adjustment for v6.8-v6.14 which used `cleanup_notifs:` label) |
### 8. Dependency Check
The fix depends on 0c565d16b80074e57e3e56240d13fc6cd6ed0334 (v6.8) for the error path to exist, and on be61da9385766 (v6.15) for the exact code structure in the diff. For stable trees between v6.8 and v6.14, a trivially adjusted backport would be needed (placing the unmap call inside the cleanup_notifs: section instead of inline). The fix is self-contained and has no other dependencies.
### Conclusion
This is a textbook error-path resource leak fix: a single line adding missing firmware unmap before freeing the backing pages. The bug leaves the firmware holding stale references to freed kernel memory, which could lead to memory corruption. The fix is minimal, obviously correct (mirrors ffa_exit()), and accepted by the subsystem maintainer. It meets all stable kernel criteria.
**YES**
drivers/firmware/arm_ffa/driver.c | 1 + 1 file changed, 1 insertion(+)
drivers/firmware/arm_ffa/driver.c +1 -0
@@ -2068,6 +2068,7 @@ static int __init ffa_init(void)pr_err ("failed to setup partitions\n" );ffa_notifications_cleanup (); +ffa_rxtx_unmap (drv_info->vm_id); free_pages:if (drv_info->tx_buffer)free_pages_exact (drv_info->tx_buffer, rxtx_bufsz);
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 EDAC/igen6: Add two Intel Amston Lake SoCs support
In reply to Sasha LevinFrom: Qiuxu Zhuo <[email protected]>
[ Upstream commit 41ca2155d62b0b0d217f59e1bce18362d0c2446f ]
Intel Amston Lake SoCs with IBECC (In-Band ECC) capability share the same IBECC registers as Alder Lake-N SoCs. Add two new compute die IDs for Amston Lake SoC products to enable EDAC support.
---
LLM Generated explanations, may be completely bogus:
Now I have all the information needed for a complete analysis. Here is my detailed assessment:
---
## Detailed Analysis: EDAC/igen6: Add two Intel Amston Lake SoCs support
### 1. Commit Message Analysis
The commit adds two new PCI compute die IDs (DID_ASL_SKU2 = 0x4646, DID_ASL_SKU3 = 0x4652) for Intel Amston Lake SoC products to the igen6_edac driver. The commit message explicitly states these SoCs share the same IBECC registers as Alder Lake-N, meaning they reuse the existing adl_n_cfg configuration structure.
- **Author**: Qiuxu Zhuo (Intel) - the primary maintainer and original author of this driver
- **Signed-off-by**: Tony Luck (Intel RAS maintainer) - trusted subsystem maintainer
- **Tested-by**: Jianfeng Gao (Intel) - hardware tested on actual Amston Lake devices
### 2. Code Change Analysis
The diff is minimal - exactly **4 lines added** across a single file:
1. **Two #define additions** (lines after existing DID_ASL_SKU1):
- #define DID_ASL_SKU2 0x4646
- #define DID_ASL_SKU3 0x4652
2. **Two PCI device table entries** (in igen6_pci_tbl[]):
- { PCI_VDEVICE(INTEL, DID_ASL_SKU2), (kernel_ulong_t)&adl_n_cfg }
- { PCI_VDEVICE(INTEL, DID_ASL_SKU3), (kernel_ulong_t)&adl_n_cfg }
Both new entries use the already-existing adl_n_cfg configuration, which has been in the kernel since v6.8 (commit c4a5398991fd2). No new code paths, no new functions, no behavioral changes. This is a pure PCI device ID addition.
### 3. Classification: New Device IDs (Explicitly Allowed Exception)
This is a textbook **new device ID addition** to an existing driver. The stable kernel rules documentation explicitly allows this:
Quoted 1 line
"Adding PCI IDs, USB IDs, ACPI IDs, etc. to existing drivers - These
are trivial one-line additions that enable hardware support"
The pattern is identical to many prior device ID additions to this driver:
- Arizona Beach (b804d7c59aea0) - single PCI ID using adl_n_cfg
- Alder Lake-N extra SKU (65c441ec58224) - single PCI ID using adl_n_cfg
- Arrow Lake-U/H (88150cd9501b9) - PCI IDs using mtl_p_cfg
### 4. Scope and Risk Assessment
- **Lines changed**: 4 added, 0 deleted
- **Files touched**: 1 (drivers/edac/igen6_edac.c)
- **Complexity**: Near-zero - just PCI ID defines and table entries
- **Risk of regression**: Effectively zero. The new PCI IDs will only match hardware that has those specific device IDs. If no such hardware is present, the entries are completely inert. If the hardware IS present, it uses the well-tested adl_n_cfg path that Alder Lake-N, Arizona Beach, and Amston Lake SKU1 all use.
### 5. User Impact
Intel Amston Lake SoCs are described as "tailored for edge computing needs" in the original Amston Lake commit. EDAC (Error Detection and Correction) is critical for:
- **Edge computing and embedded systems** - where these SoCs are deployed
- **Reliability-critical environments** - where memory error detection prevents silent data corruption
- **Enterprise/industrial applications** - Amston Lake targets this market segment
Without these device IDs, users with Amston Lake SKU2/SKU3 hardware get **no IBECC EDAC support** - meaning memory errors go undetected, which can lead to silent data corruption.
### 6. Stability Indicators
- Written by the driver's original author (Qiuxu Zhuo at Intel)
- Signed-off by the RAS subsystem maintainer (Tony Luck)
- Hardware-tested by Intel engineer (Jianfeng Gao)
- The same adl_n_cfg config is already used by 12+ Alder Lake-N SKUs, Arizona Beach, and ASL_SKU1 - extremely well-tested code path
### 7. Dependency Check
This commit depends on commit 099d2db3625b3 ("EDAC/igen6: Add Intel Amston Lake SoCs support") which introduced DID_ASL_SKU1 and the Amston Lake comment section. That commit landed in v6.16-rc1. Both commits would need to be backported together to stable trees that have adl_n_cfg (v6.8+). This is straightforward - both are trivial device ID additions.
The adl_n_cfg configuration exists since v6.8 (commit c4a5398991fd2), so active stable trees 6.8.y, 6.12.y would have the necessary infrastructure. For 6.6.y, the adl_n_cfg is also present since it was added in v6.8-rc1 which branched after 6.6 - actually let me verify:
The adl_n_cfg was introduced by c4a5398991fd2 which is tagged at v6.8-rc1~209^2~8. This means it's in v6.8+. Active stable trees like 6.12.y and 6.6.y need to be checked individually, but 6.6.y would NOT have adl_n_cfg since it predates v6.8. Stable trees 6.8+ would have it.
### Summary
| Criterion | Assessment | |---|---| | Fixes real issue? | Yes - enables EDAC on Amston Lake SKU2/SKU3 hardware | | Obviously correct? | Yes - trivial PCI ID addition | | Small and contained? | Yes - 4 lines, 1 file | | New features? | No - enables existing functionality on new hardware | | Risk? | Near-zero | | Tested? | Yes (Tested-by tag from Intel) | | Category | New Device IDs - explicitly allowed in stable |
This commit is a clear-cut example of the "new device ID" exception to stable kernel rules. It enables EDAC support for real Intel hardware (Amston Lake edge computing SoCs) with zero risk of regression. The change is trivial, authored and tested by Intel engineers, and follows the exact same pattern as numerous prior device ID additions to this driver.
**YES**
drivers/edac/igen6_edac.c | 4 ++++ 1 file changed, 4 insertions(+)
drivers/edac/igen6_edac.c +4 -0
@@ -246,6 +246,8 @@ static struct work_struct ecclog_work;/* Compute did IDs for Amston Lake with IBECC */ #define DID_ASL_SKU1 0x464a +#define DID_ASL_SKU2 0x4646 +#define DID_ASL_SKU3 0x4652/* Compute die IDs for Raptor Lake-P with IBECC */ #define DID_RPL_P_SKU1 0xa706
@@ -628,6 +630,8 @@ static struct pci_device_id igen6_pci_tbl[] = { { PCI_VDEVICE(INTEL, DID_ADL_N_SKU12), (kernel_ulong_t)&adl_n_cfg }, { PCI_VDEVICE(INTEL, DID_AZB_SKU1), (kernel_ulong_t)&adl_n_cfg }, { PCI_VDEVICE(INTEL, DID_ASL_SKU1), (kernel_ulong_t)&adl_n_cfg }, + { PCI_VDEVICE(INTEL, DID_ASL_SKU2), (kernel_ulong_t)&adl_n_cfg }, + { PCI_VDEVICE(INTEL, DID_ASL_SKU3), (kernel_ulong_t)&adl_n_cfg }, { PCI_VDEVICE(INTEL, DID_RPL_P_SKU1), (kernel_ulong_t)&rpl_p_cfg }, { PCI_VDEVICE(INTEL, DID_RPL_P_SKU2), (kernel_ulong_t)&rpl_p_cfg }, { PCI_VDEVICE(INTEL, DID_RPL_P_SKU3), (kernel_ulong_t)&rpl_p_cfg },
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 EDAC/igen6: Add more Intel Panther Lake-H SoCs support
In reply to Sasha LevinFrom: Lili Li <[email protected]>
[ Upstream commit 4c36e6106997b6ad8f4a279b4bdbca3ed6f53c6c ]
Add more Intel Panther Lake-H SoC compute die IDs for EDAC support.
---
LLM Generated explanations, may be completely bogus:
Now I have a comprehensive picture. Let me summarize my analysis.
## Detailed Analysis
### 1. Commit Message Analysis
The commit subject is "EDAC/igen6: Add more Intel Panther Lake-H SoCs support" and the body says "Add more Intel Panther Lake-H SoC compute die IDs for EDAC support." This is purely a **new device ID addition** to an existing driver. The commit is authored by Intel employees (Lili Li, Qiuxu Zhuo) who maintain this specific EDAC driver, and reviewed by the same team.
### 2. Code Change Analysis
The diff is straightforward and mechanical:
**Part 1 - Macro definitions** (lines ~280-290): Adds 10 new #define constants:
- DID_PTL_H_SKU4 through DID_PTL_H_SKU13 with PCI device IDs 0xb003, 0xb004, 0xb005, 0xb008, 0xb011, 0xb014, 0xb015, 0xb028, 0xb029, 0xb02a
**Part 2 - PCI device table** (lines ~675-695): Adds 10 new PCI_VDEVICE entries, all pointing to the existing mtl_p_cfg configuration structure — the same configuration used by the 3 initial PTL-H SKUs.
This is a textbook **PCI device ID addition**. No new logic, no new config structures, no new functions. Every new entry uses the identical mtl_p_cfg configuration as the existing PTL-H SKUs.
### 3. Classification
This falls squarely into the **"New Device IDs" exception** for stable trees:
- The driver (igen6_edac) already exists in stable trees
- The PTL-H configuration (mtl_p_cfg) and initial PTL-H support already exist (since v6.13)
- Only new PCI IDs are being added — enabling additional hardware variants of the same SoC family
### 4. Scope and Risk Assessment
- **Lines changed**: ~20 (10 #defines + 10 PCI table entries)
- **Files touched**: 1 (drivers/edac/igen6_edac.c)
- **Risk**: Extremely low. The PCI device table is a static array of ID- to-config mappings. The new entries only match hardware with those specific PCI IDs. If the hardware isn't present, the entries are completely inert. Zero risk to existing functionality.
- **Complexity**: Trivial — pure data addition with no logic changes.
### 5. Dependency Check
This commit has one critical dependency:
- **Initial PTL-H support commit** (0be9f1af39022): "EDAC/igen6: Add Intel Panther Lake-H SoCs support" — landed in **v6.13**. This commit defined DID_PTL_H_SKU1-SKU3 and the usage of mtl_p_cfg for PTL-H.
Therefore:
- **v6.1 LTS, v6.6 LTS, v6.12 stable**: The prerequisite commit is NOT present. Backporting here would require first backporting the initial PTL-H support.
- **v6.13+ stable** (v6.13.y, v6.14.y, etc.): The prerequisite IS present, so this patch can apply cleanly.
### 6. User Impact
EDAC (Error Detection And Correction) is critical infrastructure for detecting and reporting memory errors. Without these device IDs, users with Intel Panther Lake-H SoCs with these specific compute die IDs will get **no EDAC/ECC error reporting** — meaning memory errors could go undetected, leading to potential silent data corruption on affected hardware. This is exactly the kind of hardware enablement that stable users (particularly enterprise/embedded users) need.
### 7. Stability Indicators
- Reviewed by Intel's EDAC team member (Qiuxu Zhuo)
- Signed off by the RAS maintainer (Tony Luck)
- Same pattern as dozens of prior device ID additions to this driver, all of which have been successfully backported to stable trees (Arrow Lake, Meteor Lake-P/PS, Raptor Lake-P, Alder Lake-N, etc.)
### 8. Verdict
This is a classic, textbook device ID addition:
- Purely mechanical addition of PCI device IDs
- Uses existing infrastructure (no new code paths)
- Enables EDAC support on additional Panther Lake-H hardware variants
- Extremely low risk — inert on systems without matching hardware
- The only concern is dependency on initial PTL-H support (v6.13+), but this is manageable
The commit matches the stable kernel rules for device ID additions perfectly. It's small, contained, obviously correct, and enables real hardware support for real users.
**YES**
drivers/edac/igen6_edac.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)
drivers/edac/igen6_edac.c +20 -0
@@ -274,6 +274,16 @@ static struct work_struct ecclog_work;#define DID_PTL_H_SKU1 0xb000#define DID_PTL_H_SKU2 0xb001#define DID_PTL_H_SKU3 0xb002 +#define DID_PTL_H_SKU4 0xb003 +#define DID_PTL_H_SKU5 0xb004 +#define DID_PTL_H_SKU6 0xb005 +#define DID_PTL_H_SKU7 0xb008 +#define DID_PTL_H_SKU8 0xb011 +#define DID_PTL_H_SKU9 0xb014 +#define DID_PTL_H_SKU10 0xb015 +#define DID_PTL_H_SKU11 0xb028 +#define DID_PTL_H_SKU12 0xb029 +#define DID_PTL_H_SKU13 0xb02a/* Compute die IDs for Wildcat Lake with IBECC */ #define DID_WCL_SKU1 0xfd00
@@ -636,6 +646,16 @@ static struct pci_device_id igen6_pci_tbl[] = { { PCI_VDEVICE(INTEL, DID_PTL_H_SKU1), (kernel_ulong_t)&mtl_p_cfg }, { PCI_VDEVICE(INTEL, DID_PTL_H_SKU2), (kernel_ulong_t)&mtl_p_cfg }, { PCI_VDEVICE(INTEL, DID_PTL_H_SKU3), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU4), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU5), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU6), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU7), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU8), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU9), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU10), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU11), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU12), (kernel_ulong_t)&mtl_p_cfg }, + { PCI_VDEVICE(INTEL, DID_PTL_H_SKU13), (kernel_ulong_t)&mtl_p_cfg }, { PCI_VDEVICE(INTEL, DID_WCL_SKU1), (kernel_ulong_t)&wcl_cfg }, { }, };
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 arm64: tegra: smaug: Add usb-role-switch support
In reply to Sasha LevinFrom: Diogo Ivo <[email protected]>
[ Upstream commit dfa93788dd8b2f9c59adf45ecf592082b1847b7b ]
The USB2 port on Smaug is configured for OTG operation but lacked the required 'usb-role-switch' property, leading to a failed probe and a non-functioning USB port. Add the property along with setting the default role to host.
---
LLM Generated explanations, may be completely bogus:
Now I have all the information needed for a thorough analysis.
---
## Complete Analysis
### 1. COMMIT MESSAGE ANALYSIS
**Subject:** arm64: tegra: smaug: Add usb-role-switch support
**Body:** The commit message clearly states the problem: the USB2 port on Smaug (Google Pixel C) is configured for OTG operation (mode = "otg") but lacked the required usb-role-switch property, **leading to a failed probe and a non-functioning USB port**. This is a clear bug fix, not a feature addition.
**Author/Signoffs:** Authored by Diogo Ivo, signed off by Thierry Reding (NVIDIA's Tegra device tree maintainer).
### 2. CODE CHANGE ANALYSIS
The patch adds exactly **two lines** to a device tree file:
```1811:1812:arch/arm64/boot/dts/nvidia/tegra210-smaug.dts
usb-role-switch;
role-switch-default-mode = "host";
```These are added to the usb2-0 port node within the XUSB pad controller (padctl@7009f000), which already has mode = "otg".
### 3. ROOT CAUSE OF THE BUG
The bug was introduced by commit f67213cee2b35 ("phy: tegra: xusb: Add usb-role-switch support"), which was merged in **v5.7-rc1**. This commit added the following mandatory check to tegra_xusb_usb2_port_parse_dt():
```765:777:drivers/phy/tegra/xusb.c
/* usb-role-switch property is mandatory for OTG/Peripheral
modes */
if (usb2->mode == USB_DR_MODE_PERIPHERAL ||
usb2->mode == USB_DR_MODE_OTG) {
if (of_property_read_bool(np, "usb-role-switch")) {
err = tegra_xusb_setup_usb_role_switch(port);
if (err < 0)
return err;
tegra_xusb_parse_usb_role_default_mode(port);
} else {
dev_err(&port->dev, "usb-role-switch not found
for %s mode",
modes[usb2->mode]);
return -EINVAL;
}
}
```If the DT node has mode = "otg" but **does NOT** have the usb-role- switch property, the driver returns -EINVAL, causing the entire XUSB pad controller probe to **fail**. This means the USB port doesn't work at all.
When commit f67213cee2b35 was merged, the companion DTS updates were done for other Tegra boards (Jetson TX1 in dbf91ff001e54, Jetson Nano in 88d1049e16eaa, Jetson TX2 in 6895c83fda8c9), but **Smaug was missed**. The smaug DTS had mode = "otg" since v4.9 (commit 4d3457826abb1) but was never updated with the required usb-role- switch property.
### 4. IMPACT AND AFFECTED VERSIONS
- **Broken since:** v5.7 (when usb-role-switch became mandatory for OTG mode)
- **Affected stable trees:** All active LTS kernels: **5.15.y, 6.1.y, 6.6.y, 6.12.y**
- **Device:** Google Pixel C (Tegra210 Smaug) — a widely sold consumer tablet
- **Symptom:** Complete USB port probe failure. The XUSB pad controller fails to initialize, making the USB-C port non-functional. This is not a subtle degradation; it's a **total hardware functionality loss**.
### 5. CLASSIFICATION
This falls squarely into the **Device Tree (DT) update** exception category:
- It's a DT fix for existing hardware with an existing driver
- It corrects an incorrect/incomplete hardware description
- It only affects a specific ARM64 platform (Smaug/Pixel C)
- Zero risk of regression to other platforms
### 6. SCOPE AND RISK ASSESSMENT
- **Size:** 2 lines added to a single DTS file
- **Risk:** Extremely low. Adding a usb-role-switch boolean property and a role-switch-default-mode string property to a DT node is trivially safe:
- usb-role-switch enables the existing role-switch code path that is already used by every other Tegra OTG board
- role-switch-default-mode = "host" is simply ignored on kernels before v6.3 (unused DT properties are silently ignored)
- **No functional dependencies:** While this commit appears in a series with c256740c4b1ff (tegra-udc node) and 8acdb94dcfd37 (DisplayPort), this patch is fully independent — it only adds properties to the pad controller's port node, which is separate from the UDC node.
### 7. BACKPORT FEASIBILITY
The patch modifies only the smaug DTS file with a trivial 2-line addition. The context around it (usb2-0 port node with mode = "otg") has been stable since v4.9 and exists identically in all stable trees. This will apply cleanly.
### 8. CONCLUSION
This is a textbook DT bug fix: a missing required property (usb-role- switch) causes a complete probe failure for the USB port on the Google Pixel C tablet. The fix is a trivial 2-line DTS addition that:
1. Fixes a real bug (non-functioning USB port due to probe failure with -EINVAL)
2. Is obviously correct (follows the exact same pattern as every other Tegra OTG board)
3. Is extremely small and contained (2 lines in one DTS file)
4. Has zero regression risk (only affects one specific device)
5. Has been broken since v5.7, affecting all current stable trees
6. Falls into the explicit "DT updates for existing hardware" exception category
The only mild concern is that this commit hasn't merged to mainline yet (it's in a next-* branch). Stable backports typically track mainline commits. However, from a purely technical standpoint, this is an ideal stable fix.
**YES**
arch/arm64/boot/dts/nvidia/tegra210-smaug.dts | 2 ++ 1 file changed, 2 insertions(+)
arch/arm64/boot/dts/nvidia/tegra210-smaug.dts +2 -0
@@ -1809,6 +1809,8 @@ usb2-0 { status = "okay"; vbus-supply = <&usbc_vbus>; mode = "otg"; + usb-role-switch; + role-switch-default-mode = "host"; }; usb3-0 {
Signature · 1 line
2.51.0
Sasha Levin
maintainerPATCH 1/1 Revert "arm64: zynqmp: Add an OP-TEE node to the device tree"
In reply to Sasha LevinFrom: Tomas Melin <[email protected]>
[ Upstream commit c197179990124f991fca220d97fac56779a02c6d ]
This reverts commit 06d22ed6b6635b17551f386b50bb5aaff9b75fbe.
OP-TEE logic in U-Boot automatically injects a reserved-memory node along with optee firmware node to kernel device tree. The injection logic is dependent on that there is no manually defined optee node. Having the node in zynqmp.dtsi effectively breaks OP-TEE's insertion of the reserved-memory node, causing memory access violations during runtime.
---
LLM Generated explanations, may be completely bogus:
Now I have a comprehensive understanding. Let me compile the full analysis.
---
## Detailed Analysis
### 1. COMMIT MESSAGE ANALYSIS
This is a **revert** commit. It reverts 06d22ed6b6635b17551f386b50bb5aaff9b75fbe ("arm64: zynqmp: Add an OP- TEE node to the device tree"), which was merged for v6.9-rc1 (January 2024).
The revert message explains a concrete, real-world problem:
- U-Boot's OP-TEE logic automatically injects a reserved-memory node and optee firmware node into the kernel device tree at boot time
- This U-Boot injection logic checks for the **absence** of a manually- defined optee node — if one already exists, it skips injection entirely
- The original commit added a static optee node to zynqmp.dtsi, which prevents U-Boot from injecting the reserved-memory node
- Without the reserved-memory node, **memory access violations occur during runtime**
This is not speculative — the commit describes actual runtime memory access violations.
### 2. CODE CHANGE ANALYSIS
The change is extremely small and surgical — it removes exactly 5 lines (4 lines of DT node content + 1 blank line):
```195:198:arch/arm64/boot/dts/xilinx/zynqmp.dtsi
optee: optee {
compatible = "linaro,optee-tz";
method = "smc";
};
```This removes a statically-defined OP-TEE node from the ZynqMP base device tree include file (zynqmp.dtsi). The optee label is not referenced by any other DTS file in the xilinx directory tree, so there are no broken cross-references.
### 3. BUG MECHANISM
The bug mechanism is an interaction between the Linux kernel device tree and U-Boot bootloader:
1. **U-Boot's OP-TEE initialization**: U-Boot contains logic that detects if OP-TEE is running and, if so, dynamically modifies the FDT (Flattened Device Tree) before passing it to Linux. It adds:
- A reserved-memory node describing memory regions reserved for OP- TEE
- The optee firmware node pointing to the TEE
2. **Conditional injection**: U-Boot's injection is conditioned on there being **no existing optee node** in the DT. If one is found, U-Boot assumes the DT is already properly configured and skips injection.
3. **The problem**: The static optee node in zynqmp.dtsi has compatible = "linaro,optee-tz" and method = "smc", but it does NOT include the critical reserved-memory region. When U-Boot sees this node, it skips its injection. The kernel then boots with:
- An optee node (so the OP-TEE driver probes and starts communicating with the secure world)
- **No reserved-memory** for OP-TEE (so the kernel may allocate and use memory that OP-TEE also uses)
4. **Result**: Memory access violations at runtime. The kernel and OP- TEE stomp on each other's memory. This is a data corruption and crash bug.
### 4. CLASSIFICATION
This is a **device tree fix** that addresses a **runtime memory corruption/access violation** bug. It falls squarely in the "DT updates for existing hardware" exception category. The severity is high:
- Memory access violations = potential crashes, data corruption, security boundary violations (TEE compromise)
- Affects all ZynqMP boards running OP-TEE with U-Boot
### 5. ORIGINAL COMMIT ANALYSIS
The original commit (06d22ed6b6635) claimed "having the DT node present doesn't cause any side effects." This turned out to be **wrong** — it has a significant side effect with U-Boot's standard OP-TEE injection flow. The original commit was a feature addition ("add it in case someone tries to load OP-TEE") that turned out to cause a real bug.
### 6. SCOPE AND RISK ASSESSMENT
- **Lines changed**: 5 lines removed (1 file)
- **Risk**: Extremely low — removing a DT node that shouldn't have been there
- **Side effects**: None negative. If someone was using the static optee node WITHOUT U-Boot injection, they would need to add it back in their board-specific DT overlay. However, the standard OP-TEE workflow on ZynqMP uses U-Boot injection, and the original commit was explicitly designed as a convenience, not a necessity.
- **Regression potential**: Near zero — the only "regression" is returning to the pre-6.9 behavior, which worked correctly
### 7. AFFECTED STABLE TREES
The original commit (06d22ed6b6635) is present in:
- v6.9 through v6.19 (all current stable trees from 6.9 onward)
- Critically, it's in **6.12.y** (current LTS tree)
The fix is relevant to 6.12.y, 6.13.y, and any other maintained stable tree based on 6.9+.
### 8. DEPENDENCY CHECK
The revert has **zero dependencies**. It simply removes 5 lines from a DTS file. No other code or DT changes are needed. The optee label is not referenced by any other file. The revert will apply cleanly to any stable tree that has the original commit.
### 9. USER IMPACT
- **Who is affected**: All ZynqMP users running OP-TEE with standard U-Boot boot flow
- **Severity**: HIGH — runtime memory access violations can cause:
- Kernel crashes/panics
- OP-TEE secure world corruption
- Silent data corruption
- Potential security boundary violation (TEE memory exposed to normal world)
- **Platform**: Xilinx ZynqMP (common in embedded, industrial, automotive applications)
### 10. TRUST INDICATORS
- Submitted by Tomas Melin (Vaisala), indicating real-world usage and testing
- Accepted by Michal Simek (AMD/Xilinx SoC maintainer), indicating subsystem maintainer approval
- The commit message clearly explains the mechanism and consequences
- Clean revert of a well-identified problematic commit
### CONCLUSION
This is a textbook stable backport candidate:
1. **Fixes a real bug**: Runtime memory access violations on ZynqMP with OP-TEE
2. **Obviously correct**: Clean revert of a commit that introduced the problem
3. **Small and contained**: 5-line removal from a single DTS file
4. **No new features**: Removes a broken feature addition
5. **Zero risk**: Returns to the pre-6.9 behavior that worked correctly
6. **No dependencies**: Self-contained change
7. **Real-world impact**: Affects ZynqMP users with OP-TEE (embedded/industrial platforms)
8. **Applies to active stable trees**: The buggy commit is in 6.12.y LTS and other maintained stables
The fix is small, surgical, and meets all stable kernel criteria.
**YES**
arch/arm64/boot/dts/xilinx/zynqmp.dtsi | 5 ----- 1 file changed, 5 deletions(-)
arch/arm64/boot/dts/xilinx/zynqmp.dtsi +0 -5
@@ -192,11 +192,6 @@ psci { }; firmware { - optee: optee { - compatible = "linaro,optee-tz"; - method = "smc"; - }; - zynqmp_firmware: zynqmp-firmware { compatible = "xlnx,zynqmp-firmware"; #power-domain-cells = <1>;
Signature · 1 line
2.51.0
Hi Sasha
Cc linux-parisc
How did you (or the LLM?) came up with that CC list?!?
Quoted 192 lines · Sasha Levin
From: Helge Deller <[email protected]>
[ Upstream commit
35ac5a728c878594f2ea6c43b57652a16be3c968]Signed-off-by Helge Deller <[email protected]>Signed-off-by Sasha Levin <[email protected]>---
LLM Generated explanations, may be completely bogus:
Now I have all the information I need for a thorough analysis. Here is my complete assessment:
---
## Detailed Analysis: "parisc: Prevent interrupts during reboot"
### 1. COMMIT MESSAGE ANALYSIS
The commit message is minimal: "parisc: Prevent interrupts during reboot". It clearly states the problem it's fixing -- interrupts are not properly disabled during the reboot path on PA-RISC systems. The author is **Helge Deller**, the long-time parisc maintainer and the most authoritative developer for this architecture. The commit has no
Fixes:tag orCc: stable(as expected for autosel candidates).### 2. CODE CHANGE ANALYSIS
The change is **one single effective line of code** plus a comment:
```c /* prevent interrupts during reboot */ set_eiem(0); ```This is inserted into
machine_restart()inarch/parisc/kernel/process.cimmediately afterpdc_chassis_send_status(PDC_CHASSIS_DIRECT_SHUTDOWN)and beforepdc_do_reset().**What
set_eiem(0)does:** On PA-RISC, the EIEM (External Interrupt Enable Mask, Control Register 15) controls which external interrupts can fire. Setting it to 0 **masks all external interrupts at the hardware level**, preventing any interrupt from being delivered to the CPU. This is defined as:```82:82:arch/parisc/include/asm/special_insns.h #define set_eiem(val) mtctl(val, CR_EIEM) ```**The bug:** Without this line, external interrupts remain enabled during the entire reboot sequence. This means:
a) **Deadlock risk in
pdc_do_reset()**: Thepdc_do_reset()function acquirespdc_lockviaspin_lock_irqsave():```1236:1246:arch/parisc/kernel/firmware.c int pdc_do_reset(void) { int retval; unsigned long flags; spin_lock_irqsave(&pdc_lock, flags); retval = mem_pdc_call(PDC_BROADCAST_RESET, PDC_DO_RESET); spin_unlock_irqrestore(&pdc_lock, flags); return retval; } ```While
spin_lock_irqsavedisables local interrupts, the PA-RISC EIEM hardware mask is a separate mechanism. On PA-RISC, the external interrupt delivery path goes through the EIEM -- an interrupt fires only if the corresponding EIEM bit is set AND the EIRR (External Interrupt Request Register) bit is set. If a hardware interrupt fires betweenpdc_chassis_send_status()(which also usespdc_lock) andpdc_do_reset(), or during the firmware calls themselves, it could interfere with the reset process.b) **Interference with firmware reset**:
pdc_do_reset()calls into PDC firmware (mem_pdc_call(PDC_BROADCAST_RESET, PDC_DO_RESET)). Firmware calls on PA-RISC are sensitive to the processor state. An interrupt arriving during or between firmware calls can corrupt the reset sequence, potentially causing the machine to **hang instead of rebooting**.c) **The
gsc_writel(CMD_RESET, COMMAND_GLOBAL)fallback**: Ifpdc_do_reset()returns (on machines that don't implementPDC_BROADCAST_RESET), the code tries a hardware reset viagsc_writel. Interrupts during this path are equally problematic.### 3. ESTABLISHED PATTERN IN PARISC AND OTHER ARCHITECTURES
**PA-RISC internal precedent:**
-parisc_terminate()intraps.cuses the exact same pattern:
set_eiem(0)followed bylocal_irq_disable()before critical shutdown operations (line 428-429)
- The SMP CPU hotplug code (smp.c:481) usesset_eiem(0)to disable all external interrupts when taking a CPU offline**Other architectures ALL disable interrupts before reset:**
- ARM:local_irq_disable()at line 136 ofarch/arm/kernel/reboot.c
- ARM64:local_irq_disable()at line 141 ofarch/arm64/kernel/process.c
- x86:local_irq_disable()at line 100 ofarch/x86/kernel/reboot.c
- xtensa:local_irq_disable()at line 524 ofarch/xtensa/kernel/setup.c
- nios2:local_irq_disable()at line 49 ofarch/nios2/kernel/process.c
- csky:local_irq_disable()at line 25 ofarch/csky/kernel/power.c
- MIPS falcon:local_irq_disable()at line 37 ofarch/mips/lantiq/falcon/reset.cPA-RISC was the **outlier** in not disabling interrupts before reboot. This commit fixes that deficiency.
### 4. CLASSIFICATION
This is a **bug fix** -- specifically fixing a potential hang/crash during reboot caused by unmasked interrupts during firmware calls. It is NOT a new feature, optimization, or cleanup.
### 5. SCOPE AND RISK ASSESSMENT
- **Lines changed**: 3 (1 blank line, 1 comment, 1 functional line)
- **Files touched**: 1 (arch/parisc/kernel/process.c)
- **Subsystem**: parisc architecture, reboot path
- **Risk**: **Extremely low**. Disabling interrupts before a system reset is universally accepted as correct and necessary. Theset_eiem(0)pattern is already used in the same codebase for similar critical paths.### 6. USER IMPACT
- **Who is affected**: All PA-RISC Linux users who reboot their machines
- **What happens without the fix**: Potential hang during reboot, requiring a hard power cycle. This could be intermittent and timing- dependent, making it hard to reproduce but very real.
- **Severity if triggered**: System hang requiring manual intervention (power cycle)### 7. STABILITY INDICATORS
- Author is the subsystem maintainer (highest trust level)
- Pattern is well-established across all other architectures
- The exact sameset_eiem(0)call is used in the pariscparisc_terminate()function already
- The change is trivially small and obviously correct### 8. DEPENDENCY CHECK
- No dependencies on other commits
- Theset_eiem()macro has existed since the earliest parisc code
- Themachine_restart()function is largely unchanged across all stable trees
- This will apply cleanly to any kernel version that has parisc support### Summary
This is a trivially small, obviously correct fix that prevents interrupts during the reboot sequence on PA-RISC systems. Without it, an untimely interrupt can interfere with PDC firmware calls during reset, potentially causing the machine to hang instead of rebooting. Every other Linux architecture disables interrupts before reboot; parisc was the outlier. The fix uses the same
set_eiem(0)pattern already used inparisc_terminate()and SMP CPU offline code. The risk is essentially zero -- there is no conceivable way that disabling interrupts right before a hardware reset could cause a regression.**YES**
arch/parisc/kernel/process.c | 3 +++ 1 file changed, 3 insertions(+)arch/parisc/kernel/process.c +3
-1@@ -85,6 +85,9 @@ void machine_restart(char *cmd)#endif /* set up a new led state on systems shipped with a LED State panel */ pdc_chassis_send_status (PDC_CHASSIS_DIRECT_SHUTDOWN ); + +/* prevent interrupts during reboot */ +set_eiem (0 );/* "Normal" system reset */ pdc_do_reset (); --2.51.0
Sasha Levin
maintainerQuoted 5 lines · Geert Uytterhoeven
Hi Sasha
Cc linux-parisc
How did you (or the LLM?) came up with that CC list?!?
Interesting...
$ ~/linux/scripts/get_maintainer.pl --pattern-depth=1 --no-rolestats --nor --nos 0001-parisc-Prevent-interrupts-during-reboot.patch Neil Armstrong <[email protected]> "Guo Ren (Alibaba Damo Academy)" <[email protected]> Christian Brauner <[email protected]> Geert Uytterhoeven <[email protected]> Andreas Larsson <[email protected]> Helge Deller <[email protected]> Langyan Ye <[email protected]> Simon Schuster <[email protected]>
I think that I'll fix it by replacing --pattern-depth with --nogit --nogit-fallback:
$ ~/linux/scripts/get_maintainer.pl --no-git --nogit-fallback --no-rolestats --nor --nos 0001-parisc-Prevent-interrupts-during-reboot.patch "James E.J. Bottomley" <[email protected]> Helge Deller <[email protected]> [email protected] [email protected]
Thanks for reporting this!
Signature · 2 lines
Thanks, Sasha
Hi Sasha,
Quoted 7 lines · Sasha Levin
Quoted 3 lines · Geert Uytterhoeven
Cc linux-parisc
How did you (or the LLM?) came up with that CC list?!?
Interesting...
Quoted 9 lines
$ ~/linux/scripts/get_maintainer.pl --pattern-depth=1 --no-rolestats --nor --nos 0001-parisc-Prevent-interrupts-during-reboot.patch Neil Armstrong <[email protected]> "Guo Ren (Alibaba Damo Academy)" <[email protected]> Christian Brauner <[email protected]> Geert Uytterhoeven <[email protected]> Andreas Larsson <[email protected]> Helge Deller <[email protected]> Langyan Ye <[email protected]> Simon Schuster <[email protected]>
Still doesn't explain linux-csky?
Quoted 7 lines
I think that I'll fix it by replacing --pattern-depth with --nogit --nogit-fallback:
$ ~/linux/scripts/get_maintainer.pl --no-git --nogit-fallback --no-rolestats --nor --nos 0001-parisc-Prevent-interrupts-during-reboot.patch "James E.J. Bottomley" <[email protected]> Helge Deller <[email protected]> [email protected] [email protected]
Much better!
Gr{oetje,eeting}s,
Geert
Signature · 5 lines
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected] In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds
Sasha Levin
maintainerQuoted 20 lines · Geert Uytterhoeven
Hi Sasha,
Quoted 7 lines · Sasha Levin
Quoted 3 lines · Geert Uytterhoeven
Cc linux-parisc
How did you (or the LLM?) came up with that CC list?!?
Interesting...
Quoted 9 lines
$ ~/linux/scripts/get_maintainer.pl --pattern-depth=1 --no-rolestats --nor --nos 0001-parisc-Prevent-interrupts-during-reboot.patch Neil Armstrong <[email protected]> "Guo Ren (Alibaba Damo Academy)" <[email protected]> Christian Brauner <[email protected]> Geert Uytterhoeven <[email protected]> Andreas Larsson <[email protected]> Helge Deller <[email protected]> Langyan Ye <[email protected]> Simon Schuster <[email protected]>
Quoted 1 line
Still doesn't explain linux-csky?
Oh, that was there because the LLM mentioned csky in it's comparison with other archs and get_maintainer.pl used keyword matching.
If you grab the actual patch that was sent in the AUTOSEL mail:
$ ~/linux/scripts/get_maintainer.pl --pattern-depth=1 --no-rolestats --nor --nos raw Guo Ren <[email protected]> [email protected]
Signature · 2 lines
Thanks, Sasha