From: "Egger, Christoph" <chegger@amazon.de>
To: xen-devel@lists.xen.org, Ian Jackson <ian.jackson@eu.citrix.com>,
Stefano Stabellini <stefano.stabellini@eu.citrix.com>,
Ian Campbell <ian.campbell@citrix.com>,
Wei Liu <wei.liu2@citrix.com>,
Liu Jinsong <jinsong.liu@alibaba-inc.com>,
Keir Fraser <keir@xen.org>, Jan Beulich <jbeulich@suse.com>,
Andrew Cooper <andrew.cooper3@citrix.com>
Subject: Re: [PATCH 4/4] tools/xen-mceinj: Pass in GPA when injecting through MSR_MCI_ADDR
Date: Tue, 15 Sep 2015 12:34:23 +0200 [thread overview]
Message-ID: <55F7F42F.3000501@amazon.de> (raw)
In-Reply-To: <20150915102859.GB26047@hzzhang-OptiPlex-9020.sh.intel.com>
On 2015/09/15 12:28, Haozhong Zhang wrote:
> On Tue, Sep 15, 2015 at 12:17:01PM +0200, Egger, Christoph wrote:
>> On 2015/09/15 10:29, Haozhong Zhang wrote:
>>> This patch removes the address translation in xen-mceinj which
>>> translates the guest physical address passed-in through the argument
>>> of '-p' to the host machine address.
>>>
>>> Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
>>
>> Comments inline.
>>
>>> ---
>>> tools/tests/mce-test/tools/xen-mceinj.c | 135 +++++---------------------------
>>> 1 file changed, 19 insertions(+), 116 deletions(-)
>>>
>>> diff --git a/tools/tests/mce-test/tools/xen-mceinj.c b/tools/tests/mce-test/tools/xen-mceinj.c
>>> index 0c2b640..7b9dd13 100644
>>> --- a/tools/tests/mce-test/tools/xen-mceinj.c
>>> +++ b/tools/tests/mce-test/tools/xen-mceinj.c
>>> @@ -238,7 +238,8 @@ static int add_msr_intpose(xc_interface *xc_handle,
>>> uint32_t cpu_nr,
>>> uint32_t flags,
>>> uint64_t msr,
>>> - uint64_t val)
>>> + uint64_t val,
>>> + domid_t domid)
>>> {
>>> uint32_t count;
>>>
>>> @@ -256,6 +257,8 @@ static int add_msr_intpose(xc_interface *xc_handle,
>>> msr_inj.mcinj_cpunr = cpu_nr;
>>> msr_inj.mcinj_flags = flags;
>>> }
>>> + if ( flags & MC_MSRINJ_F_GPADDR )
>>> + msr_inj.mcinj_domid = domid;
>>> msr_inj.mcinj_msr[count].reg = msr;
>>> msr_inj.mcinj_msr[count].value = val;
>>> msr_inj.mcinj_count++;
>>> @@ -268,105 +271,15 @@ static int add_msr_bank_intpose(xc_interface *xc_handle,
>>> uint32_t flags,
>>> uint32_t type,
>>> uint32_t bank,
>>> - uint64_t val)
>>> + uint64_t val,
>>> + domid_t domid)
>>> {
>>> uint64_t msr;
>>>
>>> msr = bank_addr(bank, type);
>>> if ( msr == INVALID_MSR )
>>> return -1;
>>> - return add_msr_intpose(xc_handle, cpu_nr, flags, msr, val);
>>> -}
>>> -
>>> -#define MCE_INVALID_MFN ~0UL
>>> -#define mfn_valid(_mfn) (_mfn != MCE_INVALID_MFN)
>>> -#define mfn_to_pfn(_mfn) (live_m2p[(_mfn)])
>>> -static uint64_t guest_mfn(xc_interface *xc_handle,
>>> - uint32_t domain,
>>> - uint64_t gpfn)
>>> -{
>>> - xen_pfn_t *live_m2p = NULL;
>>> - int ret;
>>> - unsigned long hvirt_start;
>>> - unsigned int pt_levels;
>>> - uint64_t * pfn_buf = NULL;
>>> - unsigned long max_mfn = 0; /* max mfn of the whole machine */
>>> - unsigned long m2p_mfn0;
>>> - unsigned int guest_width;
>>> - long max_gpfn,i;
>>> - uint64_t mfn = MCE_INVALID_MFN;
>>> -
>>> - if ( domain > DOMID_FIRST_RESERVED )
>>> - return MCE_INVALID_MFN;
>>> -
>>> - /* Get max gpfn */
>>> - max_gpfn = do_memory_op(xc_handle, XENMEM_maximum_gpfn, &domain,
>>> - sizeof(domain)) + 1;
>>> - if ( max_gpfn <= 0 )
>>> - err(xc_handle, "Failed to get max_gpfn 0x%lx", max_gpfn);
>>> -
>>> - Lprintf("Maxium gpfn for dom %d is 0x%lx", domain, max_gpfn);
>>> -
>>> - /* Get max mfn */
>>> - if ( !get_platform_info(xc_handle, domain,
>>> - &max_mfn, &hvirt_start,
>>> - &pt_levels, &guest_width) )
>>> - err(xc_handle, "Failed to get platform information");
>>> -
>>> - /* Get guest's pfn list */
>>> - pfn_buf = calloc(max_gpfn, sizeof(uint64_t));
>>> - if ( !pfn_buf )
>>> - err(xc_handle, "Failed to alloc pfn buf");
>>> -
>>> - ret = xc_get_pfn_list(xc_handle, domain, pfn_buf, max_gpfn);
>>> - if ( ret < 0 ) {
>>> - free(pfn_buf);
>>> - err(xc_handle, "Failed to get pfn list %x", ret);
>>> - }
>>> -
>>> - /* Now get the m2p table */
>>> - live_m2p = xc_map_m2p(xc_handle, max_mfn, PROT_READ, &m2p_mfn0);
>>> - if ( !live_m2p )
>>> - err(xc_handle, "Failed to map live M2P table");
>>> -
>>> - /* match the mapping */
>>> - for ( i = 0; i < max_gpfn; i++ )
>>> - {
>>> - uint64_t tmp;
>>> - tmp = pfn_buf[i];
>>> -
>>> - if (mfn_valid(tmp) && (mfn_to_pfn(tmp) == gpfn))
>>> - {
>>> - mfn = tmp;
>>> - Lprintf("We get the mfn 0x%lx for this injection", mfn);
>>> - break;
>>> - }
>>> - }
>>> -
>>> - munmap(live_m2p, M2P_SIZE(max_mfn));
>>> -
>>> - free(pfn_buf);
>>> - return mfn;
>>> -}
>>> -
>>> -static uint64_t mca_gpfn_to_mfn(xc_interface *xc_handle,
>>> - uint32_t domain,
>>> - uint64_t gfn)
>>> -{
>>> - uint64_t index;
>>> - long max_gpfn;
>>> -
>>> - /* If domain is xen, means we want pass index directly */
>>> - if ( domain == DOMID_XEN )
>>> - return gfn;
>>> -
>>> - max_gpfn = do_memory_op(xc_handle, XENMEM_maximum_gpfn, &domain,
>>> - sizeof(domain)) + 1;
>>> - if ( max_gpfn <= 0 )
>>> - err(xc_handle, "Failed to get max_gpfn 0x%lx", max_gpfn);
>>> - index = gfn % max_gpfn;
>>> -
>>> - return guest_mfn(xc_handle, domain, index);
>>> + return add_msr_intpose(xc_handle, cpu_nr, flags, msr, val, domid);
>>> }
>>>
>>> static int inject_mcg_status(xc_interface *xc_handle,
>>> @@ -374,7 +287,7 @@ static int inject_mcg_status(xc_interface *xc_handle,
>>> uint64_t val)
>>> {
>>> return add_msr_intpose(xc_handle, cpu_nr, MC_MSRINJ_F_INTERPOSE,
>>> - MSR_IA32_MCG_STATUS, val);
>>> + MSR_IA32_MCG_STATUS, val, 0);
>>
>> Why do you hardcode 0 here?
>> I think you should pass domid in here so that the DomU sees the right
>> status value.
>
> If the flag MC_MSRINJ_F_GPADDR is not given (which is the case here),
> the domain id is not used when do_mca() handles the MCE injection, so
> I just pass 0 as a placeholder here. But, yes, passing domid instead
> keeps the right domain status. I'll change it in the next version.
Thanks. This is much less error prone to future changes.
Christoph
>
>>
>>> }
>>>
>>> static int inject_mci_status(xc_interface *xc_handle,
>>> @@ -383,7 +296,7 @@ static int inject_mci_status(xc_interface *xc_handle,
>>> uint64_t val)
>>> {
>>> return add_msr_bank_intpose(xc_handle, cpu_nr, MC_MSRINJ_F_INTERPOSE,
>>> - MCi_type_STATUS, bank, val);
>>> + MCi_type_STATUS, bank, val, 0);
>>
>> Why do you hardcode 0 here?
>> I think you should pass domid in here so that the DomU sees the right
>> status value.
>>
>
> the same as above
>
>>> }
>>>
>>> static int inject_mci_misc(xc_interface *xc_handle,
>>> @@ -392,22 +305,23 @@ static int inject_mci_misc(xc_interface *xc_handle,
>>> uint64_t val)
>>> {
>>> return add_msr_bank_intpose(xc_handle, cpu_nr, MC_MSRINJ_F_INTERPOSE,
>>> - MCi_type_MISC, bank, val);
>>> + MCi_type_MISC, bank, val, 0);
>>> }
>>>
>>> static int inject_mci_addr(xc_interface *xc_handle,
>>> uint32_t cpu_nr,
>>> uint64_t bank,
>>> - uint64_t val)
>>> + uint64_t val,
>>> + domid_t domid)
>>> {
>>> - return add_msr_bank_intpose(xc_handle, cpu_nr, MC_MSRINJ_F_INTERPOSE,
>>> - MCi_type_ADDR, bank, val);
>>> + return add_msr_bank_intpose(xc_handle, cpu_nr,
>>> + MC_MSRINJ_F_INTERPOSE | MC_MSRINJ_F_GPADDR,
>>> + MCi_type_ADDR, bank, val, domid);
>>> }
>>>
>>> static int inject(xc_interface *xc_handle, struct mce_info *mce,
>>> uint32_t cpu_nr, uint32_t domain, uint64_t gaddr)
>>> {
>>> - uint64_t gpfn, mfn, haddr;
>>> int ret = 0;
>>>
>>> ret = inject_mcg_status(xc_handle, cpu_nr, mce->mcg_stat);
>>> @@ -424,12 +338,7 @@ static int inject(xc_interface *xc_handle, struct mce_info *mce,
>>> if ( ret )
>>> err(xc_handle, "Failed to inject MCi_MISC MSR");
>>>
>>> - gpfn = gaddr >> PAGE_SHIFT;
>>> - mfn = mca_gpfn_to_mfn(xc_handle, domain, gpfn);
>>> - if (!mfn_valid(mfn))
>>> - err(xc_handle, "The MFN is not valid");
>>> - haddr = (mfn << PAGE_SHIFT) | (gaddr & (PAGE_SIZE - 1));
>>> - ret = inject_mci_addr(xc_handle, cpu_nr, mce->bank, haddr);
>>> + ret = inject_mci_addr(xc_handle, cpu_nr, mce->bank, gaddr, domain);
>>> if ( ret )
>>> err(xc_handle, "Failed to inject MCi_ADDR MSR");
>>>
>>> @@ -507,7 +416,7 @@ int main(int argc, char *argv[])
>>> uint32_t domid;
>>> xc_interface *xc_handle;
>>> int cpu_nr;
>>> - int64_t gaddr, gpfn, mfn, haddr, max_gpa;
>>> + uint64_t gaddr, max_gpa;
>>>
>>> /* Default Value */
>>> domid = DOMID_XEN;
>>> @@ -563,16 +472,10 @@ int main(int argc, char *argv[])
>>> Lprintf("get gaddr of error inject is: 0x%lx", gaddr);
>>>
>>> if ( dump ) {
>>> - gpfn = gaddr >> PAGE_SHIFT;
>>> - mfn = mca_gpfn_to_mfn(xc_handle, domid, gpfn);
>>> - if (!mfn_valid(mfn))
>>> - err(xc_handle, "The MFN is not valid");
>>> - haddr = (mfn << PAGE_SHIFT) | (gaddr & (PAGE_SIZE - 1));
>>> if ( domid == DOMID_XEN )
>>> - Lprintf("Xen: mfn=0x%lx, haddr=0x%lx", mfn, haddr);
>>> + Lprintf("Xen: gaddr=0x%lx", gaddr);
>>> else
>>> - Lprintf("Dom%d: gaddr=0x%lx, gpfn=0x%lx, mfn=0x%lx, haddr=0x%lx",
>>> - domid, gaddr, gpfn, mfn, haddr);
>>> + Lprintf("Dom%d: gaddr=0x%lx", domid, gaddr);
>>> goto out;
>>> }
>>>
>>>
Amazon Development Center Germany GmbH
Krausenstr. 38
10117 Berlin
Geschaeftsfuehrer: Dr. Ralf Herbrich, Christian Schlaeger
Ust-ID: DE289237879
Eingetragen am Amtsgericht Charlottenburg HRB 149173 B
next prev parent reply other threads:[~2015-09-15 10:34 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-15 8:29 [PATCH 0/4] Fix tools/xen-mceinj to handle >=4GB guest memory Haozhong Zhang
2015-09-15 8:29 ` [PATCH 1/4] x86/mce: Fix code style Haozhong Zhang
2015-09-15 8:47 ` Egger, Christoph
2015-09-15 13:13 ` Jan Beulich
2015-09-15 13:19 ` Haozhong Zhang
2015-09-15 8:29 ` [PATCH 2/4] tools/mceinject: " Haozhong Zhang
2015-09-15 8:48 ` Egger, Christoph
2015-09-15 10:15 ` Ian Campbell
2015-09-15 13:11 ` Wei Liu
2015-09-15 8:29 ` [PATCH 3/4] x86/mce: Translate passed-in GPA to host machine address Haozhong Zhang
2015-09-15 9:14 ` Egger, Christoph
2015-09-15 9:48 ` Haozhong Zhang
2015-09-15 13:24 ` Jan Beulich
2015-09-15 13:42 ` Haozhong Zhang
2015-09-15 13:47 ` Jan Beulich
2015-09-15 13:51 ` Haozhong Zhang
2015-09-15 13:50 ` Andrew Cooper
2015-09-16 0:38 ` Haozhong Zhang
2015-09-15 8:29 ` [PATCH 4/4] tools/xen-mceinj: Pass in GPA when injecting through MSR_MCI_ADDR Haozhong Zhang
2015-09-15 10:02 ` Wei Liu
2015-09-15 10:08 ` Egger, Christoph
2015-09-15 10:16 ` Wei Liu
2015-09-15 10:17 ` Egger, Christoph
2015-09-15 10:28 ` Haozhong Zhang
2015-09-15 10:34 ` Egger, Christoph [this message]
2015-09-15 10:20 ` Ian Campbell
2015-09-15 10:38 ` Haozhong Zhang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55F7F42F.3000501@amazon.de \
--to=chegger@amazon.de \
--cc=andrew.cooper3@citrix.com \
--cc=ian.campbell@citrix.com \
--cc=ian.jackson@eu.citrix.com \
--cc=jbeulich@suse.com \
--cc=jinsong.liu@alibaba-inc.com \
--cc=keir@xen.org \
--cc=stefano.stabellini@eu.citrix.com \
--cc=wei.liu2@citrix.com \
--cc=xen-devel@lists.xen.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).