[Bug]: ptp_clock_info in ena 2.13 is missing .adjfine

[gemorin]

Preliminary Actions

• I have searched the existing issues and didn't find a duplicate.
• I have followed the AWS official troubleshoot documentation.
• I have followed the driver readme and best practices.

Driver Type

Linux kernel driver for Elastic Network Adapter (ENA)

Driver Tag/Commit

ena 2.13.0

Custom Code

No

OS Platform and Distribution

Linux 6.2+ with ena.phc_enable=1

Bug description

Starting from Linux 6.2 (commit 75ab70ec5cef), .adjfine (introduced in 4.10) is required in struct ptp_clock_info and will trigger a NULL pointer dereference in (at least) ptp_clock_adjtime()

testing this change:

--- a/kernel/linux/ena/ena_phc.c
+++ b/kernel/linux/ena/ena_phc.c
@@ -15,6 +15,15 @@ static int ena_phc_adjfreq(struct ptp_cl
 }
 
 #endif /* ENA_PHC_SUPPORT_ADJFREQ */
+
+#ifdef ENA_PHC_SUPPORT_ADJFINE
+static int ena_phc_adjfine(struct ptp_clock_info *clock_info, long delta)
+{
+       return -EOPNOTSUPP;
+}
+
+#endif /* ENA_PHC_SUPPORT_ADJFINE */
+
 static int ena_phc_adjtime(struct ptp_clock_info *clock_info, s64 delta)
 {
        return -EOPNOTSUPP;
@@ -113,6 +122,9 @@ static struct ptp_clock_info ena_ptp_clo
 #ifdef ENA_PHC_SUPPORT_ADJFREQ
        .adjfreq        = ena_phc_adjfreq,
 #endif /* ENA_PHC_SUPPORT_ADJFREQ */
+#ifdef ENA_PHC_SUPPORT_ADJFINE
+       .adjfine        = ena_phc_adjfine,
+#endif /* ENA_PHC_SUPPORT_ADJFINE */
        .adjtime        = ena_phc_adjtime,
 #ifdef ENA_PHC_SUPPORT_GETTIME64
 #ifdef ENA_PHC_SUPPORT_GETTIME64_EXTENDED
--- a/kernel/linux/ena/kcompat.h
+++ b/kernel/linux/ena/kcompat.h
@@ -995,6 +995,10 @@ static inline bool ktime_after(const kti
 #define ptp_clock_register(info, parent) ptp_clock_register(info)
 #endif
 
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 10, 0)
+#define ENA_PHC_SUPPORT_ADJFINE
+#endif /* ENA_PHC_SUPPORT_ADJFINE */
+
 #endif /* CONFIG_PTP_1588_CLOCK */
 
 #if (LINUX_VERSION_CODE < KERNEL_VERSION(3, 19, 0)) && \

Reproduction steps

1. boot a Linux 6.2+ kernel

Expected Behavior

not crash

Actual Behavior

crash :-)

Additional Data

No response

Relevant log output

abbreviated kernel log:

BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor instruction fetch in kernel mode
#PF: error_code(0x0010) - not-present page
PGD 149927067 P4D 149927067 PUD 143ccf067 PMD 0 
Oops: 0010 [#1] SMP NOPTI
CPU: 0 PID: 9217 Comm: t Tainted: G           O        #1
Hardware name: Amazon EC2/Not Specified, BIOS 1.0 10/16/2017
RIP: 0010:0x0
Code: Unable to access opcode bytes at 0xffffffffffffffd6.
Call Trace:
 <TASK>
 ptp_clock_adjtime+0xd1/0x1d0
 pc_clock_adjtime+0x5b/0xa0
 __do_sys_clock_adjtime+0x8a/0x110
 __x64_sys_clock_adjtime+0x16/0x20
 x64_sys_call+0x1321/0x2020
 do_syscall_64+0x33/0x80
 entry_SYSCALL_64_after_hwframe+0x78/0xe2

Contact Details

[email protected]

[davidarinzon]

Thank you for raising this issue, @gemorin

We will resolve it in the next driver release.