Christoph Hellwig
2004-Jun-20 09:40 UTC
[Ocfs2-devel] [PATCH] fixup journal-related ifdef mess
always use the 2.6 variants and fix up for 2.4 under the hood
Index: src/journal.c
==================================================================---
src/journal.c (revision 1156)
+++ src/journal.c (working copy)
@@ -105,9 +105,17 @@
return status;
}
-#else
-#define ocfs_journal_start journal_start
-#define ocfs_journal_stop journal_stop
+
+#define journal_start(journal, nblocks) \
+ ocfs_journal_start(journal, nblocks)
+#define journal_stop(handle) \
+ ocfs_journal_stop(handle)
+
+/* these two gained another argument during 2.6 */
+#define journal_get_write_access(handle, bh, credits) \
+ journal_get_write_access(handle, bh)
+#define journal_get_undo_access(handle, bh, credits) \
+ journal_get_undo_access(handle, bh)
#endif
@@ -165,7 +173,7 @@
down_read(&osb->journal->trans_barrier);
/* actually start the transaction now */
- retval->k_handle = ocfs_journal_start(journal, max_buffs);
+ retval->k_handle = journal_start(journal, max_buffs);
if (IS_ERR(retval->k_handle)) {
up_read(&osb->journal->trans_barrier);
@@ -411,7 +419,7 @@
/* actually stop the transaction. if we've set h_sync,
* it'll have been commited when we return */
- retval = ocfs_journal_stop(kern_handle);
+ retval = journal_stop(kern_handle);
if (retval < 0) {
LOG_ERROR_STATUS(retval);
LOG_ERROR_STR("Could not commit transaction");
@@ -583,7 +591,7 @@
/* want to force our handle to disk in abort case. */
handle->k_handle->h_sync = 1;
- retval = ocfs_journal_stop(handle->k_handle);
+ retval = journal_stop(handle->k_handle);
if (retval < 0) {
LOG_ERROR_STR("Could not commit aborted transaction!");
LOG_ERROR_STATUS(retval);
@@ -705,19 +713,11 @@
switch (type) {
case OCFS_JOURNAL_ACCESS_CREATE:
case OCFS_JOURNAL_ACCESS_WRITE:
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0)
status = journal_get_write_access(handle->k_handle, bh, NULL);
-#else
- status = journal_get_write_access(handle->k_handle, bh);
-#endif
break;
case OCFS_JOURNAL_ACCESS_UNDO:
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0)
status = journal_get_undo_access(handle->k_handle, bh, NULL);
-#else
- status = journal_get_undo_access(handle->k_handle, bh);
-#endif
break;
default:
On Sun, Jun 20, 2004 at 04:40:00PM +0200, Christoph Hellwig wrote:> always use the 2.6 variants and fix up for 2.4 under the hood > > > Index: src/journal.c > ==================================================================> --- src/journal.c (revision 1156) > +++ src/journal.c (working copy) > @@ -105,9 +105,17 @@ > > return status; > } > -#else > -#define ocfs_journal_start journal_start > -#define ocfs_journal_stop journal_stop > + > +#define journal_start(journal, nblocks) \ > + ocfs_journal_start(journal, nblocks) > +#define journal_stop(handle) \ > + ocfs_journal_stop(handle) > + > +/* these two gained another argument during 2.6 */ > +#define journal_get_write_access(handle, bh, credits) \ > + journal_get_write_access(handle, bh) > +#define journal_get_undo_access(handle, bh, credits) \ > + journal_get_undo_access(handle, bh) > #endifThis breaks with the modversion #define magic. I committed a version with wrapper inlines, like with journal_start/stop.. if you have a better idea, let me know. -Manish
Thanks for all the patches so far. (This patch has been committed, but before I could take a look at it) What's wrong with how this was done before? I agree with the goal of reducing code clutter, and making it easier to be a 2.6 file system, so why can't we do ext3 style thin wrappers around those functions (look at ext3_jbd.h to see what I'm going for) instead of doing things like> +#define journal_start(journal, nblocks) \ > + ocfs_journal_start(journal, nblocks) > +#define journal_stop(handle) \ > + ocfs_journal_stop(handle)which winds up defining a macro with the same name as a function (blech! ;) Of course, ext3 had some other reasons for doing it that way too, but it seemed to me to be a bit more obvious... If there's some documented place where this clashes with the linux kernel coding style, then I'll gladly defer, but otherwise I'm thinking it used to be more clear (of course, I have an obvious bias :) --Mark On Sun, Jun 20, 2004 at 04:40:00PM +0200, Christoph Hellwig wrote:> always use the 2.6 variants and fix up for 2.4 under the hood > > > Index: src/journal.c > ==================================================================> --- src/journal.c (revision 1156) > +++ src/journal.c (working copy) > @@ -105,9 +105,17 @@ > > return status; > } > -#else > -#define ocfs_journal_start journal_start > -#define ocfs_journal_stop journal_stop > + > +#define journal_start(journal, nblocks) \ > + ocfs_journal_start(journal, nblocks) > +#define journal_stop(handle) \ > + ocfs_journal_stop(handle) > + > +/* these two gained another argument during 2.6 */ > +#define journal_get_write_access(handle, bh, credits) \ > + journal_get_write_access(handle, bh) > +#define journal_get_undo_access(handle, bh, credits) \ > + journal_get_undo_access(handle, bh) > #endif > > > @@ -165,7 +173,7 @@ > down_read(&osb->journal->trans_barrier); > > /* actually start the transaction now */ > - retval->k_handle = ocfs_journal_start(journal, max_buffs); > + retval->k_handle = journal_start(journal, max_buffs); > if (IS_ERR(retval->k_handle)) { > up_read(&osb->journal->trans_barrier); > > @@ -411,7 +419,7 @@ > > /* actually stop the transaction. if we've set h_sync, > * it'll have been commited when we return */ > - retval = ocfs_journal_stop(kern_handle); > + retval = journal_stop(kern_handle); > if (retval < 0) { > LOG_ERROR_STATUS(retval); > LOG_ERROR_STR("Could not commit transaction"); > @@ -583,7 +591,7 @@ > /* want to force our handle to disk in abort case. */ > handle->k_handle->h_sync = 1; > > - retval = ocfs_journal_stop(handle->k_handle); > + retval = journal_stop(handle->k_handle); > if (retval < 0) { > LOG_ERROR_STR("Could not commit aborted transaction!"); > LOG_ERROR_STATUS(retval); > @@ -705,19 +713,11 @@ > switch (type) { > case OCFS_JOURNAL_ACCESS_CREATE: > case OCFS_JOURNAL_ACCESS_WRITE: > -#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0) > status = journal_get_write_access(handle->k_handle, bh, NULL); > -#else > - status = journal_get_write_access(handle->k_handle, bh); > -#endif > break; > > case OCFS_JOURNAL_ACCESS_UNDO: > -#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,0) > status = journal_get_undo_access(handle->k_handle, bh, NULL); > -#else > - status = journal_get_undo_access(handle->k_handle, bh); > -#endif > break; > > default: > _______________________________________________ > Ocfs2-devel mailing list > Ocfs2-devel@oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-devel-- Mark Fasheh Software Developer, Oracle Corp mark.fasheh@oracle.com