Hi Tristan.
Since I looked the patch very roughly, I might be wrong and parania though.
xencomm_copy_from_guest(), xencomm_copy_to_guest() converts
address, then acessses a page.
The Xen/IA64 P2M table is lockless so that the page can be freed
by another cpu and used for other purpose at the
On Fri, Nov 10, 2006 at 09:57:44AM +, Keir Fraser wrote:
On 10/11/06 08:45, Isaku Yamahata [EMAIL PROTECTED] wrote:
- introduce guest_handle_add_offset_after_continuatiton() and
replace guest_handle_add_offset() in memory.c with it.
leave do_multicall() and guest_console_write
: fix_xencomm_copy_tofrom_guest
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 338ceb7b1f09 -r 25cdcc5f21f8 xen/common/xencomm.c
--- a/xen/common/xencomm.c Thu Jan 04 10:58:01 2007 +
+++ b/xen/common/xencomm.c Fri Jan 05 12:06:57 2007 +0900
@@ -119,7 +119,7 @@ xencomm_copy_from_guest(void *to, const
/s
13952:4bd0ea9c499). Can you update and resend these patches?
Thanks,
Keir
On 6/1/07 06:48, Isaku Yamahata [EMAIL PROTECTED] wrote:
On Fri, Jan 05, 2007 at 01:43:10PM +, Keir Fraser wrote:
This one is waiting for final review before being checked in. Some people
want
/type reference count because it is done implicitly
when page table is destructed.
It is x86 specific issue and it isn't the case for ia64 and powerpc.
Introduce two arch hook to address it.
PATCHNAME: gnttab_release_mappings_put_page_and_type
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff
On Mon, Mar 19, 2007 at 10:40:18AM +, Derek Murray wrote:
1. gntdev.patch: This is the main driver, and associated header file.
This patch doesn't work on ia64 and ppc because they doesn't support
GNTMAP_application_map which is x86 speicific. The flag doesn't make
sense for ia64 and ppc.
On Wed, Mar 21, 2007 at 11:29:29AM +, Derek Murray wrote:
blktap appears to map each grant into kernel space and, when not
using auto-translated mode, into user space as well. If it is using
auto-translated mode, then the grant is only mapped into the kernel.
When auto translate
On Thu, Apr 05, 2007 at 12:16:38PM +0100, Jan Beulich wrote:
While the header is only being used by PPC, aren't these two macros broken,
i.e. shouldn't the match the ia64 variants in using the structure base address
from the handle and passing just _off as last argument to xencomm_copy_...()?
size.
Note that struct xencomm_desc itself can't page boundary.
PATCHNAME: remove_xencomm_page_size_limit_xen_side
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 039f2ccb1e38 -r c4d697f2367e xen/arch/ia64/xen/xencomm.c
--- a/xen/arch/ia64/xen/xencomm.c Tue Jul 31 10:30:40 2007 -0600
+++ b
On Wed, Aug 01, 2007 at 02:07:54PM -0500, Hollis Blanchard wrote:
On Wed, 2007-08-01 at 15:36 +0900, Isaku Yamahata wrote:
remove xencomm page size limit.
Currently xencomm has page size limit so that a domain with many memory
(e.g. 100GB+) can't be created.
This patch allows
-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 6f18f5bdeea3 -r 8b6af0333d53 config/ia64.mk
--- a/config/ia64.mkMon Aug 06 15:33:42 2007 +0100
+++ b/config/ia64.mkTue Aug 07 17:25:23 2007 +0900
@@ -3,5 +3,6 @@ CONFIG_IA64_$(XEN_OS) := y
CONFIG_IOEMU := y
CONFIG_XCUTILS := y
+CONFIG_XENCOMM := y
On Fri, Aug 03, 2007 at 09:43:16AM -0500, Hollis Blanchard wrote:
On Fri, 2007-08-03 at 11:12 +0900, Isaku Yamahata wrote:
On Thu, Aug 02, 2007 at 10:00:46AM -0500, Hollis Blanchard wrote:
On Thu, 2007-08-02 at 11:44 +0900, Isaku Yamahata wrote:
But we can issue sequential p2m
. On ia64 it isn't in
straight mapping area so that xencomm_pa() is necessary.
- add gcc bug work around. gcc 4.1.2 on ia64 doesn't handle
properly auto variables with align attribute.
PATCHNAME: fix_xencomm_create_in_common_code
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED])
diff -r
On Tue, Aug 07, 2007 at 05:48:54PM +0900, Isaku Yamahata wrote:
- add gcc bug work around. gcc 4.1.2 on ia64 doesn't handle
properly auto variables with align attribute.
I thought this gcc issue was ia64 specific at first.
But it seems to be generic (non-arch specific) issue.
http
On Fri, Aug 10, 2007 at 01:54:47PM -0500, Hollis Blanchard wrote:
I don't see how we'd ever hit this case though, nor why we'd panic the
domain. How could paddr_to_maddr() ever return an maddr that doesn't
belong to the current domain? If paddr is invalid, an error should be
returned from
This updated patch series is for xencomm consolidation and various fixes
based on Hollis's comment.
Changes from take 1:
xen side varisous fixes and preparation for consolidation
- sorted out the maddr and vaddr issue
- returning an error with warning message instead of panicing domain.
gcc bug work around. gcc 4.1.2 doesn't handle properly
variables on stack with align attribute.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16660
PATCHNAME: fix_xencomm_create_in_common_code
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED])
diff -r d2f9b7e36231 -r 06791935b2cb arch/powerpc
consolidation.
powerpc uses void * for xencommized value, on the other hand IA64 uses
struct xencomm_handle *. Unify it with struct xencomm_handle *.
PATCHNAME: introduce_xencomm_handle
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 06791935b2cb -r 7b88b56c310b arch/powerpc/platforms/xen/hcall.c
struct xencomm_desc::address array
to be more than single page. On IA64 it causes 64GB+ domain creation
failure. This patch generalizes xencomm to allow multipage
xencomm_desc::address array.
PATCHNAME: various_fix_xencomm
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r c362bcee8047
: remove_xencomm_page_size_limit_common_code
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 7b88b56c310b -r d2884d84d3fd drivers/xen/core/xencomm.c
--- a/drivers/xen/core/xencomm.cTue Aug 07 16:54:04 2007 +0900
+++ b/drivers/xen/core/xencomm.cWed Aug 01 15:03:59 2007 +0900
@@ -68,25 +68,54 @@ static int
Thank you for review. I will try to simplfy/clean up it.
Probably I will split the patch into the consolidation part (maddr and vaddr),
bug fix part (page boundary check and get_page()/put_page()),
and new feature part(multipage support).
BTW although I know you need to test it before ack,
how
This take 3 patch series is for xencomm consolidation and various fixes.
I split up the xen side patch into 4 smaller ones and cleaned them up.
Linux side patch is same as previous one.
The issue is the complexity and the multi page support.
So I think that 3/7, 4/7 and 7/7 are moot, other
with
vaddr.
But it isn't the case in xen/ia64. It is necessary to convert maddr to vaddr
to access the page. maddr_to_virt() doesn't convert on powerpc, so it should
work on both archtechture.
PATCHNAME: xencomm_consolidation_maddr_vaddr
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r
: xencomm_trivial_fix
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r cd51fa91956b -r 68867379b785 xen/common/xencomm.c
--- a/xen/common/xencomm.c Sun Aug 12 14:50:02 2007 -0600
+++ b/xen/common/xencomm.c Tue Aug 14 16:44:42 2007 +0900
@@ -232,7 +232,7 @@ xencomm_copy_to_guest(void
paddr
to maddr because xen supports SMP and balloon driver.
Otherwise another vcpu may free the page at the same time.
Such a domain behaviour doesn't make sense, however nothing prevents it.
PATCHNAME: fix_xencomm_invalid_racy_access
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r
: remove_xencomm_page_size_limit_common_code
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r 7b88b56c310b -r d2884d84d3fd drivers/xen/core/xencomm.c
--- a/drivers/xen/core/xencomm.cTue Aug 07 16:54:04 2007 +0900
+++ b/drivers/xen/core/xencomm.cWed Aug 01 15:03:59 2007 +0900
@@ -68,25 +68,54 @@ static int
to be more than single page. On IA64 it causes 64GB+ domain creation
failure. This patch generalizes xencomm to allow multipage
xencomm_desc::address array.
PATCHNAME: xencomm_multiple_page
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r cf7a9141e7f8 -r b53a87fec6dc xen/common/xencomm.c
--- a/xen
On Tue, Aug 14, 2007 at 09:48:00AM -0500, Hollis Blanchard wrote:
However, there are a few places below where you call memcpy() without
checking the result of xencomm_maddr_to_vaddr(). Actually, I see the
same issue in the original code in a few places... We should be very
very careful here,
This xencomm take 4 patch series is for xencomm various fixes.
I fixed its coding style. The resulting code is same as take 3 except
the coding style.
[PATCH 1/3] xencomm take 4: xen side fix invalid or racy access
[PATCH 2/3] xencomm take 4: xen side multiple page support
[PATCH 3/3] xencomm
paddr
to maddr because xen supports SMP and balloon driver.
Otherwise another vcpu may free the page at the same time.
Such a domain behaviour doesn't make sense, however nothing prevents it.
PATCHNAME: fix_xencomm_invalid_racy_access
Signed-off-by: Isaku Yamahata [EMAIL PROTECTED]
diff -r
On Tue, Aug 28, 2007 at 03:45:56PM +0100, Keir Fraser wrote:
On 28/8/07 07:17, Isaku Yamahata [EMAIL PROTECTED] wrote:
+static int
+xencomm_ctxt_next(struct xencomm_ctxt *ctxt, int i)
+{
+BUG_ON(i = ctxt-nr_addrs);
+/* in i == 0 case, we already calculated in xecomm_addr_init
31 matches
Mail list logo