public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Suman Prakash <suman.p@samsung.com>
To: edk2-devel@lists.01.org
Cc: feng.tian@intel.com, Suman Prakash <suman.p@samsung.com>
Subject: [PATCH-V1] MdeModulePkg/NvmExpressDxe: Memory leak fix in async code flow
Date: Mon, 20 Mar 2017 14:04:55 +0530	[thread overview]
Message-ID: <b5896575bf8c77fe54676ffc4353c3768cd6d05e.1489997874.git.suman.p@samsung.com> (raw)
In-Reply-To: CGME20170320083513epcas1p3296f5fef6610fa19af50669a1465a669@epcas1p3.samsung.com

    MdeModulePkg/NvmExpressDxe: Memory leak fix
    in async code flow

    For async commands, the buffer allocated for Prp list is
    not getting freed, which will cause memory leak for async
    read write command. For example testing async command flow
    with custom application to send multiple read write commands
    were resulting in decrease of available memory page in memmap,
    which eventually resulted in system hang. Hence freeing AsyncRequest->MapData,
    AsyncRequest->MapMeta, AsyncRequest->MapPrpList and AsyncRequest->PrpListHost
    when async command is completed.

    Cc: Feng Tian <feng.tian@intel.com>
    Contributed-under: TianoCore Contribution Agreement 1.0
    Signed-off-by: Suman Prakash <suman.p@samsung.com.com>
---
 MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c    | 23 +++++++++++++++++++++-
 MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h    |  5 +++++
 .../Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c     |  5 +++++
 3 files changed, 32 insertions(+), 1 deletion(-)
 mode change 100644 => 100755 MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c
 mode change 100644 => 100755 MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h
 mode change 100644 => 100755 MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c

diff --git a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c
old mode 100644
new mode 100755
index 39f49bd..ec6af58
--- a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c
+++ b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.c
@@ -548,6 +548,8 @@ ProcessAsyncTaskList (
   QueueId    = 2;
   Cq         = Private->CqBuffer[QueueId] + Private->CqHdbl[QueueId].Cqh;
   HasNewItem = FALSE;
+  PciIo		= Private->PciIo;
+
 
   //
   // Submit asynchronous subtasks to the NVMe Submission Queue
@@ -644,6 +646,26 @@ ProcessAsyncTaskList (
           sizeof(EFI_NVM_EXPRESS_COMPLETION)
           );
 
+	//
+	// Free the resources allocated before cmd submission
+	//
+	if (AsyncRequest->MapData != NULL) {
+		PciIo->Unmap (PciIo, AsyncRequest->MapData);
+	}
+	if (AsyncRequest->MapMeta != NULL) {
+		PciIo->Unmap (PciIo, AsyncRequest->MapMeta);
+	}
+	if (AsyncRequest->MapPrpList != NULL) {
+		PciIo->Unmap (PciIo, AsyncRequest->MapPrpList);
+	}
+	if (AsyncRequest->PrpListHost != NULL) {
+		PciIo->FreeBuffer (
+				PciIo, 
+				AsyncRequest->PrpListNo, 
+				AsyncRequest->PrpListHost
+				);
+	}
+
         RemoveEntryList (Link);
         gBS->SignalEvent (AsyncRequest->CallerEvent);
         FreePool (AsyncRequest);
@@ -666,7 +688,6 @@ ProcessAsyncTaskList (
   }
 
   if (HasNewItem) {
-    PciIo = Private->PciIo;
     Data  = ReadUnaligned32 ((UINT32*)&Private->CqHdbl[QueueId]);
     PciIo->Mem.Write (
                  PciIo,
diff --git a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h
old mode 100644
new mode 100755
index 6a1c257..fa4a34a
--- a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h
+++ b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpress.h
@@ -292,6 +292,11 @@ typedef struct {
 
   EFI_NVM_EXPRESS_PASS_THRU_COMMAND_PACKET *Packet;
   UINT16                                   CommandId;
+  VOID                                     *MapPrpList;
+  UINTN                                    PrpListNo;
+  VOID                                     *PrpListHost;
+  VOID                                     *MapData;
+  VOID                                     *MapMeta;
   EFI_EVENT                                CallerEvent;
 } NVME_PASS_THRU_ASYNC_REQ;
 
diff --git a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c
old mode 100644
new mode 100755
index 2c30009..f0cca72
--- a/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c
+++ b/MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressPassthru.c
@@ -627,6 +627,11 @@ NvmExpressPassThru (
     AsyncRequest->Packet        = Packet;
     AsyncRequest->CommandId     = Sq->Cid;
     AsyncRequest->CallerEvent   = Event;
+    AsyncRequest->MapData	= MapData;
+    AsyncRequest->MapMeta	= MapMeta;
+    AsyncRequest->MapPrpList	= MapPrpList;
+    AsyncRequest->PrpListNo	= PrpListNo;
+    AsyncRequest->PrpListHost	= PrpListHost;
 
     OldTpl = gBS->RaiseTPL (TPL_NOTIFY);
     InsertTailList (&Private->AsyncPassThruQueue, &AsyncRequest->Link);
-- 
1.9.1



       reply	other threads:[~2017-03-20  8:35 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20170320083513epcas1p3296f5fef6610fa19af50669a1465a669@epcas1p3.samsung.com>
2017-03-20  8:34 ` Suman Prakash [this message]
2017-03-21  2:08   ` [PATCH-V1] MdeModulePkg/NvmExpressDxe: Memory leak fix in async code flow Wu, Hao A

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-list 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=b5896575bf8c77fe54676ffc4353c3768cd6d05e.1489997874.git.suman.p@samsung.com \
    --to=devel@edk2.groups.io \
    /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