Re: [Crash-utility] [PATCH] Fix pvops Xen detection for arm machine

On 12/16/21 2:36 PM, Qi Zheng wrote: > > > On 12/15/21 3:42 PM, HAGIO KAZUHITO(萩尾 一仁) wrote: >> Hi Qi, Xen folks, >> >> -----Original Message----- >>> Since the xen_start_info on the arm/arm64 platform is static defined: >>> >>> ./arm/xen/enlighten.c:40:static struct start_info _xen_start_info; >>> ./arm/xen/enlighten.c:41:struct start_info *xen_start_info = >>> &_xen_start_info; >>> ./arm/xen/enlighten.c:42:EXPORT_SYMBOL(xen_start_info); >>> >>> The is_pvops_xen() in commit 4badc6229c69f5cd9da7eb7bdf400a53ec6db01a >>> ("Fix pvops Xen detection for kernels >= v4.20") always return TRUE. >>> Then the following error will be reported because p2m_mid_missing >>> and xen_p2m_addr are not defined: >>> >>>     crash: cannot resolve "p2m_top" >>> >>> Fix it by using xen_vcpu_id instead of xen_start_info to detect Xen >>> dumps. >>> >>> Signed-off-by: Qi Zheng <zhengqi.arch(a)bytedance.com&gt; >>> --- >>>   kernel.c | 21 ++++++++++++++++----- >>>   1 file changed, 16 insertions(+), 5 deletions(-) >>> >>> diff --git a/kernel.c b/kernel.c >>> index f4598ea..b216fbe 100644 >>> --- a/kernel.c >>> +++ b/kernel.c >>> @@ -10745,6 +10745,9 @@ is_pvops_xen(void) >>>   { >>>       ulong addr; >>>       char *sym; >>> +    struct syment * sp; >>> +    ulong ptr; >>> +    int i; >>> >>>       if (!PVOPS()) >>>           return FALSE; >>> @@ -10757,11 +10760,19 @@ is_pvops_xen(void) >>>            STREQ(sym, "paravirt_patch_default"))) >>>           return TRUE; >>> >>> -    if (symbol_exists("xen_start_info") && >>> -        readmem(symbol_value("xen_start_info"), KVADDR, &addr, >>> -        sizeof(void *), "xen_start_info", RETURN_ON_ERROR) && >>> -        addr != 0) >>> -        return TRUE; >>> +    sp = per_cpu_symbol_search("xen_vcpu_id"); >>> +    if (sp) { >>> +        for (i = 0; i < NR_CPUS; i++) { >>> +            if (!kt->__per_cpu_offset[i]) >>> +                continue; >> >> Thanks for the patch. >> >> But seems that kt->__per_cpu_offset is initialized after the >> is_pvops_xen() >> call.  So this block always return FALSE? >> >> Is there any other variable that we can use?  or another solution.. > > Or use xen_vcpu_info for arm/arm64 machine and xen_start_info for x86 > machine? Both of these variables are assigned during the initialization > path. Like this: diff --git a/kernel.c b/kernel.c index f4598ea..4b4d789 100644 --- a/kernel.c +++ b/kernel.c @@ -10757,11 +10757,18 @@ is_pvops_xen(void) STREQ(sym, "paravirt_patch_default"))) return TRUE; - if (symbol_exists("xen_start_info") && - readmem(symbol_value("xen_start_info"), KVADDR, &addr, - sizeof(void *), "xen_start_info", RETURN_ON_ERROR) && - addr != 0) + if (machine_type("ARM") || machine_type("ARM64")) { + if (symbol_exists("xen_vcpu_info") && + readmem(symbol_value("xen_vcpu_info"), KVADDR, &addr, + sizeof(void *), "xen_vcpu_info", RETURN_ON_ERROR) && + addr != 0) + return TRUE; + } else if (symbol_exists("xen_start_info") && + readmem(symbol_value("xen_start_info"), KVADDR, &addr, + sizeof(void *), "xen_start_info", RETURN_ON_ERROR) && + addr != 0) { return TRUE; + } return FALSE; } If it make sense, I can send the next version.