| Message ID | 20200203165812.21089-1-julien@xen.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | xen/x86: p2m: Don't initialize slot 0 of the P2M | expand |
On 2/3/20 4:58 PM, Julien Grall wrote: > From: Julien Grall <jgrall@amazon.com> > > It is not entirely clear why the slot 0 of each p2m should be populated > with empty page-tables. The commit introducing it 759af8e3800 "[HVM] > Fix 64-bit HVM domain creation." does not contain meaningful > explanation except that it was necessary for shadow. Tim, any ideas here? > As we don't seem to have a good explanation why this is there, drop the > code completely. > > This was tested by successfully booting a HVM with shadow enabled. > > Signed-off-by: Julien Grall <jgrall@amazon.com> > > --- > > I don't know for sure if this is going to break a setup. I have tried > HVM guest with hap={0, 1} without any trouble. I am happy to try more > setup if you have any in mind. > > If this break a setup, then please describe the setup and I will send a > documentation patch instead. This is a somewhat risky strategy. Other than code clean-up, is there any advantage to removing this code at the moment? -George
Hi, On 03/02/2020 17:10, George Dunlap wrote: > On 2/3/20 4:58 PM, Julien Grall wrote: >> From: Julien Grall <jgrall@amazon.com> >> >> It is not entirely clear why the slot 0 of each p2m should be populated >> with empty page-tables. The commit introducing it 759af8e3800 "[HVM] >> Fix 64-bit HVM domain creation." does not contain meaningful >> explanation except that it was necessary for shadow. > > Tim, any ideas here? > >> As we don't seem to have a good explanation why this is there, drop the >> code completely. >> >> This was tested by successfully booting a HVM with shadow enabled. >> >> Signed-off-by: Julien Grall <jgrall@amazon.com> >> >> --- >> >> I don't know for sure if this is going to break a setup. I have tried >> HVM guest with hap={0, 1} without any trouble. I am happy to try more >> setup if you have any in mind. >> >> If this break a setup, then please describe the setup and I will send a >> documentation patch instead. > > This is a somewhat risky strategy. Other than code clean-up, is there > any advantage to removing this code at the moment? If Tim doesn't have an explanation, then we have two solutions: 1) Checkin the code and see if that breaks 2) Keep code we have no clue why it is there I understand that the former is risky, but the latter is not very ideal either because if we can't explain the reason now, then it is unlikely that we would in the future. Regarding the advantage of removing it, I am looking at liveupdate and how to keep the P2M around. I am trying to limit the number of "if (liveupdate)" within the code. So any cleanup would be beneficial. Cheers,
On 2/3/20 5:22 PM, Julien Grall wrote: > Hi, > > On 03/02/2020 17:10, George Dunlap wrote: >> On 2/3/20 4:58 PM, Julien Grall wrote: >>> From: Julien Grall <jgrall@amazon.com> >>> >>> It is not entirely clear why the slot 0 of each p2m should be populated >>> with empty page-tables. The commit introducing it 759af8e3800 "[HVM] >>> Fix 64-bit HVM domain creation." does not contain meaningful >>> explanation except that it was necessary for shadow. >> >> Tim, any ideas here? >> >>> As we don't seem to have a good explanation why this is there, drop the >>> code completely. >>> >>> This was tested by successfully booting a HVM with shadow enabled. >>> >>> Signed-off-by: Julien Grall <jgrall@amazon.com> >>> >>> --- >>> >>> I don't know for sure if this is going to break a setup. I have tried >>> HVM guest with hap={0, 1} without any trouble. I am happy to try more >>> setup if you have any in mind. >>> >>> If this break a setup, then please describe the setup and I will send a >>> documentation patch instead. >> >> This is a somewhat risky strategy. Other than code clean-up, is there >> any advantage to removing this code at the moment? > > If Tim doesn't have an explanation, then we have two solutions: > 1) Checkin the code and see if that breaks > 2) Keep code we have no clue why it is there It is probably early enough in the dev cycle to do this. Also, it's not clear to me what kind of bug the code you're deleting would fix. If you read a not-present entry, you should get INVALID_MFN anyway. Unless you were calling p2m_get_entry_query(), which I'm pretty sure hadn't been introduced at this point. > I understand that the former is risky, but the latter is not very ideal > either because if we can't explain the reason now, then it is unlikely > that we would in the future. > > Regarding the advantage of removing it, I am looking at liveupdate and > how to keep the P2M around. I am trying to limit the number of "if > (liveupdate)" within the code. So any cleanup would be beneficial. OK, thanks. -George
On 03/02/2020 17:37, George Dunlap wrote: > On 2/3/20 5:22 PM, Julien Grall wrote: >> Hi, >> >> On 03/02/2020 17:10, George Dunlap wrote: >>> On 2/3/20 4:58 PM, Julien Grall wrote: >>>> From: Julien Grall <jgrall@amazon.com> >>>> >>>> It is not entirely clear why the slot 0 of each p2m should be populated >>>> with empty page-tables. The commit introducing it 759af8e3800 "[HVM] >>>> Fix 64-bit HVM domain creation." does not contain meaningful >>>> explanation except that it was necessary for shadow. >>> >>> Tim, any ideas here? >>> >>>> As we don't seem to have a good explanation why this is there, drop the >>>> code completely. >>>> >>>> This was tested by successfully booting a HVM with shadow enabled. >>>> >>>> Signed-off-by: Julien Grall <jgrall@amazon.com> >>>> >>>> --- >>>> >>>> I don't know for sure if this is going to break a setup. I have tried >>>> HVM guest with hap={0, 1} without any trouble. I am happy to try more >>>> setup if you have any in mind. >>>> >>>> If this break a setup, then please describe the setup and I will send a >>>> documentation patch instead. >>> >>> This is a somewhat risky strategy. Other than code clean-up, is there >>> any advantage to removing this code at the moment? >> >> If Tim doesn't have an explanation, then we have two solutions: >> 1) Checkin the code and see if that breaks >> 2) Keep code we have no clue why it is there > > It is probably early enough in the dev cycle to do this. > > Also, it's not clear to me what kind of bug the code you're deleting > would fix. If you read a not-present entry, you should get INVALID_MFN > anyway. Unless you were calling p2m_get_entry_query(), which I'm pretty > sure hadn't been introduced at this point. I can't find this function you mention in staging. Was it removed recently? The code is allocating all page-tables for _gfn(0). I would not expect the common code to care whether a table is allocated or not. So this would suggest that an internal implementation (of the shadow?) is relying on this. However, I can't find anything obvious suggesting that is necessary. If there was anything, I would expect to happen during domain creation, as neither Xen nor a guest could rely on this (there are way to make those pages disappear with the MEMORY op hypercall). Cheers,
At 18:31 +0000 on 03 Feb (1580754711), Julien Grall wrote: > On 03/02/2020 17:37, George Dunlap wrote: > > On 2/3/20 5:22 PM, Julien Grall wrote: > >> On 03/02/2020 17:10, George Dunlap wrote: > >>> On 2/3/20 4:58 PM, Julien Grall wrote: > >>>> From: Julien Grall <jgrall@amazon.com> > >>>> > >>>> It is not entirely clear why the slot 0 of each p2m should be populated > >>>> with empty page-tables. The commit introducing it 759af8e3800 "[HVM] > >>>> Fix 64-bit HVM domain creation." does not contain meaningful > >>>> explanation except that it was necessary for shadow. > >>> > >>> Tim, any ideas here? Afraid not, sorry. I can't think what would rely on the tables being allocated for slot 0 in particular. Maybe there's something later that adds other entries in the bottom 2MB and can't handle a table allocation failure? > > Also, it's not clear to me what kind of bug the code you're deleting > > would fix. If you read a not-present entry, you should get INVALID_MFN > > anyway. Unless you were calling p2m_get_entry_query(), which I'm pretty > > sure hadn't been introduced at this point. > > I can't find this function you mention in staging. Was it removed recently? > > The code is allocating all page-tables for _gfn(0). I would not expect > the common code to care whether a table is allocated or not. So this > would suggest that an internal implementation (of the shadow?) is > relying on this. > > However, I can't find anything obvious suggesting that is necessary. If > there was anything, I would expect to happen during domain creation, as > neither Xen nor a guest could rely on this (there are way to make those > pages disappear with the MEMORY op hypercall). That may not have been true at the time (and so whatever it was that neede this may have been fixed when it became true?) Cheers, Tim.
On 2/3/20 4:58 PM, Julien Grall wrote: > From: Julien Grall <jgrall@amazon.com> > > It is not entirely clear why the slot 0 of each p2m should be populated > with empty page-tables. The commit introducing it 759af8e3800 "[HVM] > Fix 64-bit HVM domain creation." does not contain meaningful > explanation except that it was necessary for shadow. > > As we don't seem to have a good explanation why this is there, drop the > code completely. > > This was tested by successfully booting a HVM with shadow enabled. > > Signed-off-by: Julien Grall <jgrall@amazon.com> Since nobody knows why it's here, and it doesn't look like it should have any effect: Acked-by: George Dunlap <george.dunlap@citrix.com>
Hi George, On 06/02/2020 12:08, George Dunlap wrote: > On 2/3/20 4:58 PM, Julien Grall wrote: >> From: Julien Grall <jgrall@amazon.com> >> >> It is not entirely clear why the slot 0 of each p2m should be populated >> with empty page-tables. The commit introducing it 759af8e3800 "[HVM] >> Fix 64-bit HVM domain creation." does not contain meaningful >> explanation except that it was necessary for shadow. >> >> As we don't seem to have a good explanation why this is there, drop the >> code completely. >> >> This was tested by successfully booting a HVM with shadow enabled. >> >> Signed-off-by: Julien Grall <jgrall@amazon.com> > > Since nobody knows why it's here, and it doesn't look like it should > have any effect: > > Acked-by: George Dunlap <george.dunlap@citrix.com> Thank you! I have now committed the patch. Cheers,
diff --git a/xen/arch/x86/mm/p2m.c b/xen/arch/x86/mm/p2m.c index 49cc138362..961bed0635 100644 --- a/xen/arch/x86/mm/p2m.c +++ b/xen/arch/x86/mm/p2m.c @@ -688,7 +688,6 @@ int p2m_alloc_table(struct p2m_domain *p2m) { mfn_t top_mfn; struct domain *d = p2m->domain; - int rc = 0; p2m_lock(p2m); @@ -721,19 +720,8 @@ int p2m_alloc_table(struct p2m_domain *p2m) if ( hap_enabled(d) ) iommu_share_p2m_table(d); - P2M_PRINTK("populating p2m table\n"); - - /* Initialise physmap tables for slot zero. Other code assumes this. */ - p2m->defer_nested_flush = 1; - rc = p2m_set_entry(p2m, _gfn(0), INVALID_MFN, PAGE_ORDER_4K, - p2m_invalid, p2m->default_access); - p2m->defer_nested_flush = 0; p2m_unlock(p2m); - if ( !rc ) - P2M_PRINTK("p2m table initialised for slot zero\n"); - else - P2M_PRINTK("failed to initialise p2m table for slot zero (%d)\n", rc); - return rc; + return 0; } /*